AD-fuS0 558 A DESIGN ANALYSIS AND IMPLEMENTATION OF A USER-FRIENDLY 1/2INTERFACE FOR THE UNIX OPERATING SYSTEM(U) NAVALPOSTGRADURTE SCHOOL MONTEREY CA F E GROENERT JUN 84

UNCLSSIFIED F/G 9/2 NL

EEEEhh~momEI lfflllffllfffllff

-. i OZ

3322

I -~ .

de

7.

ILE NAVAL POSTGRADUATE SCHOOLMonterey, California

THESIS <

A DESIGN ANALYSIS AND IMPLEMENTATION OF AUSER-FRIENDLY INTERFACE FOR THE

UNIX OPERATING SYSTEM

by

Frederick Earl Groenert, Jr.

June 1984

Thesis Advisor: George A. Rahe

Approved for public release; distribution unlimited

....... OVA A.......

SSCuRITV CLASSIFICATION OF THIS PAGE '3MO. bANSe 410160 940_____________

RKAD MEUUCTMOSREPORT DOCUMENTATION PAGE DIPORE 2 , 0E~l FORKT. 111PORT NUMBER TACCUSSION g E TSCTAO Umean

4. TITLE (admid Sbfto) S. TYPE of REPORT a PERIOD COVERED

A Design Analysis and Implementation of a Junter' 1984 sisUser-Friendly Interface for the Unix 6. PEFRMN oRe. REPORT Muma"Operating System____________

7. AUTMOR(8) 6. CONTRACT ONGRANT UMER(N)

Frederick Earl Groerkert, Jr.

9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELELMENT. PROJE19CT. TASK

Naval Postgraduate SchoolMonterey, California 93943

11. CONTRa ,LIMO OFFICE NAME ANgAWD09~ I2. REPORT DATE

Nava ~' stg rauat ScoolJune 1984Monterey, California 93943 13. NUMBER OFPAGES

15814. MONITORING AGENCY NAMIE &ADDRES$S(I dSiteut btee Controlling 01St.) IS. SECURITY CLASS. (of Ole. report)

UNCLASSIFIED

IaD6Ct. A3S1IFI CATION/ DOWNGRADINGSCM EDLE

If. DISTRISUTION STATEMENT (of tie Report)

Approved for public release; distribution unlimited

17. DISTRIBUTION STATEMENT (of the obetroct entered In Stock ". It Elffment beet Repeef) .*

IS. SUPPLEMENTARY NOTES a

It. KEY WORDS (Continue an revere olde It neoeesfhy end Identift by block nimber)

User-Friendly, Human computer inteface, man-machine communication,computer command language, man-machine interface design

20. AIRSTRACT (Continue an Pow.,ee old& It neecemy end IdmnLtil? bW week nMeUNA

the user interface is a crucial, but often overlooked, part of thecomputer and software package. It is often the last thingdesigned. Short term memory aids and session pacing are two ofthe most important areas in which the machine can assist the user.Inclusion of a screen pointing device brings the computer closerto communicating on human terms. Basing the interface on a corn-mon metaphor, e.g. a desk top, can make it easy to learn and use.

(Cont*ne'DO I F*N71473 EcITIOM OF I Nov S5 is OSSOLCTS

S/N 0O102 LF.014- 6601 1 SCR?@ASIAIWFMSAE~DO-.i

tatua§T CL&UWSCAYION or Tool aV.s 11fl(~ e 00 SefteeE

ABSTRACT (Continued)

4n this thesis three facets of the interface are examined: Com-munication from the machine, to the machine, and the dialog be-tween user and machine. The Amiable interface designed for theUNIX operating system is described. Amiable is implemented on aSUN model 150 Workstation in the Naval Postgraduate SchoolComputer Science Laboratory. Interface design is a hard problem,much remains to be done.

Do gri 1473tlS14 sm'1d144401 2 scumi- tm, wrl oriSsA?6 Toile PAs&mbmDftf"

Ai,.roved fcr Puicli" Release, Distribution Unlimited

The Design Analysis anaImplementation of a User-Friendly

Interface fcr the UNfIX OeratinrL Systerr

by

Frederick larl Grcenert, Jr.Lieutenant Commander, United States NavyP.S.S.E., University of Colorado, I.72

Submitted In Fartial fulfillment of tre

requirements for the degree of

rISTER OF SCIENCE IN COMPUTER SCIiNCE

from the

NAVAL POSTGRADUATE SCHOOL

June 1984

.lrove : ..... .. h s Advisor

- nd Reader

Cha irman, .eartment _ oCf o; ter Science

Dean 0? Information an ci ences, o For.

GRA&I'LtAB E3

. d.,unced 0

I, icatio

3 Iv-itlbilitY Cod*53 -Avaaid/gr

'1t SpeciaL

4 . . ". =. ,_ .. , r , -. -. . .. - . • .*% % -. o .- -.... .* ",, .• . *-. .-. .- , .. "• " " " ,,v v.- L ,_" " v ," ,• ,- - .'. J* ,,- - . , - - " . "',- '' '' '-' ' ' ' " , ." ' .,,,.- ' ; , "" ." - ."

ADSTE .A"CT

The user interface is a crucial' ,ut iften overloo~ed,

1 art of the computer and software Iacc'4aje. It is ofte the

last thing designed. Short term memory aids and session

acing are two of the most Important areas in which the

machine can assist the user. Inclusion of a screen Iointin..

device brings the com_uter closer to comm.nicatin6 on uman

terms. Pasing the interface on a common Tetaphor, e.g. a

desk top, can Take it easy to learn and use

In this thesis three facets of tfte interface are

examined: Communication frcm the machine, to the machine,

and the dialog between user and machine. The Amiatle

interface designed for the UNIX operating system is

descrited. Amiable is implemented cn a SUN model 15"-

Workstation in the Naval Postgraduate School Comiuter

Science Laboratory. Interface design is a hard 1rotlem,

much remains to te done.

e..

4 a:i...

-. ~ . ~ .,jq'~ .*.~ .~, .*. .,..Ca .q ., -",

~ ~ .~ * ~ ~ . * . * . * -

TAFE OF CONTENTSI. I!TEC"J'CTICN ..........................................

A. THE INVISliLE PIPE ..............

2.DiSIJN FOR !i'AICCP.9 INTERACTION ... , k,

CI OtMPUT3P TO USER COMMUNICATrION ........... .

*A. THEZ FEHYSICAL VIEDIUZ . ............... *... 13

T. TIE SCR?.!%NDISPLAY FO?MAT ........... 14

C.TEE SCREEN LkYOUT ....... ..................... 17

F. ERRORS ................. . ......... ?r

1. Error Tyies and ~iemedies ................ 25

2. Ideal Error Respcnse .................... 2?

3. A Specious Notion? ..................... a

* III. AN-TO-' ACHINE COMMUNICATION .................... 3

A. INPUT DEVICES ................ *. .. 32

E. COMMAND LANGUAGS .. .. .............. ...... 36

2. Function Keys and Frogrammatle Keys .. 380

M. enus . ................. .. ....................... . Z9

4. Vacro Instructions ...................... 4z

I5. TE Ions ....... .. *....................*.. 43

6. Sihe Staemorth A t .................... 45

at * ith Syntax ..................... 47

5 ~ *~*. d .**- -.* *. * . . . .. . -*..* ~ **- . . .

t With Tx~lanatlcns ................... 47

2. The Pesult Sc Far ... ........... . ::

3. Suppressing Detail . .............

B. A QUANTUM hOP ............................... 51

1 .e 2si Mcuse . .... -t**

2. iindows ................................ 54

C. OALS FOR iNtERACTIVE DIALCS ............... 55

1. rinimize Ap rehensicn ................... 55

2. Don't Panic ....... ........ 56

3. Natural Dialog .......................... 57

4. Solicit User Comments ...................

5. Keyt~card? .. . . . . . . . . . . . . . . 57 "

6. First Session ........................... 58

V. AMIAPLE UNIX ............... ... *........ 50

A. WELCOME TO SUN UNIX .... ..... .6.5 ......... 51-1. Main r1enu Notes ......................... 61

2. rain Menu Selectionis .............. 61

3. Demonstrations .............. ........ 52

4. C Shell ....... 65o oma l ..o....•• e o oo . , ... .. q0•oo e•o•o••o o 52•

5. Ama.le ...................... 52

= INSIDE AMIAB LE . . *......... 52 "

C. TSTING ........... ....... . . 55

Vl CONCLUSIONS . 65

A. PEOPLE ARF ST.ANGE ......... -- .............. 65

. DESIGN MAINTAINABILITY IN ................... 67

C. DIFFE39NT DRUwMF35 ......................... 67

S-

I. '

~ .. *-.-.*... % * ~ ... **** *** *.* *.*..-. -- * ... '..* *.*.**.*-. ,

E' ..... ,.A .i II . . . . . . . . . . . . . . 6

PROGRAmiS .. . . . . . . . . . . . ...*. . . . .

'-IST OF MEERENCES . ............... . ....... 5

INITIAL DISTRIBUTION LIST ......................... 15;

I

7

41.

I Ir

A. TzUT INVISIBLE PIPE

The user to comjuter interface is tae invisitle .I e

- through which the mar. and machine con dut a lialog.

Ccmmunication flows between the ran and machine in the f ir..-

cf electroniz signals, visial stimulation and ossioly

auditle tones or lpatterns. The Jct of the interlace is tc

translate the raw data forms Into forms intelligetle by the

iarties on either end. The greatest sinple benefit cf a

modern user interface is the ease with which someone .ray

gainfully erplcy the system from his earlest session. The*

corrijuter comtines the role of instructor arl tool; the best

interfaces do it transjiarently. With aplropriate aids the

- achine can anticipate the novice's misunderstanding, and

offer explanations and guidance; yet accommodate the

exlert's impatience ty streamlinin6 his request. The

atility to do these things is not an innate feature of

exotic state-of-the-art hardeare, though hardware makes tnem

,ossibie. The willingness of the machine to lo the ,an 's

bidding is resident in the design of the man-macnine

interface and the software that makes It a realit.

Unfortunately the human aspects of users, ever within a

narrowly restrio:ted groui, are diverse sna varieale. What

suits cne individual is anathema to the next. Preferences

ar-tr?.

.

S- ..d~-.

change lertcdi:ally, end intellects grow and mature. The

*interface desiebner rutst accommicdett the widest ren, e of

users in a manner aipeallng to all and condescendin4 to

none, a formidable task! The tower of Todern zomuters is

teginnln to allow the designers to bring taie macnine uT to.man's cognitive level, ilie never before. Techniques and

equipment ere being combined that offer vastly Improve!

means :of working, learning and playing. The laser viaeo

disk player (LVD) is a prime example. It is the answer to

an Interface designer's prayer. The LVD not only offers

enormous storage, tut can be interactively ccmputer

controlled to show selected combinations of instruction,

explanation, and examples that would otterwise require a

dedicated instructor or extensive reference material. It is

the perfect redium to store the myriad error messages and - -

ex;lanations that previously were too numerous and

cumbersome to attempt. With well designed lesson material,

instruction at the workstation simply becomes e matter of

selecting the disk, inserting it in the player, then hanging

on for the ride. An LVD combined with a -cmluter wcrk

station becomes a combination library/classrom/office with

nearly unbounded potential.

The goal of the project with whicb this thesis is

associated is to design a Command and Control wcrkstation

with off-the-shelf commercial hardware. This thesis

explores the briarpatch of the human-computer interface

lursuant to that en.

Z-G-F-~~. .7° ". I. -77

-. tEiIN FCi .UMAN-COMPUTI. INTERACTION

It is not enough that jeclle want to use a comp, uter,

their desire must be sustained by tne human-com;uter

Interface tefore the machine will gair. wide acceptance. The

successful interface m.akes the user feel comfortable and

relaxed while doing tis job, and is faster and ,nore accurate '.

than ccnventional methods. A typical user will not chaue.

- h is or her normal work habits to use a new com-uter unless

it is easy to learn and sim.ple to use. Given a choice,

-users .ust iherceive vast improvement over old methods before

they will seriously consider the change. A little

improvement is not worth the effort to chante extablished

methods.

The new generation of computers from APPLE, i.e. LISA

and MACINTOSH, are exiting, relatively inexpensive examples

of just wnat Is possible in a modern interface (Ref. 1].

The dramatically new interface using icons end object .- '

oriented dialog appears to be attracting previously --

disinterested ,eople to com,.uters. Acceptance among

traditional computer users has been slower, largely because

the new Interface is un-:onventional, and they arE resistant

tc change.

Some display interfaces are easier to learn than otners.

Interface software complexity increases rapidly with user-

friendliness. An interface that appears s ir Ie ma s-ks

Internal detail and com;lexity. The software rinning behind

it is nandline, the minutiae formerly left to tne -ser. lut

-"-"

*..-... *'. % ........................... "... .

the ease of use of some of tne tetter systems dill ay for

itself many times cver in reduced training costs and

ircreased operational ef licienr . I t :ai te ;ossible to

save a few thousand dcllars Ter systen in the i-rocurerrent

- ;rocess onli to s~end ten timies that axo'int in recurring

training :csts. "

In this tnesis the three tasic parts of human-ccnputer

interaction: aser-to-computer communication; comluater-to-

user communication; and the dialo, tetween the tdo are

examined in some detail. An interface designed to heli.

* students new to UNIX is ;ropcsed. It Is an overlay that

works in conjunction with UNIX, not in 1place of it.

- Conclusions re6ardine; the human-computer interface and

* future Improvements are discussed last.

I''

I°,

11 2

11. C~IUI TO Ujt ? COM"UNIZITION

The computer communicates to the aser via Cut'.ut

dpvi' es. A tyij ica1l computer work station has one or mcre

CRT (cathode ray tute, screens anid access to a printer.

*Other oat pu t d evi e s are possible, includine; voice

synthesizers and ctter specialized hardware. A p.rinter may

be ;.art of the Installation, but printers are usually

centrally iccated to be shared with others. Zefore CRTs

tecare the accepted interface medium, printers were. Now

they are u~sed as copy machines: to cttain colies of

intermediate and final results.

Auditle tells or tones are used to sound fault alarMs,

or to darn of illegal operations. Frequen~tly, the auzdio tone

is felt tc be an aninoyance by the user comrmunity.

Voice inlut is an emerginS~ tecanology, with high rrices

and low performance, tut the cost to performance ratio is

ir'l rc v Ing a s research progresses. The Li f f Icul ty of

extracting m'~eaning iro, spoken oords and of joinin6 sounids

to forr understandable words takes sibtatntial macnine

resources and time. That additional processing slows the

dialog between user and machine below a comfortatle

threshold. Com;uter vocabularies are small, ranging from a

few words to two to three hundred words.

Voice recoenition is used as a security caeck. Speect

synthesis is also a major research area. Inexpensive sieech

12

IL

syntnesis ivicrc-hij.s nave allowed tne caiatility to be used

In several electroni: toys. ihe "voi-:e" is still too

TecheniCal and artificial soundin,. to te used in a eenerel

* lurlose interface. There is a tremendous future for voice

ir.nput inc cutput when it becorres cost effe:tIve.

A. THE PEYSICAL MEDIUM

The CRT S7reen is the con.unicatior interface. The

ir.fcraticn displayed there must te timely, accurate,

Teaningful, and concise. It is here that the comjuter mrast

condense as Ymuch Information as possible to keep the user

Informed about his current status, the status of his active

jobs, ana any other ]pertinent system information and

messages.

Opinions about various monocarome screen phos,hors are

irany and varied. Personal ireference is the determinin&

factor for which Is best. All monocarome combinations

(ember te~t, lark background; ereen text, dark tackgrcund

wrlte text, dark background; and reverse video ccmbinations)

are aleuate, thoueh the iLr.ed backgroini of full screen

reverse videc ray be less comfortable to :se on text

oriented raster scan displays. The most annoiing distractors

are reflected glare and misadjusted brightness and contrast.

All tnese can usually be controlled by tte informed user.

Means shoula be Incorporated in tne terminal initialization

sequence so the user can set his f.refererces.

13%, •..

|:..-.*'.: ::.: -

7~ 17-7-

Color monitor screens iresent additional prct lems.

There are definite col.,r com.binations that produce

remarkably uncomfortable visual sensations (green text on a

brr ibh t c ra nge tackeround is c-ne). Desi.,riers must te made

aware of the vision function of accommodation ana take ;ains

not tc inflict awkward comtinaticns on utsuspectinf, users.

Ideally, the user sh.ould be able to select tis own preferred

color ccmbinaticns, then the desirgners problem is just that

of rovidin, t".e flexibility to do so. tefault color

' cortinations snould be comfortable to view.

B. THU SCEEN DISPLAY FORMAT

There are two screen display formats used in t;eneral

purpose interactive workstations, vector drawing and raster:

scan. Two versions of each are widely used.

A refreshed vector drawing display has high resolution

• but has teen restricted to use primarily in graphics

applicaticns, where its special features are required. It

is considerably more ex;ensive than the same size raster

display. The vector drawing display's forte is line

drawin.gs. It draws true circles and diagonal lines, because

each line is a continuous stroke of the electron beam. The

number of displayed segments is limited by the beam drawing

speed, screen refresh rate, and amount of storage dedicated

to vectors. The maximum numter ot displayatle vectors is%*-*,-

fixed, whicb Is the primary shortcoming of refreshed vector

drawino displays. Vectcr drawn display text is very sharp;

%°.

P 'p ".-:'..i

'.. -77 7:: :.

but te,7ause ea-h letter is a zombination of strokes the -

numter of characters is limited. ,cr.sequently, text

inclided on a screPn frame raidl de~letes the numcer c-

vectors available to draw fi;ures. Eeiause Cf this caly

special a rlose workstatioras have vector Irawing CRT

screens. 'AD ,ccmputer aided design) wcrkstations tylically

use a vectcr display fr drafting and drawing functions.

The second version of vectcr drawn display, the LVST

(direct view storeae tube), eliminates tne screen refresh.

".ence it is ruch less expensive. DVZTs are not satisfactory

general purpose workstation isplay surfaces because partial

screen erasure is not possible, and writing speeds tend to :'-

te sl:'). DVSTs are excellent adjunct graphic displays for

line drawine.s and text. Improvemerts and cost reductions in

raster display technology have relegated DVSTs to a support

role entirely.

The first kind of raster scan display is text oriented.

The screen is nominally divided into 4 lines of 8Z 0-_

characters eacn. Screen area not contained in that Iattern

is not addressable. Sach cf the character pcsitiors Is

addressable and may ccntain any princtable character. Tris

s:reen Is best used to disilay text. The main advantage of

this display is cost; it is the least expensive to produce.

A s-.reen tuffer need only be 2K bytes. It is possible to

have two or three screen buffers to allow ra Id screen

filling o1erations, even at low data transfer rates. Lod

screen resolution, which results in poor quality 6raphics,

'5,

-S....5 -

*'',- .%* >,* .- :.j. ** . ..5~'**-,* -S "-. . . :...., . .. *

is this disjlay's r'air. disadvantat;e. Ccntrived tar and line

-ra hini are possible, as is Trule curve ilotting.

The second kind ot raster screen dlsplad is ic"'cwn as

.it-maj;ei" or "lot ad-iressable. Or this display the

smallest addressatle screen position is Kncwn as a jix 1

(FICture Le~ent). The ihysical size of a pixel is

determined ty the ;hysical ccnstruction of the CRT and the

signal handlint skeed of the amllifiers that control the

electrcn teem. Very hi6 h quality, but exiensive, tit-maked

color displays of j; to 1024 x 1424 kixels are Tade.

Monochrome dsilays can te had up to 2049 x 2448 lixels.

Though more exienstve thac text oriented rester scan

dis~lays, the bit-ma;&ed display is the display of choice

fcr a corluter terminal or worK station. In tne erajhIcs

mcde It does suffer frcm "eliasin", the stair case

a;pearance of diaeonal lines and edges of circles. But

scitware, and now hardware, sncothinj. techniques can ,.

mrnimize that effect for all but tne maost critical

a ilicatiocs. The bit-mapied display does require a

sutstentialli lar1,er screen cr frame tulfer 'a Megabyte In

the zase of the 1024 x 1024 screen). But ra-idly falline;

memory ;rices have helped to mitie.ate that disadvantat;e.

The raster scan format offers the most flexibility with

the fewest restrictions in a general purpose workstaticn.

The preferred rester display is tit-maji;el. The workstation

that uses 1.raphic communiceticn techniques snculd te

equliped witn a high resoluticn blt-mapped raster scan

15

'2 . C

o N;@% .. % % % % % % ) ,L ,( %"

displey. It Iprovides excellent character definition. 630d

qualitj drawing 1'atures, and eer' pary text legitility.

C. TE SC3.EN LAYOUT

Sost screens ere arranged like a ;aier pee. The

tegInnint; of the I-age is at tne ulper left corner. The

screen fills from left to rift, to. to totton. Only one

;ate, or iortion of a page is displayed. Tnat standarl

layout has served well and will continue to do so. One

display line is reserved for macbine messages. Thus if tte

screen displays 24 lines, 23 may be used to disylay text.

The protlem with the single sheet of paper Iper screen is

that it does not offer a large enough "Iesktop", or woricing

slace. Few people work with just a sinele sheet cf paper.

before ther. Even the most rapid paging or scrolling screen

is not adequate for reviewinp; raterial not Included cn the

current screen. The root of the 1roblem rests in the frailty

of haman short term memorl. That shortcoming will arise

frequently in the design considerations of a naman computer

interface. When the context of Interest is lost tecause of

che.-,gIn. the screen contentts, all tut a fp'# tasic facts

atcut the previous screen are gone too. The solution to tnis

prctlem lies in preservin, some or all of the previous

context when the user must change screens to see otter

Inforration.

One answer has teen to partition the CRT screen into

several non-overlalilng zones. Tach zone is reservel for a

17"

z.

7,P

rrtic-ilar tipe of information. One rniaht disilay the mcde

of operation, one 7ould be a reserved area for error

Tessages, another the title of the current file, et:. Fig.

xxx Illustrates a tylical screen layout usin: zones. T;e

zones remain or. the screen, when the disllay chanees, cr the

user changes what e is dcia,;.

SilIt-s-reen" is another version of screen zones. In

this wcade the CRT screen is divided into two cr Tore

distinzt areas. EaMb area may contain somethin6 different.

Two or mere files may be displayed simultaneously. Two jarts

of the same file may te shown Juxtaposed, to save the user

from scrclline tack and forth ,and forgetting between,. The

jrcble with s~lit screen operations is that the amount

shown cn each jortion cf tn:e screen is reduced, cften making

it difficult to retain the context.

"Windodin" displays are currently fashicnatle. They

are a trick{ way of getting Trcre surface area from a fixed

cize s-reen. They are similar to the jrevicusly descrited

zones; but are temporary, relocatable, anI may overlal. T--ey

el]4ear onl, when requested tb the user. therety saving

screen space. wuindois cverlay ana ;artialll otscure tne

krevious screen contents; but ray be movel by the user tc

uncover an area. The currently active window is displaed

an to; ct' what a]p,eers to be a stack cf pa,!es an the screen.

On most 1indcwing systems Individual window size and shape

may be set t the user. Windcws fllcw additional

information to be availatle tc view at tne user's request,

18

:-:-!:

while reservlr.d the screen zontents. The Techanism of the

request can te any of several pcLulAr methods.

. THE INTELLECTUAL LEV.EL

The intellectual level of rrachine-to-user comnu.aiaticn

has historically been biased toward tne racnine. Innerent

machine liritations and meager hardware rescurces dictated a

low threshold. Earlj interactive systems resionded with an

"UnlIirking Stare" to almost all user inpit. The only clue

that his jo or command had teen received and understood was

another ready 1promy~t -a symccl ai, earel on the scr~ee.

notifying the user that adaiticnal Input could be accepted.

Mistakes In the input often meant no response of any kind.

Surprisinely, many polular systems still offer oaly' tiat-

enigmatic stare, UNIl for one.

The intellectual level at the interface has risen

ralidly with increasing microco.'puter poier and decreasing-

hardware costs. iast, caiacious storage and modern

techniques have solved the iormer hardware deficiencies.

Hardware Is no longer the major limitation. The emiphasis f'or

iTI.rcved yerfcrmance was not directed to imlroviaK the

interface until very recently. Whole CpUs (Central

Processing Units) are dedicated to screen management and'

input-output; but the machine-to-user comirtunication re.rains

the same, with the exception of .rcre accurate error and

fault condition relorting.

1..

, ~~19..'-'

teeny of the fa-alts wit. the interface can ce traced to

practices and prc-edures that began ia the early days.

Surprisingly, .any tatch .rccesslng 1roced res are :ommon in

interactive computina. For exarple: every cornuter

* rograrrIrg languaee :,cular todai was desitr.ed for a ba tch

1.rocessi n. enviroarent, being separately corrliled, loaded

and run.

The move aoay from the tlank stare mode began soon after

the advent of interactive comtuting. It did not receive

much attention until the tricrcccmputer explosion caused the

computer to move from the laboratory into offices.

Comi.'jters texcan to be called JijOn to do useful work in

real-time. The clamor for user friendliness began. Many

ideas have been tried and discarded. Those that have dorked

have apIealed to the user's 1hysiological, sychological,

cognitive and intellectual needs.

The se.rantic level at the interface nas moved from the

bits to bytes (comiyuter words, languaee Characters), to

icm.rand words, to clumped ccmrtmands, and finall- to a

m eta horic level. (Ref. ZI The metaphor is a model that

selectively combines all tte useful and practical features

of computer interface aesi-n and integrates then, into a

cohesive ;ackage that emulates a familiar environment. All

availatle cox.juter function s are masiced behind a mcdel of

the a ;licatio. it is intended to serve. The user sees only

terms, functions and methcds with which he is familiar. The

bewildering and intimidating world of the :ona.ter sciertist

,-'. -'.,...- ... , .,,-..' ~~~~~~~.... , -........ .. ,......-... ....-.-..--....-..- . ... ,........... ... .......... .... . . .. .. .. I _1 _ I .. . .. . I -l -- I _. _I __. _i _ i __i _i __ i 5 ,...._ _ t . * Ll . ** 5 ...... . ". l* l i '

7--7 .---. ,:.j-W-

is ccmietelj nidder. The ideas that grew into the

reta;hcric interface originated with ttre pioneerlng work of

Alan Kay and his 6rouV, at Y]ROX PARC with SMALLTALK ardA

Dynatock. Several commercial Iroducts are ising taese

conceipts, among them Xerox STAR, Dclihin, ani Zorado; an.

