02-12 one size does not fit all_legal protection for non-copyrightable data

7/30/2019 02-12 One Size Does Not Fit All_Legal Protection for Non-Copyrightable Data

1/6

contributed articles

september 2009 | vol. 52 | no. 9 | communications of the acm 1

doi: 10.1145/1562164.1562196

by hongwei Zhu and stuart e. madnick

The Web has become The largesT daTa reposiToryon the planet.a An important factor contributing toits success is its openness and ease of use: anyonecan contribute data to, and consume data from, the

Web. As Tim Berners-Lee, inventor of the Web, said,the exciting thing is serendipitous reuse of data: oneperson puts data up there for one thing, and anotherperson uses it another way. b Such serendipitousdata reuse is extremely valuable. Through reuse, newknowledge can be created, innovation and value-added services become possible.

However, there have been efforts to regulate andlegally challenge data reuse activities. The European

Union (EU) has adopted the Database Directive to

restrict unauthorized data extraction

and reuse. In the U.S., Congress has

considered six bills, all o which ailedto pass into law. These legislative activ-ities are summarized in Figure 1; more

details are urnished later. The signi-

cant uncertainty and the internationaldierences in database legislation

have created serious challenges to the

serendipitous reuse o data. The dualpurposes o this article, both related to

the theme one size does nott all, are

to: summarize the range o legislationin current use and proposed; and pres-

ent an economic model or interpret-ing or recommending policy choices

that depend on actors such as cost odatabase creation and level o database

dierentiation.

As computing proessionals con-tinue to develop technologies (such as

data extraction, Web mashups, Web

services, and various Semantic Webtechnologies) to make data reuse much

easier, it is important or us to under-

stand the legal implications when ap-plying these technologies or data re-

use purposes.

eBay v. Bidders Edge: d

r f Ll cll

Let us start with an example. With mil-

lions o items auctioned at hundredso online auction sites, it can be time

consuming to nd the specic items o

interest and keep track o their biddingprices on multiple auction sites. A num-

ber o auction data aggregators, such

as Bidders Edge, emerged to addressthe challenge by employing computer

agents to visit auction sites repeatedlyand extract data systematically. BiddersEdge made search and comparison

o auction data across multiple sites

much easier by gathering bidding data

o over ve million items rom morethan 100 online auction sites, including

eBay. However, in late 1999, eBay sued

Bidders Edge and won a preliminaryinjunction in the ollowing year based

o sz f

all: LlP n-

cpld

a In the ensuing discussion, we will consider a Web siteowner as a database creator.

b An interview by Technology Review, October, 2004, p44.


2/6

2 communications of the acm | september 2009 | vol. 52 | no. 9


Co.e In compiling its

phone book coveringthe service area o Ru-

ral, Feist reused 1,309

o the approximately7,700 listings in Ru-

rals White Pages. In

the appeal case, the

Supreme Court decid-ed that Feist did not

inringe Rurals copy-

right in that Ruralswhite pages lack the

requisite originality

to warrant copyrightprotection. Original-

ity requires a work

to be independently created by the au-thor and it must possess at least some

minimal degree o creativity. Arranging

entries alphabetically does not have the

required degree o creativity.The Court conrmed that copyright

rewards originality, originality requires

some minimal degree o creativity,and Originality is a constitutional re-

quirement. It also rejected the so-called

sweat o the brow doctrine that consid-ers copyright as a reward or the hard

work that went into compiling acts.

The implication o this landmark deci-sion is that in the U.S. copyright currently

does not restrict the reuse o the actual

contents in most publicly accessible da-

tabases on the Web.

on a controversial interpretation o tres-

pass law in the Internet context.9 The

case was settled later without a court de-

cision; Bidders Edge ceased operationand the company no longer exists.

There have been several other cases

involving data reuse in the U.S. A com-mon characteristic in these cases is

that the data reusers (such as BiddersEdge) tend to be smaller rms usingnew technologies to extract and reuse

data rom one or more creator data-

bases. In many cases, the data reusersstopped their activities in ear o the le-

gal threats posed by the creators. Exist-

ing and emerging technology-enabled

data reusers continue to ace legalchallenges. For example, data reusers

that provide airare comparison servic-

es have received warning letters rom

some online travel agencies.

c

Data reusers in Europe have also

aced legal challenges. For example, Wil-

liam Hill, an online betting company inthe U.K., created a database by combin-

ing its own data (such as betting odds)

with horse racing event data publishedby British Horseracing Board (BHB),

which was the governing authority or

organizing horse races in the U.K. Wil-

liam Hill displayed the contents o thedatabase on its website to acilitate its

betting business, but was sued by BHB

or its systematic reuse o BHBs data.These cases have raised several ques-

tions regarding technology-enabled data

reuse: Is it legal? Should it be regulated?I so, what are the issues and how should

it be regulated? We will address these

questions in the rest o the paper.

Feist v. Rural: n-cv

d c a n

cpl u.s.

Many people think that the actual data

on Web sites is copyrighted, thus extrac-

tion and reuse o the data rom Web sitesis well-dened and controlled by copy-

right law. It turns out that is not the case.

When it comes to data, copyright inthe U.S.d protects the original selection

and arrangement o data, but not the data

itsel or the eort in compiling the data-base. This principle was established in a

landmark Supreme Court case between

Feist Publications and Rural Telephone

The Court decision, together with

the exponential growth o digital inor-

mation and the increasing technologi-cal capability o reusing inormation,

have induced a series o legislative ac-

tivities to provide legal protection or

database contents.

ill cp Pv

d d P

dCopyright law diers internationally

in terms o how much protection it ex-tends to actual databases. In the U.S.,

copyright protects the creative selection

and arrangement o data, not the dataitsel. In other words, the creative choice

o what to be included in a database and

the creative design o the database sche-ma are protected by copyright in the

U.S., but not the actual data records.