APPLE's LISA and MACINTOSE [Ref. 1][Ref. 2J. Tals interface

more oearli erulates a convertional desktol.. Everything on

the screen is treated as a 1hysical entity, baving an

identity and o:cuiyine s-ace. The keyboard is used cnly fcr

text entry. A pointin, device rouse) is used to manipulate

the otjects. The translation is so complete and the

implementation so natural that comparativell little effort

is required to use the system. This object based dis;lay

fcrmat is a paradigm ot ccmputer-to-user communication that

will beco e a starting point for all new machines. The

*etaphor can te modified to treet the needs of many

situations and applications, changing the nature of the

-:omiuter from, special purpose machine to general purpose

tocl. The keytoard is the final hurdle tc widespread

=Crfuter acceptance. Its elimination as tne irimari means

cf enterine text and data will nark the bc,6innine of an era

that will see comluters teco.me as ubiquitous as televisions

ard telephones.

F. THE COt'PUTER SPEAKS

Not everl computer interface designed now will emulate

the metaphcric paradif-m. Designers must endeavor to Keep

21

.. . . . . . . . . . . . . . . . . . . . . . .-.~ ~ ......-.

" -"- -- . r.L"-"-- .' :.. ." ...- ".:'-...?.""-.- "'.'."..-.'. "- '-.'-.'..-.'- . . '.!'.' .'.' ., Z..- '-.,..-. '.., -. '.". ',.- .. ," "

-, .- - - - -. °I. .

the level of corputer-tc-user :o~rvmuni.,aticn as close to th.e

.uman domain as possible. 7very expert agrees that the

rachine must respond as quickly as cssitle tc an injut

request. If it is unacle to resicr.1 In less than one or twc

seconds, then it should notify the user that there will te a

short delay, and give an alroximate time to com ,etior.

Lef. "1. If the prccessing will take a og time and

iartial results Can be displayed, then the machine should

tein responding as soon as it can with intermediate outlut.

The time limit of one to two sezords is a rule of thumb, out

fits well with human ccenitive processes to maintain tne

user's train of thought (Ref. 4]. Every utterance made by

the !"achine must te self-explanatory. Eere utterance means

any ccrruter-to-user communication. Cues irompting for

Input Should, at the user's discretion, display the tcry

(syntax) and oltions of the exlectel command. That nelps

the user twc ways. First, it helps him rememter the command,

reinforcing his knowledge about it. Second, it reduces the

prctatility cf mistakes. Ideall1 the computer will display

the comrrand line, and the user can fill just the blan.,s,

seving extra tyl-ing. Ks the user tecomes more confident and

experienced he can shift to a more independent mcde.

A machine initiated mode of ccntrol relies on the form

and sutstance of machine communication. Tne mactine usually

controls tte ilalog ojitlons in this mode; out it need rotappear to be in control. The user is free to select from, a

list of otions made avwilacle by the machine. Th.e

22

.- ,. _:'&,L.-_ .. Lq.,Z'..'_. .e.". -" ," .. * -. '.-.-.- ... v.*., .' .- "-" . ".'- -. *." *... -. " "-.."''"* "".* ''* .

restrictic.ns placed on user action ty tnis mode reauce the

cjlortunity for error, both in r.umber and tyre. Novice,

inexperienced, rushed, and careless users all tenefit frc.r

this feature. It also has aisavanta~es. It is Usuilj-.

slocwer, tecause the user :uTst scan lists of, possitly

unfamiliar, ctions end make selections. The selection

mecl.anism adds to the overnead of the c!hosen cleraticn.

This mode is more cumbersome to ase, because the ccntrol

l, ath is predetermined, and shortcuts are not easy, iernaps

rot jossible. It will affect "iser flexitility, tecause the

ccmmands and, options offered will tend to be generai

.'ariose. This rode is preferrel for sasual or inexperienced

users rimerily interested In accemplishing the task at

hand. It can also be useful in system tutorials for new

,jsers.

Machines that do not nave a state of the art user

interface are not noleless; but they do need nelp. A

scrtware interface package can be overlayed upcn the

operating syste~m. It need not adversely affect system

1jerforrance because it will be active only when the system-

is between modes. It will taxe slace and slow parts ot tae

ijresent system. The idea is to display renus, soft tuttcns

and pull -down windo s on the screen. The dis.lal contents

are cues to aser choices. For instance, a ccmriuter user

ex;ects to be able to locK througn nis files; drite, edit,

com .ile and run jrorams; compjose and edit dc crents; cc;y

1rogra.s and documents (electronically ana Irinted); and get

I, :::'-

on and off the system. Rear~less of shetber the system

ccmmands to dc these thiiys are easy tc learn and rem.e, ter,

the new user is unlikely to know the zore-zt z Omman a ard

therefore is unable to use the system. Yet ,e knows what ne

wants to do and now to do it imanually. He isc missing t.e

key tc unlock the power of the racnine. The claim is: He

rhould not need specific knowledge atout tne machine, or the

olerating system to use it. The user is normally not

interected in processing details, only in getting nis job

done. He should not be burlened with ccmi'uter-spezific

details. The machine is supposed tc make his jot easier. In

fact, present computers add a layer of complexity to most

jots.

The interface overlay easiest to inilement is menu'

driven. It :an be enhanced to a full fledged meta;horic

interface if the machine is sufficiently fast and has the

resources. A i,:k device (mouse) is preferred; but not

essential. The Iick decreases reliance on the keyboard,

de:reasine user alprehension. All menu selections are

labeled witn generic terms, recognizatle by a troad range of

users. The siecifi. system is hidden by tle finctional mask

of the interface, directly embcdying the first Parnas'

i.rincijle of information hiding [Ref. 5j. The overlay need

net, in fact must not, be static. A static interface would

necessarily te slanted toward a siecific aser erouk. To

increase the user dcmain, flexitility Is essential.

Instructional messates intended fcr a novice caise annoyance

24

!k7..

and irriatienre in experts. The same ilexibility should also

alloo tie expert to call on available ai s, if i.e so

decides. Though the mechine controls the nu.Tter and tyje of'

-ccmands allowed the user is still in ccmrnand.

--- . I. BOhS

Error handling and retorting is the most crucial facet

of man-to-7achine communication. It is easy to inspire user

confidence and calm when everytning is dorking correctly.

The test cf an Interface is how gracefully it behaves when

somethin& gces wronc. The classes cf errors are defined for

the typical computer system. Some more exotic error

treatments will te touched on afterwards. Three classes of

errors are readily definable: sulerficial, recoverable, and-

fatal.

1. Errcr Tykes and Remedies

Surerficial errors are caused by user carelessness

or confusion. Thel in--lude typographical rristakes, mistaken

selections, and out of sequen:e actions. Tney are easily

detectatle ani :orrectable, in most cases. Typographical

errors rmay te :au, ht by either the user cr the macnine.

'**istaken selections are legal operations, but rot the one

the user intended. The onus is on tne user tc ccrrect the

situation, but the system :an help by allowing the user to

quit an cperation or sequence at any Intermediate point.

Out ofP seqience actions o-cur when a iser fails to do a

necessary intermediate step. For instance, ne tries tc

25-

execute a source file withcut first ccpiling it. Tzis

error is ireventatle, t.t the designers must i ncl.ide

explicit features to do so. They rrust try to aaticipate

user a ticns jrvidIr. aj.ropriate renminders and

l, rccedures that illustrate tre ;rcper sequence. If tne user

learns by his mistakes he is less liely to commit the same

errcr twice.

Recoveratle errors are not discovered until after

the fact, but may te undone by some means to get tack to the

startinE condition. This is a much larger class of krotle-s

and requires more ambitious methcds to fix. The metaods

usualij involve restoring wcrds, lines, or small files that

were erroneously changed or deleted. Ine recovery

techniques involve internediate tack-up. Snall entities

like words and snort files can be tem!porarily saved in a

tuffer until the proler result is ccnfirmed. The stora::

cvernead is not unduly expensive, relative to the confilence

the ability to undo the operaticn instills in tte user.

Larger files -an be tacked ul, on disk or tae, but the user

must usually take cvert action to do so. . small class cf

funiticns may be inverted, i.e. run tackwards, to retrieve

the input from the results, but that is an excepticn.

The final class, fatal errors, are ncn-recoverable,

or are so difficult to recover ±rcm that it is nct worth the

exlense. These errors must either be prevented, or the user

warned of the nonsequences before they are executed. For

exar.rle: a user decides to delete tt.e entire contents of a

°.. .-.7: ..

dire.tory, and imjrudently uses the command "delete $.'"

without realizing he is not in the directory he intended tc

delete. The commani is leral and toe sjsten 'will execute it.

The system desiner must have anticipated this situation end

included a !eisure of protection. To warn tne user t-e

system shoill respond with a message similar to:

WARNING, delete . will delete all files in direc-

tory:

<directory name>

Confirm intention by pressing the left mouse cutton.Press the right mouse button to atandcn command.

Where <directory name> is the current directory. This gives

the user an opportunity to ponder his action and reinforce

or abort his decision (ard save himself considerable

anxietj).

Another type of fatal error is caused by hardware

failures. These errors are becoming less common, but nan be

heart-stoppers when they do occur. The only remedy for this

class is a cormpre.h.ensive tack-up program, expert tecnnical

suppcrt and a dose of :ood fortune.

2. Ideal Error Response

The ideal response, in ost cases, is no overt

response to the error condition at all. The mactine knows

what the user intended (based on the context in which it

occurred and tecause the suspect word nearly matches, or is

an alias for, a known word) and carries out -is re4uest,

- ignoring the user's ,isspelling or command synonym. A modern

27

'V.

Interface desien shoiili s~ecif7 a more partitioned resjicnse.

A new user, or cne leerning the system should be presented a

reminder of the prcer sjelling cr comman.. mat coulL be

done ty havi ng te system substitute the correct ccrmand

wcrd at tne aIlrolriate spot, for instance. The InterlisI.

system erielves the user's intention, requestin;

confirmation of slelling corrections it ma.es autmatically

[Ref. 51. In ojieration it seerrs uncanny, as thoueh another

.erscn is watching over the user's snoulder. Interlisp

demonstrates the ;ower availacle within the machine, but the

rescurce costs associated with it make It impractical for

most ficrccomiuter based systems. A more accom,lished user -

might turn off the reminder feature if it annoyed him, or if

it slei up his job. An even better rethod is not to use the

keyboard at all except for entering text in a document or

names into a program. Compiler errors will become

anachronisms, as more syntax directed editors tecome

availatle. The em-hasis is now turning to run-time error

detection and correction, a mich more troutlesome and

difficult problem.

To be heli.fal the machine must provide ecou~h

information about the error type and location so that the

user can attempt to find and correct it. That imrlies the

ccmruter must have kncwledge of the procram, as well as raw

information. Unfortunately, the artificial intelligence

.rcgra.r required to ierceive the user's intention from all

the data on band would seriousli cverload all tt the most

-' "'28

%V 77*.. -7 7 -. 7- 7-7-,. -

iowerful mri-rocom;,uters. The Telahine usuall has sufficient

raw data atcut the errcr to -Prcvide accurate information

atout the problem. The information is us.ally not i. an

understandable form. System desir.ers have been relictact to

provide the facility to dc the let of translaticn, tecause

it nas teen considered to be wasteful of -recio)s resources. .,

For example, many mcdern error handlers respond to an

incorrect address call with: Bus error(core du.led)". inich.

sends the user Lack to his prcgram; or worse, tc a

hexideciral core dump, to find the fault. Most systems are

cepeatle of: Address out of bounds in subroutine

<subroutlne name> statement <statement num ber>, if the

* error handler has had the prc;er eirhasis from the beginning

of the design. An artificial intelligence based debugger

might resionl as in the previous statement with the addition

of: ... expected <address name>". The iltimate reslonse

mi9 ht te a warning to notify the user to verify the

ccrrected address. In the recent ;ast the designer's feeling

has teen: A trinimal error handler is adequate tecause real

"rcgrarmers will figure it out. That is no longer an

acce,table attitude on any design team. Those resources are

ncw expendable, translatin, data from one iorrm to anotner Is

one of the machines strongest attributes.

On line detug,.ing tools nave teen so.me of the most

obtuse irograr development aids ever devised. Single

scurce-statement stepiing tecnniques (frotm compiled code)

are some of the easiest tools to use, but are not widely

29

~ .~ .*.* *~ ..f..;..4' %v.:.~.::. &K.~:.~.* ".77"

availatle That is the normal det ag ircde of interpreted

code.

3. A Spgcious Notion?

The notion that .mistakes can oe redu- el by provilini.

rewards for correct perfcrmance and ti 1tLnishine Incorrect

lerforriance seems to be airclos in learning to ise

-cmkuters. The concept of rewards and 1,-nishment is well

fcunded and widely accepted, tut in this context must te

extremely subtle. Initial atten;its will probably fail. One

prcposal is to reward correct behavior and prcvide correct

usage ,rcmpts when incorrect entries are attempted (Ref. 7].

In thecry It sounds plausitle. The particular imilementation

;rclosed by Dwyer appears to be too obvious. Most huran

operatcrs will chafe under the notion that the machine is

dlisjensin, symbolic pats on the head. Tae Issue is woo is

in control, the user or the machine? A tetter method is to

allow ccrrect performance to te rewaried with obvious

ro, ress. Incorrect performance should te aifficult or

irpossible. Allow the user to set experience level keys or

let the system recoenize the user and set an experience . -

varietle internally. User selectable oktions are test.

Various levels of experience invcke different tuidance for

each user action. The notion cf rewards must te very

c autiously implemented, and only after extensive testing.

"z""72

.. J *. . .-.- - .-.. * .:..:-.:>-::

III. V:AN-TQ- 'ACPJEN COWtNICA lION

mantcTacnine communization has teen th5 4ivid ir ;

factor tetweei cdTputer users and non-u~sers. Past (and

1,resent) corijuter command and irogrammin& lanjuaees re

cryltic, confusint, difficult tc learn and nearly impossitle

tc remember. People are not mrentally equiljLed to learn

things like that. It Is es-ecialli frustratin6 to learn d

iarticular set cf commands ar.4 their Idiosyncrasies only t c

forget them almnost entirely di..ring a fortnight of nor-use.

The reasoni fcr the inhuman lexicon in the past waas lacKc of

szorige siace. There was rnot enough spac-e to do all the.

things that had to te done, so the sacrifices had to be rrede

*by users. The former extravagance of more understandable

* ccimmends has teen made affcrdetle by the microccmputer

revolution and the shrinkin, Ticrochi;. Sufficient memory

end rccesscr s}leed exist In all but the smallest machines.

*Ncte that the conp'and languag~es transcenl tnie inere ty~.ed

code word cr ;hrase. N~ew methods of imp.ressing the iser's

intention on the macnine involve some very hu7-a

abstractions, e.g. 1ointint, to the iLem of interest.

Regardless of the methcd invclved the newer :ccmunication

forms have teen keo from wide use ty the expense of

hardware. That is ro lcneaer the ease.

thn~ lk ta I s s ecalytrstalns o ern -1i

.,IL ! * I 'E E U ..AJJ L -U .I 'E .H .' ". ." -. -'"

II'

&.INPUT DJICES

There are many kinds of rhysi:al Input levices:

keytcard, joystick, ligKt- en, rouse, d 1,itIz in ;,ad, and

tcuch yanel, to namne severl. A more !:omplete ieszri~.ti:n Is

inrl~ded in [Ref. 4] Per. 9. A trief descr llio, dil

suffice here.

The tylewriter stytcard is identified vit: the co.mputer

becajse, even tefore vides~read trteractive computing or.

teletyie maehines, key-punh "ac.tineo, ,Pre isel to create

cards that the machine could read. It Is the classical

-omluter iIput levize. As such it is a first cholce ty

computer system desiners. M-cst computer users expect a

keybcarl as part of the machine. It way have from 40 to over

1 keys. The u;per and lower case alphabet and digits form

- to 9 are ilnCluied, as are m!rst unctuatlion marks. !cst

keytoards are encoded with the ASCII (American Standard Code

for Information Interchange) character sets. The other

standard character set is EP.DIC Jxtpnded iinary Coded

Decimal Interchan6 e Code), used primarily cy ItM'. the -

* er,.cnoTics of keytoard desi,,n was decided tetore the advent

cf t.e coyljuter. It is a result of tjjewriter design, and

is locked into its familiar, bit awkward QOERTY laycut

tecaise of the costs and politics involved in having people

learn a rew keyboard layout. Other designs, mere suitacle

to hujman pnysiology, nave not shown sufficient Improvement

over the star.iard to be considerel for wile apilication.

32

The most likely fate of the keyboard is that QWERTY will te

sulerceded t~r voice input.

The Jcystick is a hollover cevice frcr the era of ar.aio.-

ccmputers. It is a stick mod.ntel witn two degrees of

freedor. Stlon in either of two kerlenlicjlar dlreztions

causes a wiper arm on a potenticeter tc icve, chanrii the

resistance. That resistan-e is convertel into a digital

signal and fed to the ccrputer. Thus a joystick can

reyresent a point in X-Y space easily,and so can be used tc

position the cursor. As a curscr controller the JoysticK is

a member of the locator family of logical devices.

A light-;en is a light sensinp. device used to denote a

joint on the screen. It senses the lighit emitted when the

electron beam strikes the screen phosphor. It is shaped'

1iie a len and has a tnin wire bundle attached to the end

o,1csite the sensor. It shines a light from the sensor end

to show the user precisely where it is aimed. A light jen

is most often used as a pick, thoueh it may also be used as

a strcke, lcoator cr valuator device. The light pen has

fallen cut of favor because it is more expensive and fragile

than a mouse. Because it is sensitive tc all light, not just

tne light generated by the CRT, it must be Keit in, a holder

or stand when not in use. That means the operator can r.ct

just set it dcwn between uses, lijre he can a mouse or

Joystik. Another light-pen Cailing is it may become a

'rissile hazard if' used in a nn-tenign environment, e.g. a

ship, aircraft or automobile. The lieht-len was once a

-kv-

-. * -* .,*** .- -. i

I....

I-reakthroigh in computer control. It Is telnb re-laced now

ty the mouse, a cheaper Tore rotist devize.

The mouse Is the new darlirg nf the comkitsr interface

designers. It offers a flexible and relatively inexjensive,

means to control a screen cursor. The cursor Is thien used

for most Interface functions. Toe rouse Is mcre exensive

than cirsor control via a keytoard; but it enables a control

mcde that Is much mcre clcsely attuned to human cognitive

1.rocessin 6 (Ref. 4]. Conseyuently, it is .mucb easier to

use. It also suffers the missile hazard iroblem of t.e

liht-;en. But a user iray leave the mouse where he last

used it, withcut fear Cf generating. spurious iniut, as the

light-;en .Tight. The number of buttors to place on the

mcuse is ar,.uatle. Three buttons seem to be widely

accelted, but one and two are also used. T*e deision s.oall

te made on the tasis of the expected user group. The

arg,u. ents range from One button may be less ccnfusing tc

"The scftwarp may te easier to write fcr three tuttons."

(Ref. 1]

A di6itizing jad is used to translate grakhical

information into a machine-readable format. It is a flat

surface with either a mouse or stylus connectel. rrawings or

other graphics information are placed on tne surface, then

j oints are read fror the drawing using the rouse or stylus,

dIi tally encoded and transmitted to the computer. k

dititizire pad can double as a mouse type cursor controller.

"...-

To'ic erels are a recent addition to the world of

ccmIuter input devices. Two ditferent technologies are

I ised. One uses transiarent pressure sensitive men-:ra.es

adhering to the surface of a dis-lay screen. The otner

version is raee of two linear arrays of laser diodes,

arranged at right angles, ttat form an invisible grid of

intersectir light teams on the lisplay screen. A finger,

or any other cpaque object that interrupts one or more beams

is interpretel as an input signal.

The major disadvantaee of touch screen panels is

resolution. To keel costs within reason, and to offer

acceptable resolution (approximately three or four lines per

inch) the touch .anel is suitable for use as a Tick device.

Higher resolution is possible, but costs escalate rapidly.

If resolution is finer than a human finger tip, a stylus

must be used to indicate user selections, to avoid ambiguity

and spurious inputs.

The ultimate in cursor control is a "nc-hands"

controller. Imp]lementation is by means of a hat containing a

senscr that letermines where the 'iser is 1ookine. Most 2,-e

a low .cwer laser reflected off the eleball to read eye

motion, which is translated into a screen position. This

mode cf operation allows the user to keep his hands on tfte

keyboard, a benefit for accomplished typists. No-hands

rursor controllers are just now tecorine; commercially

eve ila tle.

-- 75

'° _ . "° . .. . . - • e • - . . . . . .

• - • .. e .. . % . , % -. % % % %°.." %- • - • . ° .- .. .- . . . . ••

.• - -. •. ° .- .- , .. .... .. . o

*.COAND LANC'oAG.

A command language Is the xears-ti which tie :omluter is

.ade to do useful wcrk. P tyji al comrdater co-ranr lan .ua,"__

% is designed to imilement functicns. Those functicns may he

simlle or co.rplex, low level or high. Langrage desir, tased

on :ra . atical, syntactic and senantic ;roperties cf nu-an

lanr.uae tends to generate -cpater command laneuages that

icok like sicken language. In the piast hardware limitaticn5s

drove command lan'uate design decisions. Now, new hardware

ca;atilities are allowing a tctally different language rroael

in man's quest for more nataral computer use. fhe comman.

lang-uage atstraction is moving raeidly toward the human end

of the man-machine ccmmunication channel. The trend is away

from coirraod words and ljhrases, and toward a command

re1resentation that emulates the environment. The things

the machine can do are teing expressed in common,

understandatle terms. In some, words are not used at all,

3ictures and metion, based on the well known retaphcr of a

desk tol., suffice to define an unairbiruous action tae user

wants the comjputer to execute.

Cormand languages nay take rany fcrms and be irplerented

ceveral different ways. Machine control metr.as below the

process logical level will not le discussed iere, because

those details are not useful to the general user. In tnis

context a process may be thcugn of' as the lo,:ical emtcdyr.ent

* of a iser command. Txical processes are: logging onto the

machine, editiln, a file, editing ccmmands within a file,

95'

trcwsin4 a file, executing a rcgrarr, etc. Tne mcst ccnrrcn

command languae Imprlementations are enumeratee celow.

The sim~plest ccnrrand languages us,? sin.-le char-acter

^cdes to represent legal operations. Tne test attempt tc

apply rnemor.ic tae's; tut tnat is difficult with sinfle

characters, e.g. E could mrean edit or erase - two entirely-

different operations. Single :haracter codes are no lcnger

used as computer co.Tmand langua.es; bit are still widely

used in interactive editors, detuggers, and various utility

*rogrars. Single character codes minimize typin,.

The next most simple comTand structure uses multiple W.

character command names, usually between two and five.

Mnemonics can tegin to hel users differentiate commands"

tut they still need handy reference material.-

Full word command languages are the easiest to

remember. The functions are invoked simply by entering tne

raire juast as tefore; but now t"e function nee is fully

specified. Spelling and attreviation pe-uliarities are no

longer a conc.ern in remembering a function ram.e. this mcie

can te enhanced to save typing;. The machine car. fill tne

trailing part of the command word or phrase as soon as it

recognizes an unambiguous leadin,, character string.-

The factors that separate a good command language in

this ecategory, from a tad one rest in the syntax and

semantics of each. Syntax rules like the "zero, ore,

infinity principle" [Ref. 51 , and command structure

3?7

................................ **%*'.~-**.*\*******.**.*.***.*%**.**.~**.'**°***° .°

ccrsistency can Ie ircrtant alas to learning and use.

I~nor.n siYr.ple ruls liK that an result i t an

atomination. Learnin tc us. si:,j1le ccmmand lar.,uae;es is

almost entirely a remory fun:tio., the iser must associate

the tnlin, ne wants to do witn the correct function neme an

.yntax. The early stages of the crocess are very irustrating

• - because the ,user lust -onstantl; refer tc docu~entation for

help. Even tne most re,;ular languaeges still require a

learning period.

2. Fun~ctio Kei-s ard PrerT~gj Igy§-- - - -- - -~ - -- -- ---

Function keys are the nardware designers answer to

the difficilties of learnini command languages. They are

usually additional keys on the keyboard, tnat may or may not

be labeled. Function keys relieve some of the short term

memcry irotlem but they add to the user's initial confusion

and irtimidat ion. Some the keyboards in systems that ise ~o" %,

function keys are awesore. After learnin,, the extra button

meanines and locations thej are tenerally easj to use. They

also tend tc limit system flexitility, because the user nas

no control over the functions. The inflexitility is felt in

other ieys tcc. Users who initially use tne system tecause

the function keys are self exrlanatory often find tteirselves

disenchanted when they have outerown the nzvize-user stage.

Prograrmable iuncticn keys are a partial solution to

the problems of function; keys. They allow additIcr.al

flexitility for tctn the system desigr.ers and in most cases

the users. The tiexibility comes tecajse each key is no

.. L - . . ........ - . . -..',.., -.-. .*- --.... .. -..... .- ......- -... -. ,.._ . - .... - ..... ....-.. ,

longrer locked into a sir6e function. Fach ajplicatior.

1 ro ram may redefine the Aeys to-suit its neeas. In many

cases the user can define keys tc suit nis n"eecs. ieer --

function Aels are needed whica reduces tne *sers irnitiai

awe. Urnfcrturately, now the user is face d with a

remcrization prolem again, unless the desipners prcviced

for variable latels too.

Function keys ani -rogrammacle keys are an altem-t

tc cor pact more functionality into a fixed :aiaclty. They

are a :omiromise, they add to esuiment costs, and they make

the syster unique tecause of the added Keys. Scmecne usel

to a general j1urpose keyboard may have difficulty adapting.

That is not a failing cf the hardware, but a user interface

consideration. Systems with fixed function keys ten to be

aimed at novice ccmjuter users. That translates into: Oily

novices will use the system. Too much attention pail to a

particular user se.ment usually means the rest et

ne1ec ted.

3. Merus

Menus are the software designers answer tc tne short

term memorl iroblem. They have aivantages over the function

key solution, but they also suffer scme cf the same

flexibility problems. They can be variatle size and length.

They are more self-explanatcry than keys, tecause the

designer !an write ncre than a single word or attreviation

in an entrj. Like keys they also channel the users activity

into predefined pats that may te difiicult or imcssitle to

3;

- '-. i-

aelart. Desi6ners Tust te careful not to take the menu as a

final solution to r o, rr and ien ,uag4 prctlexris. The-- are a

useful tool and when ;sed in ccijunction .ith otner devizes

can hell create a 'cwerfl environment. They can also te a

nuis ance. !scale seqaences frox. several layers of menis --an

te jarticularly devicus It the user is not Kept infcrmed ct

his whereatouts in tne hierar:!y."4. Macro Instruct ions

Macro instruction is a term coined to Oiefine a

comtinatlon of comlouter commands, used frequently, that Tay

conceal intermediate results. The term was first applied to

assembly langiage programming; but is used at all levels of

computer languare now. Macro instructions are the ice. to

. user flexibility. They are an easy way for the user to

customize a system to tis personal desires. Unfcrtunatel"

they are sore of the hardest thlngs to learn atout a system.

The reascn is somewhat self ierletuating. New users rarely"

write their own acro instructicns, though they may use some

.ive ts thien without realizint; it. They usually are unaware

of the existence of macros. Experienced jsers use xacros

extenslvely. Tney alsc write the documentation about hcg tc

write and use macros. Have you ever tried to read system..

documentation? If you nave, r.ow you Know why onl.

exleriencled users use macros.;,, ~5. I cns.'.<.

Icons transcend written language. In the context of

:omiuters they are small ancomilicated sketches that

* * * *4 ". .-.'.0

represent ccrr.cnly .ncon ideas. Understanaable or.es are

sjmtols that represent a iarti:iar retaj-or. Iccns are all

around us in life. They are esb*erCiall- helful wren tryn 4,

to cross a langiage carrier, as is deronstretel by

international road signs. An unfarniliar user nay nc t e

atle to dezipher an iccn at first; cut onze he knows its

definition he will rerenter it -*aen ne sees it ag~ain. Trey

are tne easiest interfa:e device to use; but 1lace heavy

demands on hardware resources. Icons require tit-apied

screen and fast secondary storage, as well as .lertiful

1rirar storage.

5. The State of te eArt

As stated, the dep~arture from ccnventional computer

command languages is a result of vastly more jowerful

ma.:hines, es~.ecially microcomluters. Object-orier.ted

interfaces nave captured the Imagination ci tne industry.

They errbod.- much of the novelty and gee ohiz" excitement of

arcade video games, yet are aimed at entirely practical -

kurlcses. "wizzywi& is a term for this family of co.rmar

lent.uages, from tre jihrase: ihat You See Is 4hat You Get

(wYsiwy;). rhe user is not requlred to orir- anythin6 to

the machine but common sense ana a jot he wants to dc iwnich

the mac.hine is calatle of). The languade he uses sitrrlates

mer.uelk doinb the task. For instance, to produce a ne"

docrrert tne user selects tee j.icture of a tjy.ewriter from

the i,:cns displeyed. Upon selecticn 01' tnat icon the

com]pjter responds ty enterin,: tne wori processing mode, at

.4,_.

the t:. c i an ;a .e. The jier -ry befir. tyli-g as soon

es the displayed p.a,.e is in F~e. . The kieytcard is zis

typewriter, the screen cis ;a-er. Character an w~rd delete

keys are usad for tjking: ccrrecticns.

Larjer editing functions, e.,. line sacing, tnt

styles, text insertions and deletions, are mouse contrclled.

The fainction is !elected with the nouse. 7hen, tein ani

end text ;ositicns r.i ng the rante of apllication are

chosen, also using the mouse. As the rroase tutton is

.cic~ed, ccnfirine the range, the function is executed.

The results arpear in the text, witnout any further user

action.

The wizzywig editor descrited nere is the state-cf-

the-art in word I rocessing software. It is not inexpensive,

tut is affordatle. The APPLI MACINTCS' is sold with a

version that, while restricted to six inch lines and ten

iages, offers all thie calatility described above. Ncte

tho'ugh, the research and aesi=. effort that develoaed that

edltcr was the nulmination of over a decade of active

Siarsult, in1 the editor is nct transferrable to other

rI:rc:crrfuters outside the A??L! faxily of "f Z bit

.achines". It does dpmonstrate #nat is possitle in a fairly

modest but caiable ricrrc:cInter-based system. Sadly, tris

class Tf editcrs are not generally available tecause tney dc

r.ct exist for Tost rachines, yet.

S., . ...-.

-I V .. r..: 1I. . . I

'User :cntider.ce in a cnp.ter syster, de;e.ds tctall cn

hw ne views the system. T.at confi,en:e builds in stages

-ein.inr, with tne tser's first exlerience with the machi'e.

If the harlware features are close to his expectations :.is

onfidence rows; harware ui6red-es car. te ha later. If

the s:ftware is available for that ma-:.ine tc do waat ie

wants and neels, he is aeain tolstered; software is

relatively inexpensive. If the user interface is ai.ward

the maohine will eather dust. Some people are willin; to

rake allowances fer everi part cf the .achine except the-

interfa'e, tne 1art they can not toucn or really see.

People, esfecially trose outside compruter science, will ict

wrestle with a rmachine, even is it is suizsed to be better.

If the *ser interface is cricial to mahinp acceptance,

it is even .rore so for accertance of software akklicatin

"acka-es. "he feature of the interface that ,eCIe erelve

is the dielot:, eef ined as EhP ccnversationel .cde cf

coqrnunication tetween the aser and ccmoater. Rebardless cf

all other hardware and s.of tware ccns iderations, if the

!ialog is clu-rsy, unevenly 1aed too ti htly acine "'-

:ontroiled, or otherwise aa;eallne to the Intended users

thee will jerc ive the machine t. te "bad". They .ma .C

even te aware cf why they feel that way.

47

Inte rt a- : e lia c,. 7r .et a cc smrd ate -set r syc ic o".y. '-=

ctatement is sirnk.ie, tut tt.e i !,,licatiZ. ± at staterent

is not. With s*ll 5nd artifi-e the , -r "re --rcs-et

tc e; eel to the hur, * n s I , nd t:.et ii what -c~ern

interfa esi. is at ot ... "A;IC!

.TAiTIG AT THT EEGINt.IIG

TwT distinct dialcs exist at tne ccr.,iualctio:

ir.trface. a:h is tetter sjited to different situations,

thcun neither is totally wrong in any ase. The tel1in

factors lie with both the iser, his computer skills, ana "is

intent; and also with the machine, its caiabilities, and

iurpose. Cre dialog is :or'v.ter initiateJ, the other is

user initiated.

The ,rahine initiated mole, mentioned in chapter II, is

the reec~mi'ant one in a -clern interface. The strcn,

attribites of the machine are cro= ht to the interfaee to

b el -crrie nsate for the user's weakness. By dis]lain,"-.

lists, the cr.uter's a:-urate recall is ised tc nudoe the

user s rercr a'out the e.vironTent, -crands, Ic ire l

lositir., and various other details the user needs, t t s

difficulty keepin, under cc.ntrol. If the encunt c-

assistance renderel is user sele:tatle, the lEsisner has

- 1laced the user in de facto ccntrol of his envircnnent. If

scme of the levels of hel; hae ,er to Ice t'acrial, the system 7.

'e.ins to Prer,:e as rru:h rcre tha. Just a smart t;iewriter.

'a'.

44

• • °.. v

1he se7on.1 ,cle is ust ci;osite, the :ser initiates the

action via a command sequence and tr.e ma-.nine resl.o.ds. Ia

Its jure forr, this mcte relies entirell on w.-it the user

Kr.cws eout the system. It is the rcst comncn -en-machine

interf.7:e dialod. The reasons are histcrizal and econcm!c.

It requircs tower r-achine rescurces and, accc.din to Fcley

and Van DaT [.lef. E. ;, 227], it costs less, because the

7echIne nee. net store any infcrmaticn e.qlainlne ccmands

in human terms. TAe bir-en Is laced on tae man to confor-

to the machine o.eratin6 environment, by learning ani

rememterirg machine commands. Withcut the Iroler co-manis

-nd riles to rse them, the system is less respoLsive than a

jcrtatle Srrith-Corcna electric typewriter hooked tc a

televicion. odern interface des in :ractices li.tate

limited use of this mode.

This rode loes have its Ilace in a ccoq ter interface.

There are experienced users that will always rely on their

own skills. The bare system will Irotably be fister, t.o:,h

mi stakes will be Tore costly, Ir term s of lcst tIme and

4asted effcrt. Less exierien.ed ;sers 'ay exgeriment 4ith

it, ani o;ght to have the o~kortirlty to a.-=ss It when they

are ready. Everyone shculd experience the thrill cf wCrkin.

without a net.

wc. puter users need nct des;air until the wizz' ai..

•:once;t is exjianled to beneric software ;acicaoes. Taere are

rany thin.s that can te dcr.e, .Ith some effort and

ailroriate hardware, to ease the jain ,f learnin. COT'man

45

, . o . • .. ,% - .. . . - , . . - ° o ° ° . . .*.• ° * ' ° . • .

Ia za-e ! t " .'-ontorel tle m f. rm th, as or T'e firs t. -nrd""

Iredcminant, hurdle tc cverc .rn is shcrt tEr .r T7.h-

fcllOwin6 remorl aid5 -an cc 17, ee.terr en r e rs. .

ccut--_r; hardware required is minir-al. Th. =e ,ree of-

memorj ansistanze tc im Ilement is a major lesign le:isi n.

for two reasons. I. the earij stehes cf learni'. a cC.puter

system, .hcrt term remcry o.e~ s at Odd s with the ise r' s

atterpts to learn. He is constantly f'rustrated ty .is

inatility tc remremter :cnands, syntax, and -rinor details of

system use. Memory aids at this learnin, stage may be

either extensive and detailed, to give the learning user all

the Informeatio he needs; or better, the system can hile the

detail, keeline functicns atstract and general, to relieve

the new ser fror details he does rot need or care atout.'

The major design difficulties arise because the nE4 user

ices not remain a novice forever. Within days cr weeks his

needs and wants will te dramaticallf diffsrent. "i thcut

care, .reTory aids :an become vcrazios r~s,;rce -orsumers.

The simllest T.ercr5 all is a lisplayed list cf

'C:ans. The list removes the burd en of keeling several

rarely familiar comman names separate. Every csslble

:cmmand need not le disi.layed. A co~rlete sutset .ne

teneral jurgose =omranl for each available errt-on) is

adeqjate. The disjlayel list is only ttat; the user must

still -chcose the jroier comrar.d, er. -nter it via the

keytcard. Nothing .cre than tne corr.xani v'ord is shcwn cn

46

.. */ .*% ***** %'***.?-- '* -

.~ -s.. J..'P .- ..- - ,

t e list. i.e :nly ;.rocessin., i-vc lver tc suirort t ns -

Sme thod is writine the list to the :reer r ! 1tintai:,lr it

" there. This !etaod is less -seful tc te :ice. Ltt ia th.

least exensive to liilexient.

a. ,'itth Sjr.tax

T he fir~t logical extensi:rn cf tte list ae ve is

to I isj ay representa.tive c -,rTan. syntax. If several

:jtio I are associated wit- the :omar.ad, display the --st

eeneral one. For exam;le, the UNIX 71 s:rEen elitor,

'delete' :0 .rani might be dis1la.'ed as: -n'd*'. The user

can now iein to see some of the o;tlons that he tray not

have rememtered fror his readin6. A casual user, who has

not used the system for a few weeks, seeing the syntax

examrile witb the command is mcre likely to recall the syntax'

rules for the CoMMand (assumin6 tte syntax r'les are

regular; they are ir. this case). The failir 6 of these first

twc methcds is that tne user r-ust know t~e comrards before-

e sets much benefit from eithp- for. Mnemonic a.mar-

r.es can hell.; tut sc!ett:in,, ncre is neeled.

t. With rxIlanations

The next lossitle e -anzsment is to a a cr.e

icrd or short ;hrase of x.lanaticn for the =,-,and. e.,:.

-JW Delete n worls. Two fa toar s af fe:t Ing the des i.-r

rust te ccnsidered: space and interj retatioa. This for.

takes Tore sta:e. Fewer commends may nod te dis layed and

still 1,res~rve adequate wcrcs ece. Careful decisions rust

ce made to ir.:lde enouth infcrmaticn in as little space as

47

*. :K':-:.-~. .*-*.

~.cIn~e trc c.ar n itos are ir~sed t tt

equilment, now caused . ay 1sc d s iay .- :nsra~n

Taken litprally, a novire rrltt Tiszc, str,;z t1e

CCrn to rean2 it celetes i:-!c t--at tegin tt the l=t-.er

n .Ti'e desieners must make sc"'e essu rtior. a tc at Sp r

3en try level r eas on irn.; and experience. 71.-.at dcezs nct rzar

*tine design end Irlemerctatior -dill be anivErsally unlerst:!

as intended . In this rtcde the -lass of errcrs Fr-nerated ty

Ti~uflderst~c c*ues is still a i.roblein. How s houl a the

machine react tc a literal 'rndd' norrmend? A usage rte

disjlayed witi' a ii, of ommrar.zs exklainir~g that 'n' is a

1" variable that mray te omitted cr replaced with an integ.er,

* ray be sufti:lent to Irevent the error. A beci. cr s,:reen

tl ink~ will .rotatly n ct te telpful; tecause the user ies

exp'licitly f ollowing the bitdance. Testine, 1;S I n.,

relresentetive sutjects, will snow which aids ere needed and

where, identify ocmmo n errors, aria inaicate lossI b e

remedies.

2. The Result Sc Far

,that has been a::,cc lis11 , and what mas It :t 7

That dejends entirely on. one's 17int of vie4. kssumln, the

r.F.Tory aids can be irrilemented inexj~enslvely, they will li

niew and casual users t o te ,nc r e cormfcrtatle usin. cr

learning~ the syctem. (Inexl.ensive is a relative term,

mr- a su r Ing ef;ainst ti-re and p~roductivity lo^st wtile 162inin,,

and relear.-J-. memory aids are irexjetsive.)0 T~xjerienced

users w411 jrotetI; ignore the aids, ;ref'errin,. to ust, rhe

area tc e.ja-1 tne z'ri. cnte: t. T e ql:estit.,n of watetr er

to Is ~a y a mer'?orj eid ty i-f ault or t- leave It to user

discretionl lies in user- c ,le ti cn C r-, c z.n I t t h r-

,isers are -jre.4omriitell :lur.;,ed =at one en! -f "V"- cke:tr A,

r-akp tV-e ir-fault selectic. s1.t t~e -a i n i s r-r ,:rc-,;; tut

-Ta~ce the faeollity/ selectatle. 1~e Af1~xitility is imnj.rtant

Ys'!ause It reinfcrces th e r c t Icn t ta t tte user i s I r

o rt r o .

The static Trerr.cry aids discussea atove are p s sI ICle

.oi rcst -,icroccTutsr systems. They arc- em lcyed extensively

by several cor"rercidl software iroduCts. The aifficulty

L invclved in ipklPrrentation is screen rnre,Pment, which may

inVllVe SOre complex jrograrnmin ; (the men~ and work area

in must be displayed), tut Is worth the effort. ',1er-cry aids dc

notc 3 i s anyfurdarrental Com~rand 1 a nu a pe ch a ne by

*themselves. The user merely scans the list, chooses a

ccrmand, ani enters it on the -or.i1and line. The metho. doe5

I nct subvert any capability th~e tasic slfsten m-a. krovide,

such as co~rcining several conmands togetber to su ress

trntev~ediate steps . A ncvice need n ct te aware of the

featur:es tt-at allow shortcuts, t~t miore kqtcwleieeatle users

are free to take advantazge of' them. Man''s turden has teen

* eased In the early learning stabes, b-at he learrcs low level

system letails a s he 6ains rariliarity. The first Jesi, n

*irilications of the span of differing eypertise levels ~re

* bteLirflint to erree.

49

The next levsl cf atscractior erlcys intcriaticn

e.idini and selecticn. Unne,.es sar. s.stp- detail r F

concealed from the new user. "e is re.E'ted a liSt ;f

c.ereticns exl.ressEd in .ernFri= terms. Ccrmard chcicEs are

made :, tyling tne .election number or letter ar. 1;,iyi .

ar 6 uments - filenames, dire:tcry names, modIfiers, et . TYIs

is a more natural #ay to invoe nigh level 3.erations lke

editors, coTTp lers or do cumer.nt formatting rograms. The

user dces nct type the command name, only tne selerticn

lesignetor. Larger operations invoked wit less user etffrt

separate this catet:ory from earlier methods. k single, user

supllIed, keystroke causes the zomluter to zhan~e cler in.

modes. The system bides all detail involved in stoiing the'

1,reviois display, tidying up the loose enis, and startine,

the new lrora.r. The user sees the screen clear and the new

screen airear, 'ust as ne ex-ected.

A full screen editor is normally the first exjies-are

a 'user has to cursor control. Ordinarily input is directel

onto a :omand line, the top or bottom screen line. In a

screen editor the cursor - a symtol (often aa undersr-.re,

-ox, or reverse video retanelel - ma" be moved to any leal

character locaticn. Cursor control is usually ty directicn

keys, Tarlked with arrows, or function keys. Sa- keystro ke

mcves the cursor one positic.n, left, ri A.rt, u, er coqn.

Frcvision is male for zontinuous notion if the button is

held !cwn. Any screen Input is echced at the cursor

d. * . . . . . . .-.. . . . . . . .-. .

. . . .. . . . . . . .. . . . . . . . .. . . . . . * *** *.- .- *-(-

Icsiticn an. the c'rscr moves tc the iext rCsitiC!.n A

-ursor is nothir,, Tore tnan a sy~col tI- fo, s tbo user's

attention. The cursor can te though't of at a i.ortatle

pee. hole intc the mnechire. It srows the user where to loc.,

If it is r.ct wriere .,e wants t3 look, tten ne rrust move _t.

The only way the user zar access the maehcire is thrcug1 the

cursor .

Tn e features discussed uT. to this point are

routinel" fmounl ir. commercial software products. .k tyYi cal

Installation includes a iricrcprccessor, 64K tytes of R.V"

(read-write, .andom Ac:ess Memory), and one cr two seconiarj

storage devices, usually flol.ly disks. The next improvement

requires a hardware upLrede, incorporating a "pointir.e"

feature to ,:ive the interface a more natural feel for ncn-

typists. Those devices allow the user tc symbolically pcir.t

to scmething depicted on the screen. The user is able t.

select any screen location with a single coortinatel eye-

hand motion. Two hardware d.vicps are well suited tc tqis

task, the moase and llh-t-.er..

. A V.1A,%TLM HOP

The jointin6 feature entails irore thin .;st buy rn a

rouse. What appears to 1e a small enhancenent Teans r.ajcr

:banes to both hardware and sul;p:rtinr sottiare to make -

for th.e hilden details. Changing cirsor contrcl modes from

sir.le-step kerbcar! driven, to continucus :inter driven,

means much cre ccm~lez screen mar.afseent. One can not

51...

* * ..... . .. . . & %% . % " ".

exlect tc u;.4rade a perscnal ncre ccruter tz irc)rj;orate a

Touse controlled i is; lay wit. wir.ows End still ret-in

ecceitable resjlcnse. To te sur' the ccmputer -'-'ld do it;

but winter rc lasses wcu i d tve like the Cc:corJe 33T in

corr-ariscn. To fully explcit t .e Pcder cif a .ointer end

concimitant 4indowing techniques reQuies i majcr u rade in

ccriuter zaability. fluld aislay switchin e ieTrands data

rates well beyond that of 5taller macnines, as 4ell as

special disilay management hardware. The fill system ulerale

is sutstantial. The first corrercial systems that fefered

tl is ca;ability -ost aprroximately $100,01 jer station, in

1C78. Prices have fallen dramatically, in 1984, an APPLE

rACINrOSE sells for $2500, thoubh much less powerful, it has

a totally wizzywig interface.

1. Basic Mouse

As a pointer, pic,, or locator, the notion o' the

curscr shanges slightly. It is now a shuttered peephole,

oyened by some user action, for instance, tressina a buttcn.

When the cursor is shuttered it may be rived freely aroird

the screen tut has no effect, ctter tten to attract the

;ser's attention. This mode is similar to the cursor in the

screen editor, but now all cursc- control Is frcr tne 7:cu s

or light ;en. When used as a pick tne shntter is .;ened

criefly and information is transferred fro.m man-to-rmachire,

and vice versa. The infcrmation transferred .I ht nave teen

a menu selecticn and resl;onse. The raw ursor iosition .

information was transferred in. The machir.e translated ths

52

curscr csition tC the menu selecticn and executed that

?hoice. As a locatcr, pressin the roise uttCr Taans: Tack

this cirt. The mactine .news the imlied h e-at cnl ..c-ae

of the context in which the rirk is --iven.

Cne i,:ht ask why one should ,:o to the trcub1le tc

have this mole, since key control is fine r .much 1=s5

expensive. The mouse gives an intan-ible tocst tc the user

because he can now accuratelj position his lin& to the

machine where be wants it just atout as fast as he can find

it visually. Lig1ht Iens and rrice are Fitt's lad devices. I

user can ijosition the cursor usin 6 them jast as fast as te

can pcint to the new s;ot, because the conitive processing

is nearly the same. Testing has shown the mouse and light

pen to be superior tc any other locating devices [Ref. 41.'

As interface control moves away from tae keyboard, the mcise -

or s re other Fitt's law device will ererge iredomrinant, a

no hards controller might te even better.

There is a trea here for the unwary designer. If

possible, allow both ccntrol methods of iniut. Use two

zursors if necessary. Let lisers decide which th y Frefer.

Experienced ty'ists feel more comfortable with tne keytoard

and eschew the trouse.

What interface prcblems are easier ncd? The

structure of the .rograT that runs the interface is very

simple now, but the prcgram Is long, and large cver 24K

bytes of object :ode for one screen is not unreasonable).

The user is presented with lists exclusively. He may Coly

.9 53

9 9 9 . .. ". °

chcc se on F :nmmand at a time. Tre L2,recmenon 3f

sycho1ccal closure is beinr: accomodated [EF!i 71 . he

user sees definit= acticr-reaction pairs, choices have

explicit results. Tne envircn mEnT rFsented is most likely

nierar:nical, so while worcing t.e user is traversing a tree

or network. The user does not need spezifi: system.

kncwledge ,if the menu designers have done a 6cod jotb. The

-user is tetter atle to concentrate on the job at hand and

not the machine.

There are still some shortccmins. Ccntext is

difficult to raintain from one menu te the next because the

full screen changes witn each new menu. Screen manao-ement

.- .* Is a tacit desien challenge.

2. Windows

Windows are the latest eddition to the interface

desifner's bag of tricks. They are a logical extension of

the conceits of screen s~littin,, and zones, tut er,

dynamic. A windcw is a egeneralized screen or ;seudo

display. Arlthing that can apjwear on a normal dis.lay .an

a e eer in e window. Windows can al.;ear simultaneouslY with

other windcws. 'ood windcw performance calls for hardwere

and software support. rhe cleratinh sistem m.ust sul,;crt

T ultitaskin, i.e. a single ser rust be able to run

several active Frccesses an a tine. Some tcrm of raster

ojeraticn is necessary, as are fast screen buffer(s).

A raster ojieraticn is the ability to extract all or

-art cf the screen buffer into a file ani recall it to the

54

screen at ill. Raster ojerations are very fast, the wt :le

screen may be changed in one refresn cycle. Thej eliminate

waiting for screens tc fill with text of &rajhics, if

Iro]erly i.rplementea."

Two window styles are ised, static ar }1o-i, .or

pull-dcwn). Static windows are work areas, usually lare,

and contain text cr grai.hics. Pot,-up windows are used for

displaying transient lists, j rinarily irenus, and are small.

A po -u; menu is only disilayed while the user makes a

choice. As soon as the chcic ils made the menu disapiears

and the choice is executed. That implies that more than cne

window may be displayed at once, allowing the user to

;reserve context as he chooses, like flitiing -ages. Both

static and pop-up windcws can te moved around the screen at

the user's discretion.

C. GOALS FOR iNTERACTIV. DIALOGS

The user rerceives the system as the dialog between tne

himself and coputer. The desi 6ner trust make that dialog as

straightforwarl as possible, in auman terms. The follcwin,,

gtidelines are a starting ;oint fcr the interface desi,n and

im~lementation.

1.. Minimize Ak;r ehension

Attempt to minimize user apprehensica and

intimilation. Sor.e uneasyness is inevitable on any macnine,

at first. Try to get the user involved in the wcrk, thrcoh

the dialod, making the machine transparent. Keep tre

55

.5... .. . ... ..... .,*.*%S %S%5. _".' .' >-..- . -,Z';k>: % - -. '-.- - ' . ' S . ."£.:.:-.-'.' .." ' . .: .'-K' <-.4- -.. .- ,..--.:-S-. .....

"o.2

inte-fice t-sinesslike. Do not try to be cute. thogh Iry.

hurcr may have a place LRef. 4]. Keep error mpssa,es

complete, tut terse. Instead of' evalietin-! iser jerfsr!r.ce

Inr messages, use descriptive trrs to point cut the r.otlem

[Ref. n]. Avoid Judgemental terms in messages: Instead Cf

'Illegal Ccmmend' which Places tlame cn the user and .aies

him defensive, ise a more descriptive ihrase, liKe 'usage

.filename) <mode> , options>'. A usage messae e shows the

user the correct way to invoke the command. It also lets 1im

see his mistake and easily see how to correct it. A .r.u

driven system ;revents the majority of errors like these,

most of which result from careless typling.

* 2. Dop't Panic

Minimize user frustraticn and panic. An uneasy user

Is more likely to make mistakes, esjeciallj if the same

nagging fault is causing his displeasure.

(1) Panic results when the system does something unex-pected then locus up, without telling the user whathaplened. Always keep the user Informed atout,, thestate of the machine or, ,rocess. A simpl e run-ning" cr "stopled - address fault <prccess name>is adequate. It should always be dtsplayed at thesame screen location or zone. If sometbin doesfail report it to the user, then locK uP or auto-boot.

(2) The user should never feel lost. Keep him informedabout his location, either with a world view" di-agram or hieblighted menu choices, etc.

'3) Always let the user know what options are availatleto h i. Guessing can lead to disaster. Pol-lipTenus are excellent for the task. Otherwisedisplay a menu unottrusively at a merein or tcrder.

56

A"4. ..- -S.' '/ :""" ''"""'/ '-" '""-. ." ;". . . ...-Z_ -- ' . . "-"

*. .. -. . %. iW.~

'4) A. cmmcdate several levels cf exz.ertise. Try cc putt*.e iser in control here. No one lE1kes to be clas-sified, esiecially by a marline. A , for ex-perienced users Is one lcssitility.

(5) Make the iser feel he is m a'inx pr:,ress. Stresspsychological closure in ttp. design and imllerenta-tion.

Following the ;cints above will make for mo re

pleasant error situations, and if the user Is mcre relaxed

he is also rrore likely to recover with less trauma to his

ego, and his work.

3. Natural Dialo.

The dialo. rust te couched in terms familiar to th.

user. Whether it is words or ;ictiares, they should be from --

the user's sphere of reference. .-.

4. Solicit User Comments

Systems grow or die. User comments will tell the

desiners what is disliked atout the system. Thirgs that

are 6ood in the user's eyes are invisible. Annoying faults,

no matter how small will not Eo unreported. rake sure the

comments are read and acted upon by some one with the -cwer

to rae a difference. Not all prctlems will te repairable,

but at least they will be known. Feedback is esiesially g.'-"

important early in the design stages. Try tc ;et actual

users on a new system as early as possible.

5. Key.a4r4?

For ap klications that are not heavily text and data

entry, or word processing try to minimize keytoard use.

57

5..

.atatase !rani ulation alplicaticns are a 1ikely can.hidate

for a wizzywi interface.

6. First Session

If the system can not te used for useful wcrk frcr

the first session, something is wrong. A general .urp:se

set cf com!"ands and a;..lications should te availatle dith:ut

any system Knowledge other than how to login. A *aser who is

atle to sit down with no ;ricr knowledge and leave a shcrt

tim e later with a completed document, and a smile on his

face, will return. Explicit training done on the system

should be done in short lessons separated ty distinct

breaks. Ideally the instructions should be designed for ten

to fifteen minutes, with thirty to forty minutes for hands-

on lractice. KeeV the user engrossed in the joc at hani. If

he must think about how to do scmethin, on the machine, he

is not thinking about his job. The machine is Just there to

help, not dominate the taskc.

All the experts a6ree or. what to do and what not to

do. It is also obvious who among them have designed an

interface, those who have tried, and those who nave not.

They aleree that Interface design is mcre art than science.

It will remain so intil comutuers become so jowerful as to

take on human traits. Or eenetic engineers are able to grow

a computer operator personality.

The thing to strive for above all in interface

design is to capture the essence of a metaphor. Information

58

, .-.. ,'%*"-.:--. '-8%-

tre..sfer, witb the ail of a metelhor, is trers crTe: from a

spriel data stream into a troad rand t'locA. Tne user learns

by analogy instead of by rote. ie is also T.ore Likely tc

reremter tecause ne is tuildir,. on his owr. experience, and

the mental link seem tc be mach stronger ref. -.

59

. . . . . . . . . . . . . ."-%'-" -

V. AMIAB E UNIX

As originally conceived UNIX was a researca tool

desianed to run on a small -omputer in a laccratory. it

tecame ;orular because it was easy to use, versatile adn

readily ex;anlable. Now that it has been in existence fcr

fifteen or more years, the cnly feature it snares with its

first versions is ease of exiansion. BELL Lacoratories"

baby is no longer easy to learn, and to use the wonderful

facilities requires a moderately powerful rachine. Amiable

is a overlay for UNIX that conceals specifics of the

o.erating syste! and allows a new user to use the system

without first tein. Intimate.

Amiable is implemented on a SUN model 150 work staticn,

running Ferkeley UNIX version 4.2, under SUN release version

1.1. The system is part of a researct project for the Naval

Electrcnics Systems Command, Command and Control Work

Station.

Amiale consists of three major parts, a welccnin.-

screen, the working environment, and an on-line VI screen

editor help program. The wcrk station is arranged so trat

the monochrome monitor is the text work area and the cclor

monitor is for ircmting menus and graphics work area. Each

part -gill descrited and explainel. T1e final section

discsses future work.

- . *.•.-.. .... *.'* *. * * * .-

--. -- 7-

A. ALCO E TO SJN UNIX

The Welcome ;ackage contains keveral roram .rod ules

that jislaj Tenus orn the S':n color monitor. The screen.

choiies are made witn the Sun mouse, ;,1aced to t.e immediate

r.ht cf the Sin kejbcard. The entry men.i 1,resents the user

witn six choices and four short notes. The rren. is in color

on e dark tackeround.

1 . Main renu Notes

The notes ire very simple and to the point. The

first note tells the user tow to select using the mcuse.

The second tells him how to logcut. The third and fo.rth

are irnplerentation notes, and a point of contact for more

information. .-

2. Main Menu Selections

Of the six selections on the ?rain menu, three are

functional and two are under develoijment. The sixth

selection, actualli( numter four, a programming environment,

is just a title. The olerational selections are:

1) Demvonstrations.

2) Amiable UNIX

:) T~.e C Shell

A program to select graphical attributes with the

mouse Is being develo1ed for future inclusicn. The gra;hics

editor demonstration program is choice nurwer six. It was

Included as an interim graphics editor, until it cab either

be ,;graded or a new one written.

61

AIR.. . 77 --

he Tein screen is si r'lr and unclittered. ':her. a

choice is made the main Tenu clears, reilaced by either a

3 emonstration menu *r an attritute ienu. Amliatle ari C

shell selections clear the main menu tut leave the welcomie

screen or..1. Deon tra~t Ions

Sele-ted demonstrations ray be chosen frcm this

menu. The upper section Is for ironcchrone selections, the

lower section for color. The demonstrations are from those

suptlied with the machine, and are relresectative cf system

_"a at ili ty.>.

4. C Shell

This selection is nothing more than an escape into

unadorned UNIX.

Amiable

The second selecticn is the Unix overlay Amiatle. It

is a shell scrilt that helfs the user do several things in

Unix without any system knowlede.e, other than some ,eneric

comruter experience. A C Iroeram version is in development;

tut is not yet reliable enough to unleash on the 1.ublic.- .

. INSIDL AMIABLE

Amiable is a mask for Unix. It incorpcrates tne

interface design ideas discussed in Chapters II thrOugh 1V.

It was written and implemented on the SUN, tefore the

windowing software was installed, so it does net incorlo^rate

windowing, unfortunately. It does draiatically decrease the

62 .p-

aTICuInt of kncwied.e the user must tring with him to the Unix

o-erating sistem. In most casec Amiable eoE-utes c om.nas,

the user sup;lies only tile nares. It nas a terse tc, le.

When terse is chosen all .essages from Amiable jrom t r..-

input cnlj. Explanatory remar.cs and more complete requests

for input are displayed with the Iromrnts when terse is not

set.

Amiatle greets the user with a list of functions:

Brcwse Compile Edit Learn Move Print Shell eXecite lo"out

The user can choose any of the functions by typing the

capitalized letter in each. He may only choose one function

at a time, but he may exit that function, leaving it rinning

in the background, and choose another. Compile and Print

use that feature. The idle state is with the function list

displayed waiting for a user choice. Amiiable asks the user

fcr additional information needed for correct command

execution.

The edit command causes the VI editcr tc be invcked cn

the monoc ,rome monitor. It also brings ;, a jirogram calied

"vicom" on the color monitor. Vicom is a Iair of on-line

help and tutorial screens, disilayed while the user is in

Edit. Vicor" is intended to help tne user learn and renemter

the VI editor.

.p*.-..--*..

* ... * .i n * * . ~ . % % ~o..e - .

• . .-. -. .. * • % ,°°~. "

1. 7t.?o

Vicom is a Tcuse driver help filE. All selecticns

..within it are done .Vi the mcse. It is two color screens

.ontaining VI editcr cozmaards and syntax. 'he first scr'er

;resented has five general information notes atout the

screens and usage. A second miscellaneoas notes section is

irclided on the first rate of commands. They are notes

about syntax notation used cn the screen and sone

precautions about usir.g the moise.

Corrand meanings and syntax are resented it.

functional groups. The tirst page contains grou s for:

cursor ccntrol. text insertion, scrolline, the screen,

searching for strings, saving and ending a session. The

secord pege has more advanced functional grcujs: . ncvfl ng

text, usi. the buffers, and X ":" comands. Within each

functicnal group one or two basic 04eneral jurpose commanids

are highlighted in a different color. New users are advised

in the general information note to learn them first.

Eacn Iuncticnal group is independent of the others

and Tray te dis.layed by itself. Withir. each functional

groul are subgrou;s, information rressages atout the

subgroups ffay be called by pcirting the -noise at a coirn!r._

end pressing the left button. The ten or tever lines cf

explanation are displayed just below screen center. The

exl;lanaticn contains usa.e notes and additicr.al syntax

rules, if any, as well as a reference to system

docurentation.

64

7-1

- S - . .- .. .

Two exercise files are on the system: Vi_exi, and

VI. ex2. They are designed to give i ne- user m-re

information about VI and some ;ra:tlce usin thc ccmrands .

A VI expert could d: the -i.'le exercises in less t-an

thirty seconds, using the .jiekest ietho Is . . rew user

could take half an hour or more. The lessons are not

diffi-c'jlt; but the second or.e has several text eiitIn,

oerations that vill keep him busy lcoiin. at the ccirrand

screens.

C. TESTING

Amiable is essentially untested; time ran out.

Criticisms and coirments from six colleagues who used the

system briefly have teen Inccrpcrated. The system is ready-

now for some exposure to a less well prel.arel group. The

results of that trial by fire will not only tell whether

change is needed, but how much. The rewrite is inevitable.

55

---

.--t

VI. CONCLJSIQNS

Zesi,;nin4 user interfaces i s easy. 1 l errenti rg tnhe

desien to ac- ommodate the foilies of h'vren teines is vnard.

Literature on the to~ic of man-macine interface aesign

abounds in trale *ournals, textbooks, hobty m'aazines,

literally everywhere. Each author has a new wonderful p~lan

to make it easy to use com-uters. There are only a hariful.

of authors whnc speak from eiperience. Unless an author has

credentials as a ;roven interface desitrner, who actually

worked on the imp~lementation, his Ideas are suspect. Given

time and rran~.ower anything can be coded; but making a:

ccherent package that works tcgcetner is a differenit i.rotlem.

A. PEOPLE ARE SIAN'E

The tI e s t s turt in.: ticek Is -eoiple . They are

unpredic-tatle. Differences between two groi;s of pecile

that dc the samre work In two different offices ,ai st~j an

interface design in its tracks. And none of tte faults will

te dis--overed .intil the testing starts. Heckel has arrived

at the same conclusions R e t. 9. The answer is to -et

rototypes running! as early as ;ossitble and , et It to the

users. Solicit their comm~ents, criticisms and suggestions.

'ThEn do somrething atout them.

B. 2ESIG% "'AINTAINABILITY IN

An interface must be aesined tc ce :han eable. Use

every modern software maintenance technique to mrake the

- rcgrarrs easy to chan6e. 1cdlarity is a key here. If a

screen is a rodule, then rodif.ying the screen is relatively

easy. Within t he screen module make the layout cne or m.cre

Todules, each menu should te one sub-module. Lccument,

dccurrent, document! Assume no cne from tne ori-inal Troject

will te 3rouni when the zharges must be made. Explain

everything. then the new ;eople will only take half the timre

it tock the original groul tc become familiar with the

material.

C. DIFFERENT DRUMMERS

The group that sets out tc design a user interface

should nave Temters from disci;lines outside computer

science. Ideally the groui. should at least have access to a

cognitive psychologist, so"yecne witfh a visual arts

tackgrcurd ( ainter, gra hic artist, 1.hctogra;; er,

cinematograi.her, etc.',, and representative users.

An interface design impleTentation is too larbe a

jrcject fcr a single perscn, tut the erou must be

responsive and flexible. rivide the rcject among several

members (three to six) and schedule twice as r-,cr

develorTent time as is thought to be necessary. Test at --

every oi.ortunity, then rewrite and retest.4v-[

53-

Real himan-ccr;uter interface lpackages are never finished,

.n eu~hemistr for final rodu:t is: "-telease Version XX".

D. BOOTSTRAP DESIGN

The current methol cf -r,'grammint screens is not ir

effe.tive use cf resources. Tco much tire is sient epttin,.

simrle menus, and tr.eir associated selecti n moiiules on the

screen and runnint. A program or #rahi:s editor shoula be

atle tc handle that task with scme develcpmen-t effcrt.

A grahitcs editcr is the second logical choice for a

workstation after a word processor. The next project oa the

SUN work station should be a graphics editor capable of

drawing executatle user interface screens.

The visible result cf an interface design j roject in no-

way represents the amount of affort exienlel. A methodology

.ust te developed to steed the process. Present methods are

*too hit or triss.

•A 1

:--8

- ~~ ~ ~ ~ : -6 i-7 .- -.-'f EI

PROGRAMS

-

01

0 0 (U .4

4-V 4

41 S-ela

-. 44 .---

C.)~ ~ 0 . 45 *

GJ '. 4 W

F- .4 V V -4 l.

r- (L ) tcJ E- Q.4~~. 0)0 ,-

to 0 41) 0-4-1 .C) IP4 0. (19 L a

0- C)'a 'a'V 0

1. (UQ U') 1- in0

w~~ 0'. *-a , G

U')~~m a,. 4 ~ 0 4 . */,-

00, CU5. &n. 04 CA4 C.1 *00,.-a to -1 a' A 1-4 4

C, 4 4.W 1. 4. 0- 0zr o WQ

- .. in A Ua Aa K. *A' 4 iA 0. 4 30

0 o r4 &V'. - .4 ilm E - 10 t ; u i ( oFC 1..0 -4/ v E-4~ o 0

I-P F r-4 -4 * Q 6-a w to0 r4

5-4~~ ",c . .0 4 4 ~ C) ~ I

to -t- 0 t a-

* 0 0~"0 (I (0 to 40-a-clO U *"' .4'

iAJ~a ... -4 v4 10 A #4 "V4 4) -~.

C)~ ~ ~ ~ ~~ 4 a, ul' (D ''* C4. J - .i~ 4~

.41' U)0 U'.4 I 441 0EV 5ro.-. .. a~69)U

* ~ - -~ - -. - -. . I I I . I I -v--.- .- -

U6

N, cn

0*iv C3 *

0 a)C.3 r- C.;

W4 0-C.) U0 r- 0 CA - W

f. U 0 0c:*0*

CU W. ou 1:CU ( -4 ) r) W (U 4 pa w W4 *

V 13 (_, C- j P40 e r. - &

V .0

4-a

0. 0 94 *

A-, fa - 06

-- *%, 3t. Clip

* ~ 1. *- r4 , a-

0 AU *xi) - Cj u a a) - d U

o ~ ~ ~ ~ ~ I 0 q4 .. C) '

6.. (.- o ~ 84 fl-. 0 .~ Z~v?

a,.. % ~. .. . ~ 70

w (L

*C* a. a.a

** '60 =- I

co 0 F 5* ~ . 0- Up g U

r= 4-a S ..'

*d 4' '.=

v c') C. toI - a qA

*~ >b- -4 & *E

r-4 414 In W)J

a ) .V IV QJ k in wC

* 3* to'E 5. O U- N bS r .. o .b j -( U $4 o-14 w -r kI 46A ** '% Ph .i r.bo ;

1: w.. A MO 4.'LA4b )0o t

r4 "A 00 F. aV r.Q 1) c

02 .40 4.) t i 4- 0 0

O* 4. 6.. 0 0

*4 -4 4w A fl Ei

* ~ ~ ~ ~ ~ ~ i * kc * . - . -

I7

I/i

E

-4AU

UJU

w 04

0O 04

rfaJ 0W4 L

4a

0 0)0i %n

aam azkn 0(L) 0 94

.4v i4a GAiL V 40 0 .

w41 U) P :.

F~ 0-a.

(U 0) :3 0jC) 10q00 V) V-4 CV-

o a V4 pn. a VO I ) )S -4 S) CUr- 0) *a P-4 - ..h, 'C *U - 5k1 4 *a =4 *a -* a *A P

-r4 -r4 to N V uC "7( 0 0 4-a.u4 C~ Ns 14- LC-. .V -40, * !k P. 4 1 C2 V-4 1..I.A t : ! p : I. !

V- -ft r4o v L: 0 4 UV4 31 4U- f -4 X~ f N. If -4

0)- l 14 t o9 k 0'd m- CO 4 4. -4t . U , .*'P

W . oW NWS- o' 4 .81- .4 4-b . 0 & 0 0) 0 * 00 14

4- 4. -0 ;b -4 :0 *a 0 0- *. -. m 4. b. 4 m * *a

0.4.1 0J0C1GJ)N 0 1. w(0V0 4) ) ) N') C '4) 0\~ w 0

Aa4J ~ I S.. d) vU4 all 0. W 4- 4 E W .E 4-p E4-. 1,4-a E V-4:

0 ECC) 84

72

y7-

.) in,

-. a-11 -?"

'A. .-34

oil Ln,- %

U.

CU~~~ .° t

0 0 4A -4i0

to w.C.. . ,a), t. "P4 m 0 V d. -

0 S0 C) 10% €-_) 0-14 C.)"*. -4 Q,= 0)- "-"

W 0 %) W O t- wJ 40 0 w Q, , aj a) ci , ""

;w~~ +- N bo0ao

.- .-.

*,

73

* %*o

N-..

%N°%N.o '

,".".-';.-.'...".'-.-...;"-,-. :.'-,;-"-"-...;,."4"..'% .... ;'" " .'.''..'"..' "" - -'.-". . .''., .' .'. N' 2 ,",.".''.. ;,''."," ' :''.- . - .' ". - ."."."."

kn Q)

Qj2*4

*04-

N * a*-

F 10 C.-41- L

0 f- u * .

t* 0)j 10, r-. 0.

4.3

(U (D aO N*a 4.3O

-4 0 -4 *0.*mw) M*4 0/ E-4 X.-

* ) g. Q) 7% o * 4

* 01 1 !S)

43. 4j an :n I= kA* 1 '- LCZ1 =.J 014 .0114

*n - to 'Vn li l (N 4 C 1.9 C.- a L un -

=00 0 M3. c :jxA - 3 tF. L)

W -4 L.4 r- o~7 0C

* ft. W.- I~ C) S"4 * r-4

=1 IA IL E- S6-'d f0 to 400 f

L ) E- *-4 -4-4rf-4 ' -)$a

4) 1.) 4.4 0- -I) 14- E 0 %

A. r-4 r-4 C .' H 1-'t- .

cN r Q) 0) *j .-4CtoJtolto v-i - 4-D~- 0N in --- - 0)A -A- -*2i

tt 4t o. ulZ ul k l 4 U

5* s ~ CO z~a~ JL 4 4OO~~74

VC~4 -... ~ ~ Uo

r4 r

:3z - 4. CA 1. =~r

md*x *a4 04 0-4 .13 r-4 F- 9.4

* ~* .4 .94 -Id4 a

* S.4-4 71 z **

* s) ~ 1-414 1.4 04 1-4 - . 1-4 P-

W. W. ..: pa~ -4 -C .a -

V1) 4+a 4.p34-a o . 43 .1 4.30 O.1 03 *

cz *-4 -4 q-4-G -4

q4 .4 -4 .4 .4 4r J -4 -4 -r-l N, -1 -4 r-4 r-4 u-I ai q -4

-44. W-G 4 W4 .4ro 4 s W4 4 -4 .4 .40U.4 .4bw11 )* A2 A2 W 00Ao ~ C 00 A Q -4 Aa A2

0 4)Z - 4 W4 .4 "-1 *I4

4 - .4 w4 4 0 W4 .4

S- ~ - I ) kn U1Ui . inU .) w n %A n & k 4 vV ' 0 V-4 ZQ -. .4 .a 4 -4 4 'U, .r4 w 4 w4 ..4 .4 4a) ( ;p ;b .0 Vi -4. - g .a .

.4 a- &I 4.3 #A 4.4. fl -D 4.a 4.I- 4. ic- *a

w nf (S) = r. 3 = 0 i0 t-. 0 -- *.a Q) -r4J (U. GJ 4)c ) ) 40 CU 0i() 0

C- m. &4 E4 k0m F- h E k-: I hC F-. -CC. (U 4 l oX '0 O U *4 i l I - U 4 1±-

0.. cn Gjo I== . W a. 4 wa ul ul ' V n l U

11 9.4E 1 . "4 0. 4 A,- 10 0I' Al 3 I .. P) I 0-b I .- q. 1 4a& jSUJ .4. *v As.0 4. L.43 4.34. 0f W4. 4D 4. 44. ld O.3 *.3, @-3

00 GJIG 4. 0 l i 0 L0 As..A-t.(0 CU 1 10-4

co v A4A3

7S

E-44

E-*

Ei-

E- '4 *P . 3 L

C ) -~ ; -4 *. 4 - r 0 F

13 c

W to- Si 94I r 4 0 u 0 (

u7 tu 'Ar 4 -maw -4- *

;Q k. 4) l F4* t 4W A ClA

w v -I w o- 4 4 r44;

_ 4 A( in .- E- 0 *

-0 d 4 - 4. o-4 * -4 ,.4 *id CaU 0.a Ct U

Qo E P 0 %n 0 Q a 4. IALn .

.I- A2~ I (D.s- (U d) *.

O( )( Oc )9 u%

00c - F- 6-E 44 Lm 'u'_ ~ 46J !.4 .

enwl dt.L 0 10 E46-2 4J0 "D I ei *

0 .0 5

' 1. 4 C P4'0 9 % 4 fo 041 G~s.

0 .0 if'V.0 .U~ 4A toCf Q*- w*4t*a bd- 0 W) ;we mnD #.v a- cu 0 0J4 N.

% 4 4. £ .W.4 Q .1- a c ( Q 4 I.- to ou L

d UC iI *- ) o o %i: .. GJ OJ -0N. n ja~~wd4*~~. W - 3.. Lt...- ~ (U

o.4o-~4~ ~ ~ 4tf(A

E4.D 4 ~ a~ ~ 0'1& Nidi.~ iiOJO LIi ~QG~%GJCAGJO Odie*GU~0

* ... ulal-u 1 ru 00 iD-- m10U4J 76

. .. .. . . .J. 4 W .d~E ,-~ Q O

.~~~ crTu-

L.0

0

0)

* * ~ 5-4 04A- * 0

Q : - 0 di A..ai E n 0 .4.)a

4) ) 0 L 1 -10 A

5Q E

* 0 .- 4 0)0

%- (U (1 C.

c 4. 0 4 S* U) (1 W r4 ,4-q th

44 E .' -4- r-u r (

* )'4 in. O* A *&C; 4E

0 x Q(D P4 -(U) r930 rU 0 0 > (n 0r, r4) *0 EZ~ 4-' -4 - i.)0 ( -2 C-

ro- l* S.E U) -4V* d~ 0 4 E~ 0) 0.;3 W) U)4 X: 44 o ( . 0

u Al -'n' W. -.4 0 )1a )a

*U 0)C ~ 0 (1) 'a %. w . 0

-14 Ix w U)3U :3 r-= ( ()l , =-4 LF" -- 0 -E4 v) u- .- - V) 9(

*w 41U (v- ea* .

.* ~ ~ ~ T la * 0 .0j . 0U

* ~'~ *. 2 *-..-4S-eUS~77

'pu

(1.) (U

Le 4*

r94

1.4 10C

q'C

IA41 1 w r0 :D P4

V) L- * _ C)

4.o *J4 A

(U GD Sj (L 1 :.*a *- 4a " 4 *

0i C6- 4-b1 4

.14*o ~ B-'

- 'b 78

W * u*2

w* 4-a *

kA*-4 a)*) j

0 L. E E L

Up 0~> 0

-4 E W* 0 0.Q~ W -4 E '4

J- ~ -4 0'U00 (E

tv 0 10

-r4I

*~ ~ 0 fa C E.--- -

0 0. -4 000

r-4 -a

-~0. 4 ,U 0A

.0 tS. vi. 1)j C\l

r- rq.40J Q~* in P ;-L - ,

to to - -4

9 4 ton I ( U

W1 Q)..4,~ 0.01S 0 0u ~ ..

-4 41 #a AU r- r -4 E O tQ

*~F to -A)C A '4 4 I

*~~~~~ .. V1'-I0 -~~~

* ~ ~ ~ ~ A itA. tn V~NCI 1' .i

* E~4~9 11* ~ L-.J~>, -4-~in~~79

'0 t

0 (UG

P4 (U

00) 0)

Q1)

(U.

4-4 r-

N, -4 .

)44 4-a.

;w :3~. 4.)

*~~g ;24 b-,. *LO ' .p ~ . P. P. A.. a*6 :3 a~-U fN- 0 -f .4 %nl U, knnt* V) Q) 4L4 . - V*. 4 O i-i4 * *..

'4 o $m.-o -19 o- .. a (n UQ *0 C0 0 z* ~E - ~ a- 0 4. EO w -(U- *. -P -u k *; *,.I

CV M, *5Ijo '( t @ Q C3 -40 L* t! 4 %n~ r. S)-' ina -a) WL-4 I.- * ~ 5(U 9,4- ~ .)42 .V U a

W I ( ~ v IV: o t Qr -.4 C\ -V4 2"--4NC4 qW S.. P-4 CA -4 -4 0.-4 gz 0 CI.rq 3 -W r4~ M '4~*-.g *~~

4) 1:4 QJ i r-4 Q) o W :3-4Q Lc.0 Cu,,ha u 0 L43w j v w (u (Lw

I 9 r-4 60 - 1 r-4 "- * " : o*A- '>b, 44A (U 4)o0 a -4 -. 4 0 4)OG .40 ) 41,a 4 -D 40 W UQ

Ob 4 r4.) ~ 0U %- 1 1; - Aab *4.o 0J %--4 0 (U Qj 0 (L) CU 0 0 0~lj o

E.- CU- C Q 0) V) .04 ) Eof AL-

L 6-

80

.-- ,. .I-~-.- .0

00

-l *) C n )-

U) 4--

0 31* - d-4 -

C4. > o > PN- 4

E-N

4.)81

N.7

i~

* 0

S -4 0

33A3

*~~c r-o0 4- **

r.. a.

Q* > 0 (',0O

*n 0-4 -

r* 00 0 4)1,go to C4C)O

to M

FE 0) 0 0 - a 0)q

11 11 on. .. 0 A,( L *

A. to=- -0- AEb,

'k S-4 0n0 3

IA 1) 4.) .03. p-r &..~ 4

* USV4 v 00 A 4.3~ '04-b14) .

* .r-4 4 0 aj IPo 0 4 0

E0. 4V U U W 4.44 4.0 0 0- d) 00V

= C: * r~4 * E v I -

3*~~~ r- .. ~ -4*~t (a0 -'. -4

82

**

L

4.'40wlU

4.3

bC

C43 V3- n& , nU

A2 4.P 421.Q

oq. (L =- l -- o e 0 - - C

0) A,0 UV 0 JNI 4F)q0 NC q4 V

Ig 4A 1 141 ! I.

fo tv ga k a P- * im 0 41It) At-N.

U) .j 0 0.l 0N~ a)* 43 *-..U 0 ILI. tii* 0

............. U 4 ' ~ u.~ ~ .' ~ -

CU Uk4.X 0 PQ

v*V, foIA 0 cc *

E- QCV -# C) o 4- A- ADP4*P4 0 fa

n k'-L-4 -r-j -UJ -&A*

>*0 it, .,j X4 .

*.04 4.b q

M o- N C LY0 CV0 CQ0 NQj 0 W4-f F- r-4 -4 1-

0- P4F 4- -P E *-

-484

t .0

4, 4U11

N 0 -f 0

* n

* p = 410-

g io 0dN'~ -4 44

LO -4 0. V4

a) 00 .. 0d "4-

p NO 2 -4-o 0

) 41) A q- 4V

*l 0o =1 4- -0

ada a r1 0 paC1 0

A Lo 4 'O L

(D P* :4.0 . 004G C3 -- 00

-4 O43 . Q;- 4CU 04 1* .44 ' l 0).1 Pc* 0 43 t

is 000 0 4.3

v'i 1-4 U)o4 o. 4 . % s4

,* 4) 4)

r4 W3 .0 =J4. 4.3 ) O4)G

*0 C.) -r4 1-' V 3 ov o00-00

O'~ UU -P 03 4&43~0 aA ti t4t i ti

N' 001 0'.- -.4 Fa~z~Or~ 85

.1*.toCDN) mWA

C) zc*

4L i -4. C&I te) S) *

Aj .s v s) sW r

tr; -4LX) I fuoo . .*

* u1

0*u -A3 .0 c

=I IN00

*a4 *a. *a 4 ** V

q- J=- = c upb

O A 6a*4-a**a

#- - *D6 4aOo 4a" 4 * -b r .4 to.

IN 0 GJ000 )L l 4 V

'TI * ~ ~* U.86

>44III -4E

4.

4- 7

0- 0

J; Z k 3 . 4

-- a 03C - d d o

- 04 (U .L 0. O* to n )=9- 1

ft zr4 I -0 1& )In 6 a% n % 45a 4

Ij - . - 4 -= 1 4 1f4 I I-_ .0 -0

ca- o 4- .4.- r - (U 4 (L - )

M 0 t &a 0-. -4t - CM1 W &a ~ N )C4*_4W- S*4

ulA d u~CQ 4-0. W ) J0 wv tJ I0 * 4 .1f *i A s -1

q ~ to n I ~ 1 *4 k I 0) '. . -,2 C3 . 110 - 1 n $z0AaC

. J4 j-4 eW -'N E C (Lo '-0 (U -4 to '4 fo .i4 4 .0 v4 o

~~~~~ I4. w. Q) b-1 W *0a.-~ v . . . ~ L ~ ~ . w- t4 I3 >6 N- W '-. fla G,, 0 0 W 1 413 *a P a,: - . 0e4 0, *A r. -I -to

4) d) a f.CU r- S Q Qr. w4 w.. ,f0I 0 (U ef W ( 03 0 0 (L C.%n % uln t viDL C0'L.) V)l kn 1 r lE aF

87

S-4 (U t

At -4

,IP- .10 4

0 4.)

kn4 Ll~to

0 4A

Q)o An .a)

An Q)

CA Ej * GA.-4AaOl Q) .4.4 a a

to 0 (U-Q

to 04nG W4 -4 4." 0r I

0 E1 Awio~ cu OGA (U. 0 aa

ox 6 -IF4 0 * Oo 'q fu~~3 - -~~.',.4 ~ l * 4 cfl

4j r.4 * -4~ (U W ~ AD 1. 11.

-434- 00 0A *14- 000 .G e-GAG to) t.-4 6.- , 4

En . Q) '- Z A 4) 13 0 4

04 0- ad :3-

~~~~~~~d W. 4 ~ A ~ i 1r lr -4 +a W to W #iJ-a' vo 1 0 6-4

U)U0 4f J 24 r14 4.) W

w ~35. t.)~ Q.I.-4 to r. Q)S.. 0iGA - 1Q... 4 6-4 *4A -

-4 C.) .4.E (a ~Io 0. tip j4'1 0 to (0 .

GA E r'4. . ~ I.b'-4 -. *6- 0 .4 to t- (U -0- :3 C\I O-D

GGo 01 0 U q- 04.G OOGA 0. qI

0n 4. U) Q) h-164) f -04 44 0 f.AG> 0AG

to0

144

88

W:. 1-.-- m

- -'- -. ~---.-. ..- - s- * - ~ -~ -~--- ~ -

*b

*O

4-4

VI*

39 4.4

-4 U-)

-4~1- -- *-4

C,-4 *

~ 4z

E- E- U*):- 4L10

a-4 Ok be= r-*av a) a, u *0

Cj~ C.) CJ 0 0 o( U()A 3

a4- 4 4. e a C)0 . 4 AA -

> M C w w.. j ag to. 5-" "1 C\ S*

L) (I (U IO* UO j - 444444r 4 l'

at) 4 T V? afC. &0 t.C -&3 0a. 0 ;3 *

*.-.fl Q)- (L. Q*() pko t n4 * 0 I n -

(a 94 24 2494 . Ia IJ 1 3*.0 Z

() a- 4Jl. r-4 *q-4t o-f -4 j C 3 *

%o V) a *& 6& ".. 444w *-b 4.2 !* : =

(U - 0 00 *e-a r4EOw-4""a q ..P .

.. r. = C: Q =2 *

4 W- I 4 @ @ q-4 4 -. -4 L.4 tn u 3ov,% m&

.~4 A UU (OO V~ ~ ~ 0Cf *

~ 1 - 4CU ~ * .- 489

V7

r. . wnr4

04

0. 6A -

4K 4.) (U(94 5Cr- U -

%o to. #. 1

.4~# 0'.U.464

rW &4 i4 -4

%A 0 3cQ -4 * .) 03

cu 'a .4 (S - -c

F.E 0 kn- . (1 U

00 (U 4). U4a 4

1.a.. 4. 4m4 0) NEJIi 4) £L W 00) uU -4W

to 00

W 0 04 W.kw

o ow .4P O CU 9.QiC

oj -ef E- (E4'1- 44 U )

1.4 44 4.4 14 = IV

04 1 C) 0 44, 4.a

4) 900CA0 vG.0P-4 4nQ

14 (1) 4. 0. ()

U~ W 4A~4 4J 40OA 1.4 9...L coo 0. 1- 000 441

Q) -a4)OUC.0 % -*

E-4 O-4 0 .4 6-a

93 0%0 93 4.'41 t4 t 4 -It4t %t U4t 4. (Ut £2 o aV - ' iV

0 ~~(-J.A-3 el 090

Ud)

4j1-.

V10.4 c

cu~L >1r f 3

C1)- 5.4t

to Q) 0 4) r.

ro- (4o

4- >b IV1 (v

.-. 0- E-C

it V4 0J (1 0S 4 n4 -4 E-4 -P

0 1 0 % 5n.4

.0 CV (4 - 4.

Q U 0) 0 444i vI l

%n r4 j( -f tv4n ) i/i4- I

v 4.C 4!b 1-4 o.0 o 11 X3 0 4

VG o .0 - 40 1.

U) L0 -4 H~ Q) W.E(W~L C) OJ-J C3-,I ~ ~

4 .4 4 0) A:3.- 43 0 1 .0 ) '

o ~ ~ ~ ~ o W1' ~ .41 - .4.k3P4 CI ~ .4 t- EQ 4-GNO. 0 P4.E) 0

0t o4 0 ) (. 4. J- 4 -cJ 0 C4 ML WU-L 4-ca--.4

-at0 (4 U 0 a) l 14 f--0

C) .qf f.. Api't 14 190) -4 -40

0 (u OW .

LA 0 1- d dU 3 14344 V)J

m~' 40 U. I~ .U ~i)U~0-~ )0 u1

0) . .4 0 0 ) -'-4 0' 1)li, 91,

b4-

4.4 .4

V1 -4 v (a 4

E..*a 0 l- (

.0 0

4. C-Ar

4..V '0 t4a.

0f (

C'- 4.2 0 E 40. If> M

'4 C (4 - Ocu 1Y 4 Q.D.4C

er' (U 'U 'U'

If3 A in- '0 =!-.1

0- V) P.0 Ar"0-r do C 14 0

4A0 W0 ul Wi S0~i 4)4. -- 0 fl

.00 0 4 511 0 4 i

0 U Hf 6-4 4- 44, Q ) J -4

% - 4- r- 41. ;;P4V 4r#-. tn a i o l n

lo 0 0) = -do W. ) CC0 4&U* ri1

I Q Q V. . -4 *- L'Q = a) .U 4.4 rCj Q4

0 ~ 0 110 1-Q .A 4-4 A' ) 44U W r 0ft 0'.. LO( ) n-&. 0) 03 064 C44 4) W4 (U )5.

4 0 4)V. in- U t0 'UC- %n.0Oin Q U 4.6 93 '44' ~

;0. bd U(I 4r). 4- *'U -4 Q'\ 11j) W,/-4 to4 q-- .J i, -, -4 ,

s-. IV maC- ,a ' ~ -. V - I 0CJC- 44 (Lo .0-s ft ft .2 -41t .1:.U i

.0 w, cu- 0--. II \ I -'*q* 0(U (u .. ' il .-. i- i E~~k.- in4 .a~. hE H-I'-il. U 0-

toO (41o4EI G.- 0 Vi . .u5.00 5'

II 5. .- 4 ~ n- O.. 92 .-

0u

'341

A AoU

/N C)

W4 0. 4.)

IA- C, Po EU

(U ) -4.- Q1)

..- Ec P4-

W4 .

%-4 0)r40 4 I ( .

In- = ) u -4 q)4.

0)P )4 %A 4 a.r4C44 ,

4- ) VA- AdU- Ur 4 -V). 4.) 44s

in. 4 -I (4t 0a#(v0

ku 6' 6-4 i

v) 5-. Lnt ) Ato -- got ir 00

".4 Fd \/ I 93

4Art

4.4 (1)

o 4.o

(a a)6-

o 0)) ~ .44E L -E

id4

o Qtoa.) Cn- r-45 4fl

d ~ -4 .

(4 a)N It d)t- - 4 - t

4 4. j -

o~ ~ &4 0. .4te. v

-- w Hj4bV- .(j 44a0E

llC 0 1 dW*- -w 0 IL.4Q -4 l.

ii 1 001 1-11r4 4

.. 0 in -r4' 'co 0) 3144V u 4&

"-4~4 ".4 to '4

u-4 - -4 ~ "4 E jcuc

I ~ ~ ~ ~ ~ L pgI c Pi I - '

~s..aj S-c94-- \I

0 .

Cu r4

(UU

PA.u

E-E

,-r4 r-4PA~ W4 -

4) 4 %-

vi (U CU 9-4 - 0 . to

X. '-4 toU)V

o) 'o 344 c:L -

U, 0

to to-

Cu Cu

AD-Ri50 558 A DESIGN ANALYSIS AND IMPLEMENTATION OF A USER-FRIENDLY 2/2INTERFACE FOR THE UNIX OPERATING SYSTEM(U) NAVALPOSTGRADUATE SCHOOL MONTEREY CA F E GROENERT JUN 84

UNCLASSIFIED F/G 9/2 N

U.0

L2.2

-IjL2 L-

I I II .11 1.6

MICROOPY R STO 3ETCHRNAIONL BRAUO STNA- 1%

% 3

11111M~ M%

1111 .-% ,

w4 TI U0.0 ca.0 )

116.C 1to 4 1 * -

too. 44w4 0 '

94 0141 UN Ln

4-11 S3a

ILI 40~ 0 pi 0 CI .5

5.4 0 4.h.6 in 102 (A * * *'

.. a41~~r I4 CDMaa

6401'~~~10Y'4 0 ti)q.~ .C) a

0 C* *4 to a 0.0 icto -

W4 6,U 4 0 r" - p- -~ w a a

'- dj'o a W4 .0e41 ) ) %A 0

40 03.4 ;30 4oa CA .. ~ *o

06 ai a11 a IAA ' 4

r-4 W' 41-010 r.-4 3 0 (LI .4~ 21a-e-- So a

W4 t) 944) 0 ,14 '4,.4. F-4 .0 M~**O~D ''$* IL U'.4 IL 6%A 40' *a ) 'sie o g o-

YJ wJ.4 'Q La OW4 03 Ui il -ra 0U M I -W,-4~(U1 W4. upv (Lo *0 -14 4 1 WGE 4.01~ I -..-. f I

4.Q r..04 '00 4 I .

~4 4 ) 4 4 ~ 6. 4. M A A.' t-p I I -r.qi-

00 W4) Poll P4' * - 4Mr fr 00q4

h. 041 1- 140 S. *0 "N -4 "44 4, W44-410

cUw 0,~a U-' 0 64 016 0 64646464b4W'l0

-4 r14 L64 S3~ COO.0 4~ %Afl

04 ..S An4 .5' 4..4 An P4 4. a 'P0044Dp

O4.0S.hU 0 0l 0 00000 40 a000

E- -.4. 4q61 r-4~ P 4 4 -4 r4 04 0-4 4 0-4.-44" qt64 q 1111 4 1-4 4 44

OUUUU0 00 V

6- 4444-- 4--4 44.0 44.*a4.0 I4. 04 4P 4a

96

- - ' v.-

* --~.w. . .

oil - . -

e* w- 4)

sk t0&

E-' too .q-cV 0-4

01C 113 l * *14jU"C);1 Nr-Z

14 A. I * 1 Q*0 ci. C) iQC)

&0 ~ ~ ~ ~ I " C. U 54 4- - a "4&404 , 44D4V0

ko -P4 V) tI il gfn %A0 4 ' *eU )V i aW

. .I . .i .I .I .l .I 0. .- .. .. . .' .O . . .C . .. ..

7. - 77.

14-

. PI C. *a* 0 .l0 3 V9 *0 *0- Ia .or Va* ul*. * o. o*0.* 0.. o*

cI II ii n i ril V) c' ul V) ii ii 113 tt V-4 CQ iij V) Zi E EIH H H H f Ii

I. XXx n UwU)U cnU 02~ #:5.4 v- c')a. i rv4C\ ) Q

WC45 vi (n Wk'1 E ..E. -4E w c - t- c rLI3 Lc4b 0cU cfUtv 0

ti0 0 0 ~--I C4 I- - E-i-E I I 'ClV X IL 1: P4 oX-2 . 4b4wmw wuL*&nE U)) In .10 4 0 I p b4 m b4 10 a Mo ;M. I0uwtn P4~ 1 u~u 1a0 0

Cnan* if 111Z 1:0 !W- ZCL P)- V-40~ 4 ~ 04O Al 1 'lW- 0 00000 00

A; 4 a$ zI~ 4a co q (4 )( c nu dc .4 = 0-3 v4p44..4 . 4.444 ::w4. Mot.A..4.04 04 .4.A

u L .4..Fee44"..4P .44 .4-. ....4 s- - c iu 04A .4 .4 4 M E 3 E u.4.u.4.4.) .44U

* ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ I Q1b~ c a..) 0 .J vb. ~ ~ 4d~ ~J.b ~ J 4 D~ ~-r 4-4w 40 fr 4W - 4w v P 4W W 4-4' 4r 4V 4W 4V4- 4W 4V4

98

A. A A, *

4-b-

to * 4

v 4 ,4 '1

*V N' .4 4 I4) w~.. I .4 M "'

* 0 go ob -* -4.e5- v 9.4R

0 k11q 14..

N' 0w *4). fa &

to 0 4 44 .to

4)~~N A4X1 3U Ec

0' %-' 21 U. .. tn0 0. 's 0b

0 0 (U 0 0 Id> 9 -

* 0f i - 4 0 j S. l 13 ( N

-V% C) 6-VU *14 N 41 4) U

-1 41M 41 Q E L 0E r4 .. -4 (JM d1~i..' V W %24 0l e. co- nfa lE iAu-'~~ mj *4-4 w a=k .- v - r 140 V P .

*. 0 4CU vs 4~4 *4 f-4 Ci -a 4141(U 0 w. to 41 4% 04Lff

, Q& rI-41 44% so a =4 -.) -r4 .W-4f'l x 0o " 4OU L4 *U 0 5.. V4 PU 4 rqviW t $

-" 'JC .O f 1 -W ,10 0 * a - 4 %.- % O "* -41 0 T.. 20 ". 144 44 C. 044. A Q%* L W4 4'.. "4 m P 0 ) ) 4 %q W. wei "4 pq'

r-41 .. 4 04 r04 0 4. -U. S-~fl QJ0-40 w W-4) 1~ 5

* *IOO~f--r441 ki =- 41.P4 14

to ~ w41 -11U C) U ~U.k'f.4 (. .la 4

* ~ A ~O% 41 41 4 141 4199

iI~

3 OU

An-I

al w

at g to.

0~- " 0 -44O=0.

0 *J1 0 (U )

-r-4 .4j4-e

a GJ'. (-

RD E A3Q.d)(Uo

. -4 :1 0 t4it0 -V"4( "r4 r

In S.. 0' ba00

f4l 10 t of6 0 4r o %-r 20 0-1 1s . tO tn-~ P40 0

o4rI0a, A A- 42 4) in G #' q3

-4 - Al l u4 cudiP

~~- -~.4 Ad d o03 p3 Q) .0 4# % PI4

00 .5- *a~ *a 64 *af'. ~ .

ul %A ul V, A V0 ow v4 1'-c- - , 0

* a .-4 100 I 4I

"

.

-I

464 *,a

N '. ....

.) 60 6

tnn

9A 6$a A3 (6 tn "-

.Cd Enq

4.p (L Wf o4-3a

E * "Oj '-

4) ) :4"e . 4 o" i%-

rA S. r.) V4 . ,4 A 90

i 0 : = v -aC -

4) dk mCU#0C!W

o .t 92 . E4 pa

4) 1..9~ r4- o t t nof4S

it.. .... U -V .1 .-

?~~~36 E40EnIGJU *

a) . A o7) .- if X AUl. 0 '+ N. ." ""

MR: 4f;V)do itdo (I) C) pqA si t) 4N

t) (L' u 'u %nr4c-M4

o 4 -D.=o M ,,,. rv W .-.. ,-&4. N lot C!

0.' cu S. in ~ (LO 4. 0

0 f .t- 4wJ. . ,= - 4, *a- rf "w

a)) S.~ 4) a, s vi o - En ' .4t,

VinF .% 5 -U .4 E %r4I vi . (

• .. D -- t En - * ' 0., ' €.C E., 1.'..- ,-t , OE-n 0 ,--' -. ...- " ,,,,*, GJ.- . .E-4 E• Q *--* ," - 1

• 0 ,, S.e-aJ . . d f, Oj -.F '4 It.- . ' 4'- .- . •I---

1 o g. w C/ .. v i*.., o) * ,4

ooo.-4 K 5D ., ,C .-. # (I ' , U. I A'".'-

C) .. -- C.'.0CJ ) 1 .... ,o-4t .

LI E 0 "*1- * "'"

101 1. -43EI '- O'-4 I5s-. ..

4. 4.) """~N

-..

'% o,

% % % .% • % " . .* , . . % • ".. -i"0. . -. . % %- -%- - . , .. . . . .101. . " .

L; - e % '*Z'.', _*'', ". .%.. "= %-.- '. "!". ", "Z. P% "J ""'° ""•",-, ". -*-% %-% ",."

ol

0 00

W ul10go 0 0. P4 400 0 w

0 0 4-A0 J 4W404 .4 k..W

0 O - 4 40 1c wIm

=~ . 0.4 rI0b4 -0ki E- W40 ifC) 4r0.~a1(

-) U0.1 MR .504 t UP Of WFO 04W 4 0- 1m. 09 na 04',~4~dm

40 t .;O Z U r EI P4 M

p. Gj o =) w- 0.-0- -40 W

:3 0 0.1 Ma p.r 4 4 4 4 V4

.4 p r 4

.4 W4 -1-.4

0.1 IJ F0-'4 4. 4 4 4 -4 4 44I .94 r-4 r-4 .94

0.1 ' 0 . 4 *O .4 w 4 4r4 w~4v4 4 -4 w4 w4

-. 4 w4 W 4

04 (U 91 QAC0 %

to Own0'~I 94- -4E EE E Eq-4W 4- 4

?4m 6)0)0)0)0.to)00b..

IS~4 (n Ad-L14' .)aC 4

*...94 In Go' - 0 I -U4' (Ud Ua W4 -

q- 0j 0 n10 0 " go - 4 - 0 #n W)L.W nL) )% ) A n

cn - -OQ9 ct b I=.1 - Pc - . . ab * j 4.

+ ~ ~ ~ ~~~I In 0..-I i Au L niA

CDL .. -a#4 4a

t) V -- 1~ 0 0 4 >% (U - C, fu ( = a)

in C E

0

1-4 In 0- W -n 16Ia '04 *a CA W 404 W~ tn 0 - (0 *C (3 *a W a In W) W 13 c..

En ~ ~ 0.4 =2-1 ul W~'( 0UaCla~ Oft -Q - )04 4 0 M nb4 ( n0c n W P14 0 to644o5n Am 0 " .4 r-ar4 = c- P 1 U) GJ0.4.14 Ca1.aru404 0

4 ~P4IhI(0.44( I * r& )-4 V) U~t l 04 q 4a r44~ ( U N49aN nt

.4 -C X: 1"4 )(ac1: 0 * a E4 . 4 E- a: K

*U) 04 -4 :m- 4 *Ei"~' u fl cu aC.o t P 4 u nN a 04 0-C E4 ro rU A4P

- n .;z CJ i E- ocn -m.p do :% E ~ *- I - t -

* .. 4 :ZQI-~4 r4 vi. fl W=q 0=. P4 u 4.4.4% N.4 .4 1=. U4.44.4W.4Inmz , 0.p 4 -4 c -i. w V nc IA.44 4 n4i.4

rt 4 4 .4 4 4 w4 4 %4 .w4 r4 w4. 4 4444.4 w4 W4.4W4.' 4 4444 4r- r- 4 - 4F4r404 - - 4" - - P4P4r4r4P4fI- - 4- - qr -

5wS I4 -4 I4 I4 -4 I# w4 w4 36- I4 4 r 4- 4w 4-4-o 4-4p

.Abj'Aj DA N0A l 0A 2A "p 4.3J4 fu 06- o A.4j Jbaa AN4.4.. 3*40 0

bp P. :p.0pDo.................. P

WGJ~i~U'- I.~ off,, s. IW J J j ii 604* a i i i f U .6A4- 01 ni4- fa -0in69410 b A 06Oan 0# . - . 0 4.0 4-P O innnnnii u 0.

cJ C:U' U'

4) 0) 'U( u4 )c #0 )Q u )wQ )c ) )w a )aE E E i9 E E EEEEEF-EEEE EEE

103

4%

En 0 N. j4

E- 0 -E L .

A- w) - W MZA 11 ( E i n m ;

w w ma wf 1.4 Do . l- 04 M0 C4 Cn 1 =. =a W" . *-da Ecn-at *

3C .-f) rA C3 0 -4 P4CIj N u- m. u~C m-(l. a

5 ~ ~ 8I~.4 ** :P >1~- -- ft ;p b:b;tt%> 4> 16

"4 w w~4 v4 lfl * - 'C *e= 4 4 -4 . a4"" .4 W . rgk P.4 -r -r4Ci w o-4q -4r 4 r-4 * -4 -4 F-4 '* -4 E - - - f 4 'U r- . . -* IrIr r4

W4 *.4 -- 4 4 .0 .0W4 * - .4 -4 0q qk Vq- -" W4 -q 4 .AD 0 .r A) ~ c. A- O A& P.- *E- A2 A -PA80. QAj j Qg Q w. 0 .0 10A2n-0 .E -V4.i-q Z# W4.04 'UZ W4 4 W4V4 k W4J W4 W. ".4 W- 6-W. -u *z W4 4 =

*- -4- 4 - W- - 4 -- -4 - 4 4 - -4W . w -4 wr. -" 4 -4 4-4

4- .0 4-a 1- *ab 4. -a* *j . 4a +m OI4. & .a o 4. a0.40& 2 0

a -4 4 44-D -- 4- 4 4 4 O -D 4 .e v *a 4).4..

-4 -4 C4 -4 -40 ad -4 -4 -4 -4 . -4 G4 -4 -444)W(a)C 6G F--. 0- -- -&0 -4 a)a)a www'kq4 q. 04~. tnv i nv L ) r n 4oo l i nw 4 oAa 0 v u 4

Q ci

GJiG ODV LGJs ia iJO OGGJOJJ G.-I~ ~3 JGq104U~ GG

IVV

or. 0. U')t ;

E-4 0 -0 P4- AC4 3 ) -4. P44 s ca- t 2 c 414

4Pa o.-4U1=1 E-- '4 1-.uP% PA E-.' A FCI *.A E-

WE E. -0 ~a, - 0 E * I AC.- E%-4 -0 "1 -E- 4' acIC24 t..fr4 !M- P4 -r4 .34 (n L A4 b-4 CM Q r1 'A m- Pi 4~

;E-r~J fn E-c d IEaT' cn zP n -4 1 E-akr- C ' X Qfr 0-0 QZ ,0~ * ;A P14 .2~ - 4 'U 4 - UI P$A :::) Z ) - 0- V4CO0 ; wo=)=u "-!f ZJ4COa) o = -f1 CAAn= a -4L QC

W.4 C4.3 64 E 1 K .- 3 P14 M f&b 4)J4 i& -c -AC *4 U04 . 4 .N .).

00 Wa 4-a *A #-a -D .4 +-p4.0 A44 4a'a *AW .

.4 .4. - P4 .q4 4 -4 q -04 -r4-4 .4 -4 4 r N .4, 44 " 44.

-4 w444444444 .4.w4 -4.4.4.4--.4.4-4. 4 .4 .. 44.4.-4 .4

I I=* I = 1c = fi r. C: Q *3 r. I 3

F kEEE E E ESEhEE E EE E EEE E

'0 to:0di9 4 U0)4 C)a 41 ( (L w ) W.)a4 W w 4 . 4 A . QJ * b 4 w 4 . w

Vd) Q wi U 0) vt 1 0 0 i 4n .n 0 xi v) o / v) ) en Li v. vfl v) A. in t n o ) viv #A i kn

a-441'-

0.kG '

los

77 '.-

*PAC

un op .0U

imoiIll. S..

CA 1.QJ #1 A 0 f W . rtA ..a1 ,

910 .4 Cut ci' -1 fn C4 ~ 0 1 0L 4-

0)6_ ra . MI.l C> = = -b

~~tr ix usO- n

04 -4 ~ l~fi *faXe 0 M f ir;1. (n 44 ;C CA 04t-0 C w )u n

a An z ." -8- - 0 4 WN cn l *l N.ra.Q 'U 184 FR E.fl 1: P4 kQ i -.~4 )4acM

C-4c toC 0 oeM -N N/ P4. 4 b3:bb>11 >nZ ib 41 :09 > 1.6 6.4 inicn n r

4 0- r4 4 4 4 -1 q 4 -4r_ 4 E00o 0.' c- cn m~~ rz t3-s

W4 r - enf It in 04 P4 - X fr4.44 4 4 4 .4 .4 4 .4 4 0t sib ul wr cnww

-4 4 -4 .4..4 .04 .4 4 -. 4 - -.. 4 0 to tb 0 %Af " / = d=im C4 =(

4-ae O.A *0 i/ -P b enesi Mne 02~ .- =-.') cr .4 0 >b >1> 1 -Cic := r = C q% a; ~ 4.1 4- *0 4-, 4. lb

e.. .4444.41.4.. EJa FVJ42.G E * _ EEEV4 _ _O4 O

"aa GJG0 Va*a e"Gta .4 , .- W4 4 *U -1

.oa 4 ) )( (UO s 0)'bDa 4) C" .W4 -r1s .a4.4.4.4. 4IQ

III _.~ I/b i/b sV1 0/ 111 en V) i0 kn UIM 0' ii I 4-4 -4 4 44 44

:( cl ~--.-- .4 en W) en n en n e

ene Q %A Wi V) %, n Ln Ul en en %0. A.. k54 to in V, op en 0 0 . i I .

en a.. 0. 9 s &.9 . .a w 1 0wa

fu a, 0.C0-4 -4 -4 -3N nnonennnvi

Q) 4.4a.' (Ua a' 0)IL w 4AUel.4 o o sitenn one .41

106

%.

I.;o P4'. -

-d ad (- .4 I. 1D

-n C44 15*

*. ra al N N 4W.

1-4 r-4 W4 Ge-4 1-4 -4 . tn r

I ~ ~ ~ ~ ~ L >4f aSa* a.a

C,- C4 - -) .-r4W 0M*

-4 .. .4 -4 -4 -i Z -qta *- 11 a

U1

Q) 4/ ) r4 #jmP40 q

.,u Lo 4 -4). *1 EC - 41 ba 44 0.0

-U (LJn %i0 0 l% A0VI4 I-i. PQL = C= = = ..

o ) (U 4))) ) :V V))Q 0 4. U U U U

cu V) di QJ c, V) a)0 (U %n W 1 Q) Wu ui CA Wj~ F- 0. -- 4

V) o lA u An 44%

Ln ) U*)

41) )

107-

* .% %

- -' -*** *J -.~ b . " 2 7 - -o -. "_'-. .- ." r--r .- ,

tn 5- m = . -

#4' pz, re

;v454ra

a a a * a ";-5." 5

- -o 4 W

- . 5 5-* ~ ~ ~ t W V/) Mr u fl Cl5hC c#

X: ; T- 1 u r- 1 V- 1 2: I ! -

- E- - - :? - '4 - E-

S-4 -. - , 4 . .4 04 .pa4 .4 4 4 CIO

E-, .>,:, E- E- E- E i4 E E -4 E E--/-'

• *., ..-- u UD .4 "4. 1 "r 4 5. 5

r s-45. 5-4s"

4""

;Z 2! =, =I m.== P

, ..:-aJ .~3,. ~ ..

* II5 -4 I=

w4 -4 . v4

14VUiU2

r- *- r- 4 r- 4

cn• 0° -.' Q.

rw~ AB B &a A2 E!0 1"

-. -44::.. :Ds D1 r

.n on 0n 0 6

cr4 QM.10 -4 s.G 1'3 4@ 4-a4-a 4 3~G 4G

-"::::

00 .. 0 0: fu I I ot

I'

....... . "-"/",,-..," "- *a", -" " " ,"" ' .

4-- 4. .4 4.. . 4 4

cn 4.

-dC E cn0(

E- Ci , .

)-4 ;P4

E-i E-4 >4 )(a

rAflAntpfA ra .4 "4 -i .1

f . 0 *n 9/) 0

I= E-- C13o-fr ~ 4 -44 0-ik4 cn04adt

r4 M4 co -. n 4) 40 .a 0

W a ~ k4 :lb-a

.a4 "a IJ 11 tp~ pP in l4=4 C- -4-

-W %AW W4 4 i-fif IV-4 r- - -

4.Ai. ".a wo q.. 4) w9--4 a IV'

14 pt, ad m-v aP!w 42 O_ aQj a Q - 11 F- #A vi (u *., o 64.

t.)kQ J .1 V)..ai AJ ~ m' F- F- to4.q4 LA4V, A

vain~ ~ c4 --.4-r W-4 'aD 4*~~~~~~~~( 0~'~~~-(j~ 4 ...

4 I JI

E, E Ps' Ul F- Ir 0QQ clV ot

C~( cu 1 4 4 41VVn'00 JaUT-DI 4 I

-~a a -4 * a

0) a) ( cu 0 f )in lfl in'UU Ln0

109 -

U, 0

%n 1to Q)

9G.

ubA

m' Infl 1

to U2 M atinn

C-1 *'0 1. U 0 "-

w * 1.4 = a 1 1- 04 -cNIf E e-a-I d .*$ **lb a- 1.4 9

4) -4 ra -3 E :1 * 104 LIrzzE * a I C wWja -4W

-4 W m c; ) .1r. 02 4.# a V

I~ I . -MR PA ;14 04o a)". w 0 - 3

be-) i) wn ;vQ 4a .0 ~ 4noP *aU) o 0tF - E "" WU =). t*4.. a rl (I V4 W-4 . *-O

V 0 rL 0) W4 r4 -4 q '-f U- -4(L. t t.4D :C- Sn o IIF4r1 "I If W4 ) W--"

4A4 '.4 OP Wl~o4 k a 0 i V) Q U) 0 .9CV tr Gh 0-11 tu) v -, E rm V) 10 V4' ~ -

0 aio. to -1-LO C.4I Wb

flgfl A q3 'a )D v C~IU~, VU'- _4~ -uw

a a -0 46 -o- w4 E E .r .h)I W fl*.W4 U4 .hONivd

4.D~I W~ 42j ..4 04~*E G--- W 0) 41 ---. 'e

%4 %n %nof~ wi vi vs ~ 104%

.fon J j, j ..04f 1). OC1 1 1 1 -4 -~' 1 ~

4n.r4 f' o4S L4 %A %n %nrbw i ln

Gj~ja) a)0.. a 4 M a

efl~fln ~110

W ~ ~ C *)(U

Z: 04 '.0 'U V0

W 4 *a *a sa It wa..0- -4 " -3 -.

ItU 00U04 40

4iJ P 0) -14 f~f r. weE 24 %n #A pa W4

4 4 (U E- I ia ox Zt. 4 C4 -r

L"- ..4 P-4 E- a - -i -a u4n:3 0-4~n a l Q)~ * ici, ti. t-;% %>b-4

t.8 4.A. --- - 4-aU' " a-

- ~~ Z-e- f4 r-4 4 -4 U f -.7VC04 ) v #A4. a)4 W4 -4 J41 .4 ~ -

-4-4.. w ( t W- - E a. C) do 44 Cv A1 .410

V4 OA.. 11qJ .~ r4- 04 4-.

4.1 0. 41 4j Go* 'O 4 4) 0 14 0 . .- -e4 4 46 (U

0 A %4 :0 WIIj i ( Wr_. ton .

FO K)'w LO) f a-4 110 *'a 4.4 4-A b&p a i . tf qV toT' to ,f (-*- 4.3 pO~ . Q 04 4) vsa III0aw*

U. V. C. 1'4£ E.-'wN NO Ni L'- E30 C) 0C

N %4 %A 0 0U~

-4

LO C1 E- 911

toll~~ cn(AtnO

'9 494 4 Cj%

>4~% P" P4-

E- E- E- E- A %A C 1 4,4 dWI I I P. #Ain 9 tn Ww w

>~ .1 a .0 0 c -,r4 - P%

LL rq W- 4 -c- 4 (q-4. P4 pa'C4 J

#~~~~~ 4 0 'A u t t 4 c 4 4 .4 4003W )4 i 14A 4 W4) C rj

P4 129. 4. 0A "3 0 4...w 1. to

(L L 4 P l 4*jt o@ 00 00 l I i a)S E E A IA 1 S.4S 4 C1 0 -2 4-4-..

40--4 -lo- v "D %n 4 W.-0 -- --%9unI % X

&I ) -4 0) L M- w '0 4 6 ED ECE I 0CI "-4 -4-49A-C4

I Uti I to r- to to4 =1 kI k3 to,\ I3 to Vtql

to E) Mo %A- 4. 15=.4 t F o.vgot

W- G) W 4) 4 84)44)3 3n 4~ . 4A. %A -r7

% .~ . . .. .

0

Us :3

. s 0 *

e. -1 u 31 "a.Pra ~ 3E- ~ fl 0.NEf

P4 V44' W.3 ** .C. r

* .s a Pill %b4 4o o ~ o 'T. w* Ic w :

E E E E Ea 11 ;!a P4 U3 w 00 00

4.15It 4.14 0)0* sic-. gig.4

>4 44 0 :P4 OP 0 -" W V-i .44 4--- - 0ll wj 4) a) 4)UJ J I wO 4011 w ~ -4 -4 I"

- .454 4 .- =E 1- *3 ,'D o v 040~ .4.-Cd 4 .54

A: I-4. to uaa tosn tl0 o -4 jrv ...4 W I a A.4 -4 .4 CL)Qo 'Ill Ul )V l ) l s W' .4s .4.

o I - I o In Uri %n An 4 4-~ .vi ..7!S4 ..a GJ C.) (L 46 G4 410E .4.4 .. r- !i% 4.1 '*%5

0 51 -4 .4 .4.W4

e EE E 3 0 11 oi-4 W O 40 d) CC C- 0 W-0 $ -.4 E E Eo 40 wl 0-4 $3 V 0 V* 4CV"Ul nln v 4

toI -4 Ito 0 **t 0 N, 1 1- a a a4 ,Vi I% I t to vul*5 0) 5 QO 96. 12C n943 .0 C :a 4.8 60 4--a Ad4.Q 051

UI0 Us.L. C54 %AL Aj r. Q -4 - "r - -4 4- -%, W 0) W) (4u %A s v MA UsO v /i 00Aoa) W -0 04"0000000 E L hE'u m m0o0ot

n toUlt C) U

0)Q W 0)03-1In Ul ln In0000i

-4

113

%

0-2 .4 IA 1. :3

04

*4P4W '4 Iw1- '-- r - C- Vol W .. E*4 cnc4 -*r -4 fn-4 cd to s

* - -~. 133

0 P4 04

cu E- V

00 ( 00 a0 00 00 Q -I 0 0

I-f t toL X Io J) I I do--- q- I I% to r-4 to 0 P

42*041 *4a d4# #4.p. C 4'. &A. a-& -E *a4 O'U 14 0 I

w '.4 I- Go4 .Ua "4)WW04 .jW4) 4 0)40 .1 of I. V 0 0 E%n InIn. A.0.0 % A2. IA.Q VII AW.A4. - 0 A2 E4 oNP A

to4.. to. to .. .4u IV .- 4 4). UCj =0 ~

_:: -4 *.4.aj .4.4 A3. oa -.4 .4 - 4

to In 04 ' % 30@ 0

OJO AGI IG Ij Q Q *~ 0414.

A Jr. M W

cj U *1

fa-

-0

*~ o )

0 0k W Q0 I U

to E U)a 0 . 0..I .bi v0 .0W a *aMCZ

0 F- sf 1 IQ

$A N 0 0 " -4

4n ;h Wa 4) ad . .

P4 00-- 92 0 62 0--C - rW 1%4 w4 Q- tGI an 4U'i 1 W6 4 V ( CYaI. -

. 0 0 w I. 0= I ~ la t- o to I

go 0 in 04~ 4*~ .4 o-44a * .4

EU 04 04 wn t- m~ : 1- 4)Aa

kI E- 0 4 0 04U94%n6 so~' 4) Ia.. w & I~ 4 0 4- Go . j .*

0 o0 *a wi .4 .. * 0 W 0 w 0 - ) 0

0 ina Un E n 0 EI 0 .0 F* E* E F

-~ ~'.4 S..~ ~ *e * **. U

'! - -; 71

0 4.

Ill

.00

00

C4 A4

0u *. 00G & .1 MIP i4

t,>4) 4 4 $ 0. 094 64 .,in 41&0 1 --

0 vi0

3k Ag W4 En- a. *~l *Evaja

(U -44 di Q) ~*j 0w %.0 01 u *~ do A2*s

e;. ) %.4 1".- 0 sm 04 o -i Mtn-OV Ld 0a l O .. . 0. W4. to 0) .O2~4 #f)

le ;p An Ln E. 1, - V4. a 10 tQ. 5.4

i'. ea ta, s4 %01C 2 0 0-V .0& V 4)O A ~ co 044. o-s NC~ QCQ -f0 4 - M -4N -

xo v C o9 23 A* t3%~'. 14 E O 1o. 1. 1 :0 01

V f-4 ab . 01 or.. om %A v4~N uf 4 m.LWam0 "0.0 w~ 0 wCot w0.- i 5.4 a) % 00.-. to gUA:~ S...~ .* - -ot* 0 64 a1 -a A 4 $ 4~-a 1 4 z 'Af

4-P 01 434 4) 0, I .. _ in P0 N4 C3p 40 to0 .& to 4- 6

17 ;C " D N0 00 4.&.0 -0" in W 0 a0 ai )a a Q

5.4 4 cc) 4 4 0 04 we .4 01 .. 4 94 E~. 0 Es %n

F 46 4 - Ev'U %n v/i0 U

0,

116

Ia.

03000

A3

44b.

0.

0 &

-,4)

to! ? oE4 9 nE44 A

Ah (nA;. A2 A3 v-4A4.)

0 .to - 0 to41 v IV m 0 F1O.A -a * O- (U .& *Abp -D 44 * 4-f = U

'A F8 bm I-- )' 4039u3' t. t0 >=4)

Wa C)~ &s. * - so 40

0 -.- f - 4--1- 1 40 -- 0 1I - 0 - 1 -a t *A it 0 wq v c

0) ( U 3a 0 Vj 1 4 1" * a) -q 4

4 v4)) S-.4 03£4 )O )q -4) 4 4) I .414 03

A2 ss oa A A '2 4 q o .

£4: . 0. u£4: ul '4o

'.4 '.4 44 .034 -re -4 w'.4 - 4 4 #~GJ a £.U

* ~ ~ ~ ~ m Ia to* *a e - 9*a* -* * W ' 3.-4 AW -4 M- -4 W #a i4) UI

*w u af *r- U ~ Q 40 03 421 ) ) ci *. t44 *iUWL G ' ~ 00 4.)

lb6 1 s 00 0 *0 0.0 ob- *a£4- . P p . o*..if k3

~ 4117

C...-

*a 4's 4.

ro"%-

E ia). .... A.. . .. .... i i

o )) U C 4 4 W %-

'4 6. J 0 • . . .

w 0r 00 0 m q s 0i

o___ a ~ A .. aU 4a to -* vJ -*a it Ca ) )it * -o(

CI - qN 0, -#- V 4 N -4 N -4 C) I . 4 a \I -4 w4 0 -\

) *1 A 1 0 o # f I • "N , iM 1 IA I -*4 I I I ... .0 0 4A 0 X F-4U *a 1-4 4 % - Wa 0 -- 6 M "'"A * s-

- U 04 b4 4) c . - Z c. Q) , .m -UC -4 . = ... M4U

o CU ' ,.0J 10 al 144 W. g4 4a" 4 '0 4 w I I*& -,

4.10 E v, 4- 6 U . E u) An *. 6 ii .' = v,. j 4 E nI. E i ev 4-0 0 rE e .

.1*.4 .o

118

,,I " , " t ° ," e, ' . . " . . " . - . - . ' ,'." , " ,* . - . ' o " . " . - . . . • % , , .

-4 W4

o 'in

W4 ~ 14

0 4a

4b4.p

-f

* 10

1 PAO

4.0.1,-i

w44

C,, 0r

o 4-4.9 Co.. 4)4-4f". L (U0 0

C3 93 a -

F- i__ W 4)4

63 .- -.- AC Go, 4. -dwr4M 0,r4A u - 4aW 04 U- 1=4 PEW0pG . 0NU a 4I ut

Go to -r. 4 "-4 A4- 4) 4 )& I( vF

(V I - 4 a (9a I a1UtP94 m 4- ;w *0 ps :m 0-0 04 op0aNb,#ap

w )Wwap 0 0) 0~ a) a)0a *04 jc )0 W,. 4)'En in- kn *a- I- %4aE A St *a E A i % 4- I %

~ 14 0~ 4 ~ i ~ *0 ~ -i ~1 43

01 01Utfl ~ ,-4 * 119 U ~ , 4' r.0

4. w V k*42 %

4 *1 %A &aD %

4- 0 ILA Ili -4 I.0. 6 6. 710

S. 0; 4. w U~.o, 0 M4 AU -

(U~ 4,4 I E ' .

)4) ILI 0. :104

.0 '4 -4 in

4' V V '-

0

'*4~. 0. 0344) 0 tv 4.' to 6 .

E- :3 0.) E 4, a4A 4.V) W S.. r-4

)4 4' . 00 E-,

4- .14 as 4.p.41=' 0m r4 E .(0 U

4) to -4 Eun E-4 0 .

L. 0, o m) 4, F Q; 00.. U E 4' .WU 4 .

I 0 -4. 0 r- &P- -l4 4 S 4r \ 0 - .a - 5-V. ) .t . 0 &a .1 .03 . .4 0 (k .1 . , = IL

w 16*!IEG a 6 m *) 0- = a/ 4, w, 6A CAI -00E- . -4 4 .. *C ata4, r0 -L O0P

to &4 E- n %4o o) .l -' -6 04-D MJ a#. t ~ 4 C~ l I L 04 i-I 'V j 5.V do'G.. 4) 1 k n 14) l.. -4 IV f' I U 1 1 %

a -C 'I0 r- 0 0 r- 4.' "9:' 0 4, - - C 4w - -Aa0 o W0 0 *j *4 Wo 4) Wn 4)& . od V L 4 -.

P4 a t 04% 9' .. (4.-. OW: S. 1" S.. 060 9 ~ E-

4-'U .j:.m: 4 -. 4p 14: m: 0 . C 4 DoP4b 'o6Ik0 *wwo owwo 04 0 0 1& 0 U 0 4.' Q

E *C) rm EU E. E4. Um4 -EF

-~ a ~ .' 4"~U C120E

a) 0

-b 0) 0 ~ 4a

4-;'.u- .-

(U W1 ( U

4. *0(3 5.

Pt. W) d (1) 4.' b

0 41-A4 43 .

_O w

Il l 4-P -'0

E- 4-' 5. (3' N

(1) cu 01 43)

-0 03 cc 0)) 4' (0 -

Idl r. ) 0r4

Qj -0 *;f Q

P. P -L -) a)- 4-' -\ E O D

'a Inq 0 4.4 -4 4. 3441.c.(P34 b4A adr4 4C

+a 'lV4e - .r-4&.. - ;4 rIn tv 5.4oil 0 4.6 0 v% Ill 0

r- " . . -r4 . 0 V;00 -0 & jA -

m Idt s PI)& % )( 9 t s .;. 0 >,

0 -~ 0 C- 0 oAam ) 4 nN04 4 NQ 0v :3 N. -0 CY

vi 4j 1A I0) sk 11v 15. : 1( 0 t-' E4 s. IAA

0 d 1-4 0) A - - o k - - -- 4 P- - i$3k .

~A2 .14 w 1 En w 0 -0 w a 4 )E )

"-4 to (U4 .aQ . ) 44 . Q)a& * a a) *& 4j M . go 4))

0 SC.)OM WO 0 0 01 w ... 0. w 0GoE- EI -'If -_E

E F

4-'~ ' ' -~ ~ 5. -121-

F-

IL %n.

1.4 AA 3.4*

0 0i U 8 IS'

%. Ul C, . 4

:3 r-4.- D ,

ko '.4 " kn 'a

xi S: (U 4 0 0

4 A ." 0 0 .8

04 ;p -4 IL 'f A

GD 430-0 0 D :

40 .00 IV 64 '

at 4'. 0j j 0 tf)4,0 % 0 -4 to Q) G

0 --- 0.' 39 to to

0 ~ ~ ~ -0 D *0 40 .

.0- 4 -90.. E r .4) 0a ( in4.E a-.: 4A. i - C/) E4 5

to-- C. ~ E 0U. * 4 6J GD 46 -5 0 1. )4 0 4- 04 *

-4 o 5- - 4r4 040r4 r4* - i )+ r - --GD D GD a I.& CJ .. 4 .8 go I/ to ( )f 4 .P

~-4 a k GD - A 0" $4 cm A4 53 4b M4 -4-. 0 CU 0 D go-, 0

GDin 4) Go 0Y.1 ) .YP04 -. n t 0) Ia N' .4 CV

0J is $a I. 4 GDJ a4 4l a 4 U- is. 1G ainf1 0 -49 4 40 0- 3rf4 : 3 0a ma 0.P 1. VI0r40 f"Gr-

Cn4 - 4 24 a oA4 4) G u8 W0 %4-~I.r 4) L ,5P2a 4 . .4

*~~r CU E- 61aU ~ 4(f a

m4 :0 P45 ob *0 04D. m ow.8 04 o0 GD)Mg wP h . o'a 4) 04 O( 04 0 4) 0 D -400 s 0 0 W04

F- E1 F. 4-P E .*- E .0P F-' GD 4- in04 C E a E . *jE0.

0 0 . £4* 43 n e 2* 5 a . 6'4a ' 1224.t *

0 J 0 D I G a6IL iG

A) (U 1

£4 0.

EU "o ~5 ' .(

(U (LO r-4 4 to $u0 = 0 -4. *4

4-a (U (U go4 '0. in ~£40 (U1 Id

(U 0 0 to S..# 4-P 00 (U (U %

0 0 f0. 4) c uo

03p4 In W4 . (

(U E C: (U wl 0£4 (U.0u ~ 4 .0 E

inb 431 a)) U 0 ~ z Lsee-E00 u0. £44* .0 F U

-4 U)' 0 0 D 0

0 aC.G)a l00 .10 0#4 (U T#4 (U4 o,4 0,- *j& in ~ .

sail r. (U .4 -2 '.aM..& o0 "D0O~~d 0) 0) E'U . U ( )

w cl N~ to. -0- r~ 6 0. .I =ClMN )t Eit %4 0)n 110 1- In E- -r. C.) 0

40 ( -6 -40 0 1- %nV o4 0 4 04.'a %n qu.0 ~ ~ F 0.; In utlU)0 0. 4" 0

(U( v (U 4- W 4Aj 4u4P )&.)* ) AWbd P-.-4-I p P, b m 0 o :0v " w 4 :0.m-w " b 04

Q~ ~ 0 0Q)0 0( 0 ~ ~ I(.. (60IV0.U04)Q(

In~ E0 a4n,40 E e e U Ii'4(U(US4U '4 I~. ~ ~ 0.~4(U (U U (U £40

£4: 6.~ . E (U: 4 £: 4: 4:123~,

* ~*.- .--... *--..*.*....*, *,** ***

4. 04 a4

-1 54 4 4) .0 64

t1.e .-44-

LGe &0 %)00

(A) 4) -

W .0 4

A3 f0 4.4t

to 0 to4 lba 4a C". 0)0d) 54 Q 5= . . *

r-4 4) .4 0)4 V) W) 5

Ere CO Q4 -0 E- d

.0 X. 30 (. :

t 4 EU U *0

00 0-0 &4 F- 9- 0

L~ E go log .- - 4P L4 kA .E >

0 4 4)e o40 5 d4-W. 4, v .- % U'h-- A 4,, d)e

I+ r4, a: 4 (0 04 00W tn f C: 0--Z. " ~ 4,e~o .~ to. 5%-k IV 4 vi f

0. tn w, a 0-19&" ) 93 Wa) E 0) a ib&IEM4

4 -4 0 0 .4 A, =4. Z E J.4*fl -4 '0 4, 4*.r -4 -*- m 54

*a CV CU4. L" 6\ ).-4 0) -- C.g4t C 0 ' *

- I V - 4~ -4 .. o - -4 to~f 42 0.V -4 -O4 '4 (.4A el E *- doU Go W5 OA1 P:4 Il.- vu * 0 ) *5

to Qj te pa. 05 4 D. m to m 0. ) p , . 44 ) 0 W 0a Q) o .-- O-410 od.4~ W. -0 WU. 00' W WQ

.4CIJ4,NE 60 F_4 nCi4EC F. f- E ~a

V) tn 4 4~ g 04' 0(~ ~ 4

~O0 4 ef~i~4 'a 0 54 .J~4P.44

uulA ,.4 oaamJ,-G)14 4 f'. l54,0J,

-. in

1-4-

(U ~ 4- b2. 1 - 4rM 4.a to- * .

94 4- 0

*1, r-. o0L .00 c.' F_. #-a 0.IL1i

0. Q O 4.41 (a ..4 "-4 0 S

E~ ~ 3 F_ ' .-E In C;

0 r-4 93 0. (U 08A r4 0

A3 F. *a 0, 111''a4-a 1f &1. in *" 1.4 a-53 - (U -4 4. S(J C)1 0 ~ ;c

4. " ' 5. &. I 1 94Q* u- 0 4.a Cl 04-

v -.4 :3, 93W

W 0. to. 04 .0 0.0.

of 'a o qO~0 V. (fl Q, As to E

r4 0 4- W 5.. 01 1 C DP4 .a .a1 ft &Z - V1 0NI in 0 0

4W A Sf .0 'DG 4. M 4 CD. 0 3 M I-0 04 :s% *. C3 4. nU 41 0. 0. W)

w. 4 w in &n cu o9 0 914.-M4 E o04Mq

*N 13 v t 4.v #a CY c -I44a 0 - . W -4 S.m1 01) 1 in 0L 0 .0 4') 0& 0Ua s

r- - r44 Vf 0- M2 'q P-4 43 r- a) doC la 4) Il J . -.b4) w . ( .. \0 a) (U X aj 4 )p (uc A 0 C 0.

m .w Mn lo- m0 ;w M 0o1t 0.8. 0 .0 0 'l1w 0 a W 1 *W M W W In 0~. W oC W Ua 0 w 44 5 0 .0 S.0 00 01

E-- - EL- E-.% 4-... E Eo0 E1 0 ~ ' in QEt

-'~a ~..ea'-.I... %.. %..0#..'a~.-.~.4414. 0 *In~010

.- 40-4&.-q~-4 a~- 3- .- 12S~1 C N

LL

-4 - - - - - -- d - - 4 I

00 0 we. fI- . a. 4. . * . o 4j

a4 W 1= 40 ( Lo cu 0t C4a0 o q: ( ,M 0a M- 9:M -4W 9 9

I I $A I I go I9 $ t .) i I I I IV 1 6 . 1 4, -

-40 0 &f4Mr40p4 M 0*Pr4A "4 b .0 r- AdCrfA 3IWA

.01 Jt 46 ). MW0. 1UI,4 * L 1 41 U = W I@M9 0' 100 lEa

P- 0- -- -A 04 00 -a 40 - * 0 04 P

N fG 6a 041 to A :1" ! 00 4 P494 lf0 w 4. GO .0 U) 0 40w w wwI 0 ww0 4

rM V I9 *n W h k 0% _W 4Ev *a *_ .4 4-P n S1 a) 4-P F . V) '# F. #n W *n E ~

I-i * 4 * 4~ * 41-- 16 . 0- ~ - '

% %-

4.- .- 44

on VIP"

41 .0 A2 ~~ICAE 0 4

0o to =

--- oA. 0S

0 0-oj $u

.14 4. op *

= to

n~~~0 41 An0 n1 .

WIG %J 0r) 41 9 10go . 0)04 0 v4 us 019

"I v- j Iv. . '"0 0 .4 .4 - to

P, n bM I 2 I Ql

r4 CQ 0 4iEN04 0 N 0;- w CM -4 Cqf P00N g 0 * 61

I* i s $ 1 1 .1 t o o 4 ) 4 1 E4 -4

9- -tc - dto 4 . -4 1- 0 04- 40"- =

00 O 4P & % n w-r '414 U

*6 vF 16 v. 4)% j o1 4) * n cn E. 41 0

" w 0- " .- 0. Dot at ." ,

*w m o w r w . 0 (U., 0 C) 0- 4 1 0 1 0)0 0c

E *0 E44~ 4b 0 *a S 0

* 9 06J. $ ~127

t- *. G))1 f

IV Q1 -.

a Al I

to 4.4L

0 *0

0 kA En E

41 c 0 9-2U

in to to

to 31 6A 0 U0 4-31 'U AdLi*

ri~~ IV4 w 0-4psto to 0

In .0; 0 0D 4A

AI 0 A.* E . 11W, 0 0 14 >lb a~ tn In 54 t

RC to 0 40 04 '0 441:

P4 r.5 41 1 q4 Aa 0- LQ 6&%4!=% % .441%o In 0= toe 0% 0414 - - 0. 0p A- -

#&, In tp W 0 m 41 u 0 cM 9I o M -(4 U) 0 0 I ' U 0 4)-E- . =4 6 . 40

C0U A pq -4 -d laP 0 4.S3 s oM p5

to. t* Li,. C) 'U=i0 0 _P.

Ow 00 o l 4%1 a. 14 '0 a0l = I I4). .a V.-- CY0 S aa00 0- a 0-4410 - qk.# 4 .0 E

4) c ~ ~Jn 14 53 bp ea A lw n4% 4% C 440 )E % 4 qG

.(U I V) Oai 0 . V 41 # 41 0. V 0 a# en OPG 00

E'O 10P-4 14 In 14 E- E.41 U ' E~%~~a ' 0 4~-~P. 4 is"'

*atV 6 I ~~n i 411 a 0 U 4*a6 0*. .- 4*

*' 6 '0 0 1 . '-''- 1- U 0 41- i128'-~

A, -4

C ao S.

-4 4.

o3 IM 00 IL.. =0c) "4 'V u- 0 A

* ..- SI 4.9 (

o ''V4 'V 0 0 k- .

0~~~ iv00'V S

o~ 0W4 0 0 . 43 .9

%n V) 4.3 to 0 0 0 .- 4

enn (U A

.0 Go - 0 0 0 '

r-0 0 %A..31- S4 e > 0 fa

o- 0n 04 04)'& .

o9 to (v = 0a 0 A3

(U 40 40 -- 3 %n-

In: 4) -r4 'V 0

UV S. V '

a) 0. w' cu (uF0 F.:- 0

'-' 01 A -4 04 ' - .'U % 0 0 0 n -4

-4,> + -4 -0 CV r-4 %). r- r-4 04 -~~..4r- &* F-4 43EU

ri. 'V4.0 S. # Do .kcl .'V ..0 . to&0 5.4qj S) &IC e r 4E *A) X . 'V !9 N

%-'-.4 'P 9* lVA FA . 4.3

*at *e 03 A2 0 * a *aqUN *4 N ... , a N w0\ *- 0yla I*& *A to 4) .... -40 :z 0 0.. 1 -4..1

*a-4 -4 wmU' 410 00% WP 14 F-44

.* 0 f-4 44 -4 3t.--4 92@4 - .9-4 P.4

1 *0 * 4i -P4 L"00 l.o 64 &*.. k it-l $4. go ow. &4

boo Do V~ m 14- 04 ON-v w .3.~ p w o-- m .DoK* !1

cu 0~ Q# .0f CU 0-4 r=' 0 9.1 oww E 0 ww 0.-w 0wkiC .4Id4E E EV EQ 0M4,4G40-4 -EAE-E

00' VM. ~ lA '9£ V 4' 1' 01 VE. V.4 3. V f1'

* S..: .4 .M 4 .fl 1 294 S. . 54 1. .~ . -z 5

ILI.0J.3~ %. % 44 V'*'.0~'*G*0.'

-a-- -. 7 a- - , -

00

"4 9. -4V

4 %bB 94

4A V)4jr4

W4 E 0 V 1P4 0 3;( -

C -- 3B 0A 4W b 0. - r4 3 U44* 6 .

4 r-4 %n * o

0 In 4l * -4 54 2"9 4.8 0 go 0r-C 3- -C - ,Iavi 0 vr 4;4o - MW .'w 00 -V- .N l )LJw -6

tv 54 4.8 .9II= 1 4 to 1 4 CV.4A 10C r

14) 4A inl .9 An 54P - - ) . - j4 1 ..) 9'4- (0 -4 0 WOO -N .- -p o 4P )P wc4 4.) Adc *da )4

o )* 0- 0 GJ to4 04 0 4- $ 4-01 IM k

eli 04. FA 4A An * 0 *)G ) - i0 00 40 0 %b 0

4)0 0b41 n 0 EW Q %4-21lf E W 0 V W #0EW4 -4-4W'W

E- iE Q 0 %A0I- In ~ III~P" 9 i a-'~. Q~4 .. .~4~J.

.8~0 G I Pi~e A54-4 ~ 54.P.4 U-4 ui. 4 Gj0.

'4

r4

ffl ~ U *- ~ * 'U *el130.

54 *V 0Htfl M.~.. .-. ~*--~ :.n.

Cool 4fl 6.V

AA C3 %n . .

in 0 i-S

Si(oU (

54~~~~ %4A4 i S ~ 0

co"4 a) P- Q, to -(44 r-4 A-, 0 .A , .

Q) 544 54 4j o &U f 0 C1. 04 5

p ~ C 0 0

In 0 k .54 5

4 Wfl 0 *0

54~~t Si i a t o S

"4 4 ~ "4 N 411 to

Q) adSi ( to I.. ai

(U W- 0 n(4" 'iSi~~ 4 "4do'0 *44 0

k* 0 0 0

04 04 V4.'CU)6 Q) "4 fu" i i US

0 Q4 U~ ;u0.~P

U)~ i i a) S 0 '0 4

(U ~ ~ . a' c. 54 E- S"4O In)0 0 Si * 0~ (44 0o

.4~V 00. 0 ) toS . Si . ~) "U Si 0 0

1A 0 9. C,1 94 t0 .4 ;PW4 4 4. A@5. 4.4I4 + 54*e.0*4 $1 f4 *2 4 0 r-4 Ea q0 %.. 0- 4) to

0i 0 ~ 6a 4. N., 0- - I

P- PO "4 A alr G ovs 0to C-E 41. '4QJ"-4 40 -I L) -- 0j C4 * *( - - - p . &I -

o n ai M." I I3 v 54 ) 0 0 #d I- 6-4 O. I go It

%n- ( ovO "4 .0'4C "' W. 1"E- v P4 -' -- ^-, - ?-~%- IO -# :4-4

a4 to *p * 0)* 410a *A (U *a lb 6A.4 (u . u .a4u 44

in4- W m : w" w t :04 i~)- I* " tb p m to " "Si. $0 ACfli 0I"4Soiw U)GUw Si0Si 4 Siw 0IUCJo

.040 E4~ Ei0 E4 E.. 54 E4 E4 541 r4=4 5: 5

* C

4.4

-4

0 E

0

* 'U

k A.IV 00 d .

5-. 0 u0 9

-w 0 im W ) - ; i#

ov- aa C

(u ;-I (U a*di ap -0 92 p.1 as10- aNIImV)1 -s

(u F- (U *- a 4A tA *a4# #4 10-

0na n o% (U 1.e 4 F.: 9C. LoU - kC) 9

14b 0- " 4.J &0Ci- -o w -O i4-0L . .4) L"h# ) 1- ) A4PW A A I %,. " .*

U ko. to p *.. " w IM D . 14# D :

04 po 04 0 Cw0w 0w 00 o-wm o J

ejC al V) kG 0 E Wa E

*-0 EU %A In

UEUI~L,- d GJ .W~.4.4 ~ J' .. (j..q

C4.d 63 -4 * i ~ ~04.-q132-

.~*. .-

-

..

4. 3 10L - .

0% 4z E3

o~g AM .

o0 0 q5 P" A

-4 0

00 aa .2.

;m0- -v4 E- 4 0

4, ~ ( 0 0

;0 r4 0' 00 a; "0-4 In

a)A*-4>t 04 01 (, 43 E AF- o r4 & . 's s- do E 4a..

a ~ ~ ~ ~ ~ . 04- V) a 4A O -. D-4-0 -icsCUW O 1) %w 1"s 134 r- V in 1

LO i Q) O;pb0 0 ll a 0 a) 04(U0 Q4 w .w 4 j I 4A 19 (U

0. AaC Q - Q ( BW v(6Q4 A (.4 1= P(1) ~0 #A 0. 4-9 :- - :- -- t- !-4

ol i-p LA p a w 0 4 aC aW4 ( . )* )b%4- * 0 4. :0 W awOpm.p. " :I. ".Q) E 4Q 0 0( 40l w 0 v 0 w 0 4.r

0n 0 -n 0) %n40 4-E E a E E* -

~ 4' 9133

0j 01

0 -4 *. &4.SlJ ta -4 #a 0 z)

*s4 0 -4 4.0 6.n a

4 (U u4 to u~

X.* (L -4 E 0.1

04 45 'a 0

om O 0 0

o b4 to .to 0

31 01. 0 0-4~.4

EL, 0 ) 'a 41t

4.4n 5/1 'a £4934 00

.04 so 0 0I .

4) r- ( 005 ~ ~ - 01 e 4A -

i n "4 *A4 vMa E- .0 Qjt 0 -4 -

('1 inl E-t im ma 4 -4 W1

ce) X2 ap 0.Cl .4 Lo 4

-.4 0 1-4 4-a to 'a a3g -4 5) t-4

24 0. -4 in 04 in nS4 * 0 5 13 nO. .

0 A ') ~ a 0 I-.

~ 0 01 A 0. *-* 00

go 0. *a C: 5a l 0 cE :00. k a sE C: * 4 0 01 E1

EQ'S E -- 0 01.aU 01 00 v) .- 0 *- .4 m*4- Ak 44 44 X- 4-4 *- aj S S

ISnn P-4 r- A 4) r-4 -4 r4 V4w - 1

v* ul = - ( 0 ,4 .- *4b -0& kn .' ma (U *(U5 . 0 cm *a 1 MC\S =S ti P 4 '-*- 0o1

kn 0 -14 o-4 4a * C:, - w t F'a* - (u0 Sn n

Cy C\ CyN-C. NL3 u i C vs. 0-4 W 0 CM P-4 to 0 S - ~ 4..S 0 A 0 saa a# A0.. 1 410. 1 *a 0me

0 &0 Sn 4 4a P-4 ..v- es 4 -4 r4 Q r 0 0.1' 0-i k2SN~lA A (.U 0Ef O 1 00 0 " 10 sicS. a %.. >6 0elo3 m j' 4)f 4: W.. 94: 14: W.: &8: ": W . .2, 44

F.4 (U 014 04. 0*0 a 04u *a %n 0 0 '

0.4 :0 *. mo":wmP "0*4

134

4.) *

f- - -cu (

-W . 0- 2 i

4.) V. "

-r %n 6

E 0

*a P. l) .

r.4 VI SN4 N

0 00 0 "4 1- M r-4oa $4 ;r 5 U(U 0 L

4.) 0-. to r- Q 0O

+ - . r-4 04 (.4 # )4a O

kn 3 . o 4) to s 3 GJ" n -

* 0 0 CJ 0) 4 Q h( 0--4 fQ .0 0f fl(U- 0 OJ 0- ("4A 0)- ~ l

W ~ ~ ~ . "4 06 cu N NImCJU0 QIC Q - r 44 o A

03' IVr44 -AF4UP4> - 4* 4 .b r- 4 i#0 41 0 a) 0 a i) Q %n C

oa 0- *a 0 0 " UU

op K) P-. M owl " p0 oM P l 41. :w sbl

Ea~ E 4- E J E a 4- A

0G)~ ~~~~ 0, w 0 w 0 4 (J 0 --

~flU.) ,4 0 .4 G £4 4.' .4 *. -4 -- k

0 30 O~i .4 ~ .- 44. -.. '135q

* . ~ P4 M

.- 4.4. Le -44 #a %V 4.. .4vi4 &3 i Ga .'4 04. S. w t

6n 4. 0 A -r4 (U -. .. G C

-410 *oi 19 -VM4r4 m.- C.4 M 04S f )'a 40 -0 4.2 -. ..V.%.00 1

W - C 4a V 4a C\1 4JC~ -V44 ~ . j C C 4-44D*6 1 144 0 o I I *o- tU 4 - as I' d I taj 0n . 0 o- *a Wa IV E- X ;.v- %e *-4) r-, %d (L) ,

AM.' P I 4 )( Q -4 . 0 Gi 0 10 -) E- 4) (: 0

leana: k *U- &o -04 %4 41 t. .1 S.c:* O. J 4 6d 1.L

0 to op *a b *a,-40 *A% DO. 00 4 bb' 41 a o0 .m P *ldi ov

G) C)I I4)0Q c0 Ij ) 0 4) (U~ 00 4) o (U 0 i 04 o (U o *

tn 1- .0 0 E~a 00G woG 44Eus4 EV a R ~ aGU Ea~a E ~ EE E 4)Q

t' a: S. a. S...g S..-. .... 4 .: .: 4 S136. . *D~

Ln %a % ... ~ '- . * a

0 !j. (k .

01 to4.

Li 4 4. n 3a 0

94. 93 Eh C: 0L 0) ow0

94 a ca 0

a (U& 4 C9. in 44

4. 0 r- r-4 '(a 0) .4

04 0 0 0e -

m 4 Ag. k.

.~ 4.& 4A o0 .a

in to *a 0) Q) (U 4' 'E4. s. b in. 6h 0 6&

(U 0 4 - 0 11 eli >.b i

0- 0 w aL --

0 in 44.. coo G3 Cbo 9&. (U 0). .4 ##) em. 16

4-P elif IV in 0- -d .0ga-0

~a u0 -4m ummam44 W" t/ En U. 0n Li En

(V CV (V III A 0 04 .. CV W. CY ,ia

V-4A 10J Cn (U E- 4.3)W 41 0) III a) w' .' i4ta aob0 0 0) IO It" k: e. to:e &d. .:ee in p

E *a h 4j .. 34.0 ILI 0 a 4)41 Ca - 0) 4) 4) 4~.10 .0 *a *a (Lo 00 *)#1 ) 4se cu :M. 01 N. H~ oIb 0 - -p m P. 0 4 . - 4 ;w m so =P -, 4D :. PQ 4 0 w/ e 0 ~4)a0 lb. 0 (e 04.I%.eOWO ow*'m.'. Ow.b b e % Ie*b

r. E. 4-ae 0*0* 0 E 40) E o a a - E o

0) 1O .~. *6 *Gi U~0~ 00Jj~i137 0

7.. *

A N 0 .- A 64

C4.4 ~ G

0 4. 0 0 0..00 :bb

01 00 '. i ..

In :3. Q40

.0 4.4

Vii GJ W

0 IQ

-4 E r-

0.~~ C:0 loIL

0l 1 44 00 V0) s' 0

*4 r4 0 .1, &

toe 0 -0 48 0 UV4 *a .- 0 . 4..4. 00r14. i 4 0 s- %.4' - 4 f

Ad4 O0 0 004 1+ v4 v4o4A - - P-4 w' r-4 i%

o0 aI M0 e V to v W 4 !w.

) PQ.8i n W4 06 0 0 .. 44 .a) t ..4.0 Q 0400, S sI. E

0U r4 %4 Pb do a,-%oG-

a~ ~ 'V 41 k Vi o 1 %$ 1

E, u P 0 r -4 '* 14 F-4 0 P -4 W4 r4014 0aj 0.4 t ai Y PO !b414 2m 01 O Z O 0..

to41.to'0 4, to sop 4 1 r W sot S.: 1. Ad: 1

WOO. %A Ei Of W* Vso4 a4 4.' PW*0W VP . I 0 .,1 .. i" : t P14 14 P ** " 0 .Cgo l0 J Q) go 0 S. o~ 4u o a 0 (Uoq ,ac Ao4's-,I

Qs0 6.4 U'. E~I 4 EL *0& E E E *A E@ 0.

bE *n 01 OQ138

%A'J %1 In O .4 0. 1 0 0'.. 00

.9.

I.-.D

*0 0

-4 0

0)4 W

cu 04

-4 03

Nu .4 9. 5.4 0

to Oo 0.

0E Eo: a. Q, 44

%. (U Cl Q

C4. 4.3 4.

4:.a . 0 0 4)

+ ~ 0-4 24 004

0 a. >,, .3 Q -40 4-a Co.a 4 m k ~ 0 C;i

Q 09l 0 4 pa -t (UP 00P

QJ 32 oC. o

r-4 r- U-

4.& ~ I2 -3 W to 4.34-4 0

a.1 (U 03 ..- g4.., 032 .4~0

ro (u oa (v *-V *a 4,a)i 0) w. 0j 4' r4

S. 0 4 0 .9 0 cc0 .(U *e- U0 .1 -4 I

;,., 0)4 -,A' j 9

4. .4p..q. -

94 (~ ~ ~-*M t5.1 0 0 ~ 0

0.. 0 ~ .~. 139 '

30 30 * 30 77.

r41 C-)4 10

A4 W. .-4 I4 30094.

*00 0 .0 0 Obi *o~t L* 4)---

5. .~ W GJ * * . 6. ios W4 C: d) 0

1 0 * 4. PC* . * U *

4- * - iV -4. '. -14 A2 *4 004 *.O L a W N u b *S kt 0- 4.0 04-4&xS.

wJ r4 *q . j *~ to 0f v0 -0 0. a) KJ U) 1. *t -4 *ai * C-- 0 * m r 0) M . *aG) N0 *) a -- 0 C- l- 0 o Q1 *. -4 -ui N 3 M-' %-4 W. d%

cu 4 w -40 V-* *) N \ r-4 0) N ~ C4 4. W4 N .- Ad N4. 0~ C%3

I 1- a I I * I * . I 4 N I " 10 I I I N. I I 1 N .3

-~.4 1 Wr- r -4 G)N.. WJ~J. -40r- C -4* -4A ( v-4 bd C2-44 Mw N W 0 0). Q I4) t 4.3 (L .04 4) "s- a C1* 0 A * 4) %401

i4P IQ , , N O"- - Arii. iJ vO "L -LU G J U a(L 0ft GJ 0'31-0 1*& .Al 0,"

(U ! * w :-a -6 ) P4 4-bG t4.3 !4AJ3Q q) 14-1 !@ 0 4)O 140, a) 4.3 1 LW0 - 0M * 41 40 bNP 4M P - &M4P:.0 40 a"P - 440t

oaji Liw wo wo AO( 0100)0. 000 4. ,0 a)0 0 Wwo 0w.' w 00 , 6a .3 4- W .3jjIn4. E v 00 E. 2 v) &a E I v- 4A~ 0f %A *a E " E 0n %A v'in E

140

*~ -j - - -- - - - - -

0-4 cu cu 540e

a) IA -2 Inl I- Li n

0. e n O- 5. . -

-4 -49O .

o .4

cu4 54 -4 0

fu4 4.4 0 e. 01 b4.3 0 0 04 0. to2.-

93. I0 -

CU ~ 0 0C:

q C. -4 L 4

4 ) 43 E w4

I9 . . 0 (U Q.- 0 54 Q)tV E- >- 0 0 &a 0-I,!-

~'Ve0~ 0 '0 N m a p 4.2

Ir -14 41 54 1 0 % - ) 0)E- En 0) la. * .0 0 a~ .4 0.

E- -m 4A 4-a E go'

:. 04 0 aU 0 to 0 ~~0 4

*0 0) E4a 0 >, (m ' 4 . 0

-E-4 r-A4 0. 0A A)E ~

E4 W 6-PE ;p -44 4An 0)V) hd 4 '

en0)-.-46 In~ enP e+ 0-.'-0 ). 4..-. e .4.

4-Pe 0s 4oC. " 0 a.) 0)e) Q 14aw

tO P4q5 *0 *0 *E305*0

(U bA C- '0 kAP 06p(U

Q) QAa V O'C\ 34.4r4 t..CJ 4" qe j4. 44 C4a)C- - 6 jC4A a)In-')0 I +4ei 0) F- 0 o4%) - -e.) r4 C* : e. c- r-0 r-- 41

14.2 in eno W" U- -0 - - r4 04j 4 4 u ( A0.-4 . m) p) m~ G3 0a) G) 0) m 4

(U 0 0. :4 m)' 04 543 543 94 C4: 54 5: 5:

04a4 W'0 040 M. , "

-40,) CU ( > 4400)00 10ft0"0)0 OJ10E-O 0 Ila 1 -4

w n

14

-4 Q# Q) 0 Le .5 * * - .*0. w. 0 w .0 0w 0 0 w

.****-. E5.. E.. , % % . . . . . . . - . .E....................

Co ~A a

141-- .- - S *

014 0.1 0~

0 s.. %. :3

t -n ;P- W

'a :3 'a '

4.. 0) 0 L6..

in . i

'.4 D4 0- * o 0 -A. .- ~ C) .r J - ;.

PA. W " - J) , C3- CU

0 n0 40 # . 'A- 0. 0 4- V.4 4-D%-C-) ~ ~ ~ ~ @ IA, I+u-#ar4V - q 2c. - J

.44 - D W 0 C - ()>w) '*G204 C . %% *a - 0 - - -- ,X4

CM CU M0 I % . aC

W i k-4 W (Li al 4 ) (L )04 ) ro . : . 64 & 69 5 m . W 4

bd0 t ) U*AG4PW 0& N) *a * * -4 10Q, , V4 m P.. '.4 m- p -":0m wm*0">

E #A E 6- E ' A IE 4- E 0

142 w

A. ~ ' £ \£

1.74

TVJ

4.0

oi 0

&4 r.

E-4 0

.0

4.0 1-4

A 0 .4 0r4 0) 0

c-' go 0 0 )

Nj0 "0 A-0 Q)

tn di' kn.4 .

54 . i ul4 c1: 0to.

". 0 'v W

w4 E 1.- %4 0 r.0*fto 45) 0, S. n 0Ps

1= " 4 4-

0o . 4 .o ~ t 0 (U

4la 5.. A. 4) 1 *a'i 00 1

%n U) 4- V) r -4 .0-- / - 0

0.4~k (U) WP0. L.. J

20' L4.,~ 4~.04~ - bFJ .

(3) I C * e0 / a) (0. *. *- *~4 4-G

Q*~~~ ~~ 4~l., WO1.O**.4.0 to Qj o)0 t44~G ' ~

0U w 0. a) 0.j CJ~~W\O~ C .UU 00.1t0 WjU 0) *' .) 0 . 0

~:-4 v4 rOr4.. .44f

~a1 0 s s.:~ .: 1. S.:,Q .

r-4 .4 4-

143

IZ

cu a- fl - fl

'4 64 64 64 6A 1. 6(U W4 N U64 0

a) W) 4E >4E E ES 10 4-a 0U)b Au to -

o .4) 4E 4 '

I % - n 0) h 6m& t - 4 go 0

0 PC0'Q44s

r- P4 au o- G) 1 . 4- G4 -- mm M Z -C -- 0 -- __

- En .S 12 CIO a -4 .-4 -- "a4 0E~ ma- 40.0 ** c4O 0- qou~ c ~ . Enj N~ Enj *.. N4 sc N s-c sc fi N * s-V V4 r) 0v, on 4.....

0) IA *a.. F- G-4 .0 *C'J .4 #Vr4 Cq3r ) * r*-4 *( 0 ~ *- MC -3 M c P4P4 O 4

EPu6 0 0~J4 U4 0C~. C 4 ~C n 4 V W ~ ~ . .N)' 4JIC *q W 4) 0) AsU to Jro 10 CU)

I Eu 4u: 64 43.W: t 5.4:q 4-: 5. 4. .: w" .~ S.w4g t .4 64.m-: k 54. k 64-4 ~ ~ ~ 4 03- 1~b~ -14.-' - P- IA-- A ;N- .- o.-

4 -( 4 &'G 40 : 44 o) ) 10-0 (i t*G 10-0 60 #-& W '4) :40 Q : J

a) cu OL () 4 a 4 aIO) 0 W0)404) do cu40 w0 w w0 ~w w o

144

Ae

-- o.a -

. . .4 . . . .. .

.4 -40 0 S" di

4 0

04 0 0 -4to 0 u a

4 f4 04

. 04 r * 0 tu

to 0 to w

r %n4) & 45 04 (U

"a E to W 0 = 4.-e. 0 4 o

A I In0

.,o o EnI 4) kJ . -'.

a., . -* Za il 5- 0

• - • 3 O O 0 ' o -

0. *) to

m 34 .- -. 4. 1 0 4 40 4"to .- ,

• ..,,s ,,=o . .- =, £ . 1 .4 .,- 0 0

ej a 0 4 4.

0a* * n . 410 to 0 0 0

* Tl..- 0 0 0 -

14 14 C) 0 0o ~ .40 ~ 0In0

0. ., *M . e '" O . - Z.- r '- E-4, ,- . 0 . . .0"".-

u -- .. $4 0 1 0 V0 0,% 0 A.

'~~ Id goi 0 0.

U',-- v : IV . -4 d . 4 0" 0. " 0 4,.5I '

0 0 C. r. 0n k

£ ,0 - , ,, > .' -. --..4 --,' -E .- .-.. .-- '-". ..-

(U 00)

to 0 0

, ,-,,, = 0 - , £.4a~ . o o..fl:: .

*0 Wa. A 0 5* ~ (.U E N-. v1 £

E4 C0 0o i i oCQ ~ ~ ~ ~ ~ ~ ~ ~ £ .0. qA%44 -o(U14 - :A

I 4.In to * to s 4. # 1:

o, In u0 0 1+ E NC E - -4 m e-4. r."-4 .4 oD ,.. 4 *v0 4 L4 A.- F Q - E .- oe o 0. V .4 -. ,"

OD o v 4 n o Cm 0S G IS) 4 t 4sA. CU *0 C! C.o *c V~ * 0o *0 *o * .

* £4N CYJW4.4. >4 --0 .4.4- C.I 4 3i 0o too r. .4A CU 0 04 0 04- .0 N' 0 cz *-&' Al 0 .\j = N +-

M..4.'m (U r-4 0 Il L 14I) 0 a) -4 1-4 31 .0 a0~) I * 9E- Isf444(u0 noW( E 4u aQ - 0 S. 4lfO,-4 0r-4 r-4 P-4 =r4 r-4 4

£34 l£a t a) UW c (E-O 0) 0 .....S a 10 Go Q v-4b4 Q)D MI- £4: £4: ". k4: $4 6.4

.4 .04361 D f .4 1- -- - - -V- r-Np 4.&4 0. I al n4 oa 4G. 4A (4-a 14- # ( U 4-D (U (U o

000000.4.4

14S

777-.. "

0, 61 ;j 3 .0

.1 04 cu r-4 0 A ..

0~w d3 o 4

.0~~- di4-~6 -

~4 U) 0 -4j.06

in a) w 0 0) V

W C3 C.4.1 "4 1- 6 - 9 14

Cs q-4 0) 4

00

(U 0)4

10 03CS W4 so

fu 04 44 oEU 4.0 0l 4-A A0 S 40 $A a) A.' .3.

0n 0 6

~4 04 \01 ~ .0 .

~~,( Cs iC.: -'A$ .V4 C3 *. 94 #0 0..

EU4'i.4AI U

fu 04 .0 r-4 V / (U 2v~ i L. f"' ah U).W EU05 014O 4.' E co- -.44 5 .

rq 0 04 E F.-..4. E 0

EU 4. - as44 19 n 4 )a* -4 46aQ 4 rw V0* e%1-4 r-4 -4 r- 94.4 0 9404 Q +5 C- 6q(o - -

I&-- S) r-4 G) 4-vto0 00P-.v 0 0 CD .0 oso4G(U W . ) 3. a t1 0) -46J0) 0 E~r9- 6 -9 0

Ph 90 40 0 Cu o *0 0q .o W( go 14 N N 0Ob ON S.. * o- .4

I fu. 13D 1= 10 W A k u -w I *-V 11 4a

P- M -4 f4% .41 P-1 0 4 0) 0) 4 05'i0 -4 (A 4 rJ-4 F-4*0 W 6 QA (-4 * ".4 A- 0 . 9 4 -W. *.* go c

4) 4 ** ) ~*0a S4 q aC 4 0 W 00 to i 6 0 o O 4Dlu4& o w' 4A EUi 40

-. 9.O OW M 4'- : 0 0 ko U-o ovIL (\ EV E' afl a' E E 0 = * 4 C 9E4s'& E '. 3.A E - E ' C .

l4 i-I ..- .4 ~4 qv0 0 6.VT4- -4 . . .

S % 9 e2 -, - 0J 4 94 - *.- p 59 -' k 9

a) 094 44p94 W4 4 U.

0 01 0 aW14oN .a 1r.44 .

'~~0

aaO-

'* 4 0 0 .4

cJ % a 0. 94 IE . 9 4 GO 0)

r 4 S. r- 03 73

4A * 1 . 1 1 .E 4 k '- 0 0

0 V 4. r4 a)..4

%4 C v.. Va-4;)

4A t- 0 -n .1 4 )0 *

P6 4 . - toV0 0J -4.40

O .4 4.' a aU)'

w~~ o- * 94-, a~~,a 00 05

Q --l 4. 4) .4l W--s "a M3. r -4$a 0 4r .41 -0 ) O* %nu4 - 4 .04" 4 03 -,

P- rqf4 - F-4 r- 4 W 0) +- r. r 4 4) -4 5.,Er-A U an .4 IQJ .5. *

(U. Ck 4 CG fi% 0) AZ W4 0 E r.*--. :% 0 Q 9A) 0a4 * 0 0- ) ?*a .a V Q .4 t - U) -. *a * N. - -V: .-.v um.04N '-c --- 'r. a4 41 ..0 MV4 4\1 0 Q li 4'a--

a .I ao 1%.4 100 039 0fl Ifl CU 0 - 4 'Ola0l ' 1 b1r-44 %4 4 W4 #-4 w a1a 0 0 Adi4 = C 4 -4 q~ 50 4

P. 0a (.. F-4 -) *U Wa- E- (.- V .lr U 5.'a~b (1) .41 A.,) ~ 0~* 4

W41 doooWw0,- 0 k'a a w- a U'- w a- 0 j w -4.1 .N0 '41 C k~ 4- E.Cl'0 E 4-v FJa

*147

aLA.. E L30C * * .

.~ 'a s~-4 C2 *

Cu4- 0 -0 . In CL'=

0 -4 -L4 a) ,.

10 ul 4. . .5 -- 4 p(U. 0 2 02 94

0 0 xi0

in InI E4 E &0

i 0 -) $ U 0 ) 0 -

0 k. A% 4.n 315

o 4 IA to a) $4

-1 04 Ag E 4 C. (D C

E4 A 4. E 0 0 f"o~~~ A a~E6

w. E i00 i- u ) 6

o~ 0 0~C ~- . 6d 04 0- It 4) (i 4 0 r

%4. V; 44 0 W~4 C3 4

0. 04 0 3 0 a r-4 40 4 ~ q 45 -

k . r-4 r~ 4. 92C

in OA C. IA fa 0 0

4- 410 a C) -if 04 0).0 4

4.5 E E 1 1. 4.0 0- "4 64() (U. 1Qj 4~, (U AE a -4

tv C4-4 4. %n4 0 4. 0 4 C6.4 0-4 94. 0 .o) c. 4 0 P- 94 .0 1 )c 4 * 4 -4 f-4 0. - nC

." 0 - to 0- A .to 0 . .. kn .* 0

M ) A :S) .- tiI 0 I- In .5 S. "G )I

$A "-M0 Is .' (L :0. un '.4 pt-" 0.M - C .0 k%6 o V4 .0 im -4 01 1.0

V W 1 0 - 0 QIIn E(U- 14 'T E -4 I -f*- F1 4 r~*a-4 2) Ma kn. r4 Al~ V- r-4 . C.4 V 0 *- 0. r.-4

$3 Aj iL Ad ~ U) Q).4- E- --- 64-(- a) W a) Q-):3 b0 00 4= In a I.: &@ &44 & w*

(U"wONW6 0-0-44n *Aa w WW~~.4. *a4 ,4.-4')0 P. * *0 h-40) ; w.. :w In. *o 04 :w 4 P -n 4 *4 m

In') '0 o0 cb a. 4)0 0 0I 0 4C)Q 04 4 k 0 C6 0E.- -. ~ .. . E E E. E E- 4 . -4- i .-

0.~~nC.eC~00 .-4 Pie~4ICJ0

o .'4 A *'0i4I 0 s ' ~4 . , F' 5-J

'.4 '4 '- 0)0 In --- o 14n 40- 4 . e ' .

-. ~~~ (U-*~* -0 '

E- .w 4 Cd

;P4. 0 . C.' 44 vi

'j +.P '4 4-a 00 a,

r- S. 0 4' 0 4-' .4. U,

E V 0 .44 Ei -# '4 ..

di 0 i'U UE- ILO r0 fa ( .

C) C 0 4- 0 -..V) E ) -

4- 94 94 fu d 0 V),l . 0 r-4 ' Q) E -4'4 (

.4 r#- 1+4 ( .

t, 4) r- .C P4 4.a

Q) aJ , -4 %.4 L .8.

4-' 0U r-4. 0d

.C0 4..~0 *C0 4-V

'.' 0 . 0 - WU r-4

Id 0n -- a) C) toA

fu 0 r4 0- E44' 0~4 U

04 $4 0 0 U 0 i '-d

to #j 0l :3~ d

IV r= U4 90 0 0 j 0 a.- 4 1- Q) 4. d 4 '

a)E~N 0 VO..' 0 0- 0i w4 (U-a) 4- di 4U -U. h--d~

'4 a) '4) %A E.U) 0 . 0U~ U40 O w V. Q4-8q 04 r.0 41- fV .C 64

~ ..- 4 di~ C C) d Ok-4.)# P- 4. t*h4to) r+f C.4 g v) W4) *~ t-o 440c- - 4

r-4 r-4 *a a-t r-4 'A3 1+~ =w ~ -44. * J *C' 0 C4 f-*

V tQ r-. .-.- * 0 %nC .14 0 ,+ C%~ 'aI - diM -c OnA ",4 .L * 0 a -4 U) U2~ d~)J * j 14 -4 * * =s

Q)U W 3O 0-~~s C Q) r4--\

fa . IM 10. W #0 4.0d ;3 *dn *.4 IC V ' I *n-4 MO CNP- 4 InC.. I" Sr~0 -4U

* (4-.4- i- a-) .. o4 02 64 wU~ a)di(U-'-4--V .C

O' P w 4- wNi' O~ 00 0 k - fu CU0i.4 0 4a )4- ( 43 ) -wV . "U iC) P.a &~ to w 0. w -m ~ .0 P. m s -K .4

-4 -4 Wi 0 W. -4 k - **wo '..4o 0 w*qIF Ei 0) Ei '4E Ei E4 E.r~j Ei Ei E

- ': .: .. 4gN AN di~ ~ ': 4 ':14: '4

1= 6.04 4

c 14 rU -44

to 4. LIr. 40 E

6 -4 -4--S

oj EQj to ~ U

.44 0

ko 0. .0 .3 . 4-aEn E-~540 E

in :0 %A- - 5~

.4 0 0 iv3 to 4 .~~~.~ U~ .E 4

-S S2

0. V) 04 .4

* 6~r4 -4 0 U 54 .

-4n 1-6~ a En E 0 4.

in En >0( 4

fu cu P4 =4 EU 44100 0£ V 4 41 4k. 0) 4 En C

EU n t 0 0+W4c 3 - 0- EU , r- C3 -4' a - A 01-4~~ En EU En to ftU :

-,0 (go :', w )p 5tva v 0

0 ~ ~ ~ ~ ~ 00 0i P4 -4m; E U , k4 aSt a a

-r f4. to 0 0%WEU r. iv 0 06 .0 &.1 N q - jQ% 3 YA" A

14 04 0) I.4 14 -AI.o1" 4 1 % O-

4) M -. 0 k A O - 1= 04. 0- r-4 04 (0- r4 WO

Iz-r o -1 )E 0 En d £4 .104C~. vi to EU4 f"t: w 1u 4 k 4 9: &Z s

;I " * 0 F-- Ov wt rmp w " ;p -4 K-o.*t.En W.4o Ef to ILO 0 w 0 Ow* o

E O EUP F.. El) E E .~)

~,aj~n ~ 4E~,.'- U .3 .c p- OjU0

3.435~~~~~~ % .4 n.'0 3 ~ ~ E

W 0 di 4- .00

;*a Ui - 0 0ED w4 0.

V M ~ 0

W4 0 It,

-rq ek3 *a3 0 lk :s(

on 0. = ~ 40

54 gm4 dc . t

a, w 04~ U .3 43

0 w w F J

04 4) J--4 in..

ka 04 C" b 54 E a E(U a '- 0 4&G

0 k 40 .00li .540+ 4 E . 4 1. 4P-4 ) 5-4

.-fUt Qn IDF4chs 1 0 4)s 00 04 o 4.r 3 0 2 4- 4 .0 ) )Ui

0 4 r- 9 V r 0 40m*u 0 s0 a0q

r- -0C VCY6 ~ 4 I 0i N, CU .q C4 J3 IV004 An 42 14.3 4)o4 a1 1* i 4 14- *a29 k

k n r4 rI -I F- F .4 r-4 r- 4)qd. 0%n#AP4 -Q n -4 Q) Q u 0 c.4 4 Q Ut AC

0 -. 0 w0w0)0))0 o ) 0~ W 0 )Q o

r- ) 1E40 bE- *AF C6 4.3 0 4- kE E4-W0 2 0 r4. 3 5

* Ut , ~ '4 0 4 40-4

Le 0 t*

in Up 14 do*_

0~~ .54 . G

ona

k n .4 w. 4- e 0G

O&A A3 *- a) In 11. 14IV -4 V* P. *4 0 5 '.

4 4 .a A 4-D a to 4

4) In 04 =- .. _. 0 _.

a C3 . til . 4.4 0 ~ S G94 C) do ILIW

(a. to' 0Q L. U 943~~q ~ .o ) 4u ) -.4 0) 6-P o.

kAV2 1. .0 kA 654) 0 AP

0 4. 2 C:4 03 Ga W-1 (U ,

&0 0 00 0 *AW

U-b ~ : -A 4) W I 5 a 0 )a0 0 5C 0 a C2- @

0 Ga S.. 0 . 00 .-0 0 Ga. 0 In. iP U

0 6-M 4 /P C .4 -0 md .. "o. w 0 tn ..4 00 M n ~ ~

.0 . 0 1.. w a- E 41 D

P4 Ga c' on v 0 z 0 ( f A bO* Z *-- G a 00 E 0 -4

IL.-0 S. 0 r4 m r_ 5 w~ w.. %,

0- 0O 41.. Ga IS 0a 4" '4 In 0keU.4 i n 0 a4 r-4 0 %A 0 00~ Ad FE 4

0O 50 0 14- 0 5. 00 so 4) 0 -

r. Fl. o.- 0 n I-. n kne-0 .* e--, E ..s

2h A t.'.4.t-'too-.C-5o 0~ -4 M #a- U4 - o-C - Ud, #a.n l '.4* w 4 so a . - a d oI* 1ql * C. * to - .4 4,~k* e+CC=.'a\j -4 0 -4 C'.j v-. 4 .& 4 9CY0 14.v-4 U C\.-4e-4 A 4

enl--" -- -~ -eIn - (u &#a *0 o 0 *ooZ -. 5a -0 o a - & - %An

II -4 a# 4- - 3t Q)~.5940# Ga~ .G.. 02 00G M a6 .5 IV0 j 4, 4 'GaI *41 . m *-'a(

*4 4C3 k4- la ~-l 0- F - t 2' ,"ca. - 31 4- 0*- ev- Ga'-WC ( Ito N " c i P". in IV-4 C\) - 4 N r"I4N W' 0N C~i , ~ z.. C\?JI~ 0 .5.'jAN W0-b I 1 13C I O 14)a .0 . 4) d2'. i.-4 6 C- I r. I a 1, 10 It) I, s I f41I oa W "4 40 A3 -4 f_4 -Cr4 r_4 r.4 F4 4g4 .454k..4 "Ol .

4e.P4 P-0 0

94DNo4 S..4c5ILI.G qu 414) a# a Ga L (34 111l~ GE w 4)10 a 40 *g:go4(: Se4a 4u s W ~ W .k .. 6.4 W. 6.4 6.4 fs. W.. 54 k.: k5..

Ga DO5 N. M, :P 0Mt ,- P- 0 %-~ P4 ObM - " ~ II-M oiao 0 46 W 0 a# a )0 )Q( j04 0 .5Go.5 4. GJ5Q.5 . 0a4.G.5 .Q 4)4.Ii a o 4) 0

152Z

-- ~~~~~~~ 6,W -_Vg-- * . .. . . . . . .

*vi..

uJ. 9 a *

S.. Z0 U0

b4 0.

b-44

4~ ) F..~ .

00 W--4

0 .1 :! .:a

LV 0.

IV w 4.

3 %G In

G CY' P4(Ucn4)C.~ I W4i n ( ..

P~-- . tL-4.-c 031.

U)EU tGI -.0 IL (U - - A+

Qj> . U0 .a .A1 - w (U a)0 )

r-.~ 04. fu ~ 03 d- A c0 ) oS'0*tiO0. R C.) #- .mE t

o ;3v 4) V1 p * 0d A ~ 4A6 iv4 o 1

IA @-4. r- V.1 !" U00 - 0 (L o6n% 04-4410)0 *- 0 1 0.-- 0)0 L Q V -

0) ) U b4W4c CV w ' 0)4 4 0 03-5 g

W 4A vE) 0 -CV( otorC~~4GO 030. Q ) M . 0-;. .1 i~E

(U ~ WV4 to go-t P4 U)IA

E- V. L6w -9P4$ 4 Utk I2 4 0 .'t 4 w ~ E-' 0U-.

to.1U 6.40J~ %O4i 40 4-' a)* ,E t u41c

16 bg 03 pbG .64 q .0 0 W 0. )4 >

W0 * 0 .. w t0 Q 0 064 w0 W94 4 94 o W 0 G034. OO )P LJ.q'01' p %n vi0 o wo: u Z

r- 0to 4-Dp t 4 N

153

14.

In 03 *4 0 V. 0

j n*0) 0 (1 . I-.

*~~% .3C) . *

4-P 00IM 4) u3 W) 01

40 A. -43 o (U CI 031 mg * . 0 -4 43

kJ 9 ' 3in '

.4 0 .4 -r .14 0 3 0

-P" A d 4A .~ 0-, "

4n 03 01 ao 0 O

00 0 (V #A cu cu

0in 4.' #An

~ .3.~ ~0 r ..

'a .4 OW 03 !ni 0(k 1.. =) u ~ ft 0

03 - S. -4 .3 la u J~ IL n -

0) 9d ' 0 knNn* k3 0

S.. to. +f 030n 4 * ~ i -

6- 0) 114 L2 kf ) " 63 'a .-.I

4. 03 04 Q) 0 toEE -. 1

to %-4 41 0 34,g)a0

0 '3 E r- 0- CfL.. r- 1.) F- . ok.. 5. 0. . -~ a 03 oE(

03 =3 , Q s. -.0 as+ f=. rp-"6.kn42ir A3 sU . 0P t Spi *. "q to $43

'3'. . o -. If0 5. in in- 4). C3 r4 -, - C 4 . .

-u 6. oi 0 0 'a -4

e104- Ad -4~. 40% W~i W) .0.4.-Ir4) E a ZWiMi

Go&0 40 4 6 0V ) 4-P 03p mr.~ J p..~I "~f *l a) oi 0- v903 :

wowG cowowpS 11 0 1 w4owors va iE. 16-... *- 0& :1 u Fa- E F3-~- 9 EU~

~inC~in~J~J NiN 'a 154'

.1.. * @5 s s ~ ' in 3 S. '. '3 :4.

CA A *a .

%n. -4 0 .* "4 - 0 3 - E

0 9i

w W4uI.

I. -P4-9-~ ~ -4A*p- j . p * .

--- v--I---.~ 7 ..

In 0

q. E -0

4 0 .

o r4

i aOA 00 0 0(

.m0 cm LM

12 0 140

(U E- GS

w oe w Q4-0

0. 01 '156

L 1S T 0k F FER E N

1. Ar-le Coraters Inc., r'ACITO.QH sales troct!.re,

P2. S!mith, D. C.; Irty, C.; Kirtall, R.; Verplank,B. Har1sem, E.; 'tesienln6 tne Star User In-

terface; BYTE Mag4 z I n , April, 1982, 1pp 242-

Z.Sr'ielderm~an, B.; "Humtan Factcr§ Experiments i aDlesigning Interactive Systems , Compuje~rDe-.egrbr 1979, pp. 9-1-W.

4. Cal Stuart I.; Moran, TomTas F.; Newell,Allen; The Psjctiojpg;( of Huvan-Com~pgt! in-teraction , Erlbaun, 198.

5. MacLennan. B. J.; PrIgg-i~1es of Prcg;ArnmngLang~iagqs: Design, Tyajuation, i~d 1~em en ta-

Hlnjolt, Rinelbart and Winston, 1983, p52?.

6. Teiteirran, W.; Masinter,. L.; "%ie lnterlis 1Prcocramiine Environmert , CITmppjpr 14:4,

* Ajril 19E1, .25-34.

7. Diier, Barry; A User Frierdly Alf.orithm", Cor-pjg~ ~ £f th ACt1 September, 1,Vl

24,nicto. 9,E1 ps. o-5

8. Foley, Jarres D.; Van Dar., A.; The Funrdamen-t~l of Interactive Cc mji uterkddisor-Wesley, 1982.

P. Heckel, P.; "How to Design Lser OrientedSoftware" *TD9 glements Of FrI endly ScftwarreSjm;osium' by 1,uickieiw Systiers, Los Altos, -.A,1962.

15?

%'. %

INITIAL DISTFIBUTION LIST

1. Defer~se Technical Inforrratior Center 2Cameron StationAlexandria, Virginia -22Z.14

2. Ljitrary, Code 0142Navel Postgreduate Sch~oolv onterey, California 93.43

3. Ee~artment Chairman, Code 52Department cf Ccrputer ScienceNaval Postgraduate SchoolVonterey, California 93-'43

4. Curriculum Officer, Code 317 1Con'~uter TechnologyNaval Postrraduate School!Ionterey, California 9394-3

5. Professor Gecrge A. Rahe, Code 321Computer Science Department K

a ~Naval Postieraduate Schoolr'onterey, California 93943

6. LCD1 Fredlerick F. ;roernert 2IiS L-35Naval StationVayport, Florida 92228

1 SE

I.FILME D

3-8

DTIC