Although the U.S. has rejected the

sweat o the brow doctrine, Australiaembraces the doctrine or its copyright

e U.S. Supreme Court, 499 US 340, 1991.

Note that Web content, such as news articles, music,video, and such, are not data and are protected bycopyright law. The ocus o this article is on data, such as,

in the previous example, the list o items or sale on eBayand their auction prices.

c See Cheap-Tickets Sites Try New Tactics by A. Johnson,Wall Street Journal, October 26, 2004.

d International dierences are discussed later.

f 1.

f 2.


3/6



law as evidenced by the appeal caseDesk-

top Marketing Systems Pty Ltd v. TelstraCorporation Limited.g Desktop used allthe entries in Telstras white pages and

yellow pages to make CD-ROMs with

several additional search eatures. The

Full Court ruled that originality doesnot require novelty, inventiveness or

creativity, and a work is original i thecompiler has undertaking substantiallabor or incurred substantial expense

in collecting the inormation recorded

in the compilation. The High Court oAustralia conrmed the judgment in

2003 and maintained that Desktop in-

ringed Teslstras copyright.

The dierent creativity require-ments o the U.S. and Australia repre-

sent two extremes. The Canadian law is

somewhere in between the extremes. In

the judgment o a Canadian case,

h

theCourt decided that originality need

not be creative, in the sense o being

novel or unique. A work is original iit is more than a mere copy o another

work and requires an exercise o skill

and judgment that must not be sotrivial that it would be characterized as

a purely mechanical exercise.

Despite these dierences in the cri-

teria or testing originality, copyrightlaw is quite uniorm internationally

that one cannot claim copyright pro-

tection or individual entries o actsstored in a database.

h d Ll

Database creators have tried several

ways to protect their non-copyright-

able contents.i A commonly practiced

method is through access control,

which oten requires user subscriptionand authentication. But this does not

prevent data extraction i the user pro-

vides identication to the aggregator(or example, a user provides login cre-

dentials to a nancial account aggre-gator or it to gather inormation romdisparate accounts on the users be-

hal8). Enorceable contracts to restrictthe extraction and reuse o the data are

dicult to establish on the Web unlesscumbersome click-through agree-

ments are in place. As a result, some

database creators eel existing law doesnot give them sucient protection to

their data and their investment in cre-

ating databases. Consequently, theyhave sought means to protect their

data through new legislation. See Fig-ure 1 earlier or a summary o legisla-tive activities.

The EU rst introduced the Database

Directive in 1996 to provide two kinds

o protection or a database: copyrightor the selection or arrangement o da-

tabase contents, and sui generisj right

or the contents in the database. The

sui generis right is a new type o right toprevent unauthorized extraction and/

or reutilization o the whole, a substan-

tial part, or systematic extraction and/or reutilization o an insubstantial part,o contents o a database that is created

with substantial expenditure. Lawul

reusers are restricted not to perormacts which confict with normal exploi-

tation o the database or unreasonably

prejudice the legitimate interests othe maker o the database. Here the

legitimate interests can be broadly

interpreted and may not be limited tocommercial interests.

The Directive has been criticized or

its ambiguity about the minimal levelo investment required to qualiy orprotection,5 its lack o compulsory li-cense provisions,1 the potential o pro-

viding perpetual protection under its

provision o automatic right renewalater a substantial database update,

and the ambiguity in what constitutes

a substantial update.Under its reciprocity provision, da-

tabases rom countries that do not

oer similar protection to databasescreated by EU nationals are not pro-

tected by the Directive within the EU.In response, the U.S. database industrypushed the Congress to provide similar

protection to database contents. Since

then, the Congress has considered six

proposals, all o which have ailed topass into law.

HR 3531 o 1996 closely ollowed the

EU Database Directive approach witheven more stringent restrictions on

data reuse. One o the main concerns

is the constitutionality o the scope and

strength o the kind o protection or

database contents.1, 7

All subsequent U.S. proposals took

a misappropriation approach wherethe commercial value o databases is

explicitly considered. HR 2562 o 1998

and its successor HR 354 o 1999 pe-nalize the commercial reutilization o

a substantial part o a database i thereutilization causes harm in the pri-mary or any intended market o the da-

tabase creator. The protection aorded

by these proposals can be expansivewhen intended market is interpreted

broadly by the creator. At the other end

o the spectrum, HR 1858 o 1999 only

prevents someone rom duplicating adatabase and selling the duplicate in

competition.

Following the reasoning in theNBA

v. Motorola case,

k

HR 3261 o 2003 hasprovisions that lie in between the ex-

tremes o previous proposals. It makes

a data reuser liable or making avail-able in commerce a substantial part

o another persons database i (1)

the database was generated, gathered,or maintained through a substantial

expenditure o nancial resources or

time; (2) the unauthorized making

available in commerce occurs in a timesensitive manner and inficts injury on

the database or a product or service

oering access to multiple databases;and (3) the ability o other parties to

ree ride on the eorts o the plainti

would so reduce the incentive to pro-duce the product or service that its exis-

tence or quality would be substantially

threatened. The term inficts an in-

jury means serving as a unctionalequivalent in the same market as the

database in a manner that causes the

displacement, or the disruption o thesources, o sales, licenses, advertising,

or other revenue.

The purpose o HR 3872 is to preventmisappropriation while ensuring ad-

equate access to actual inormation.

It disallows only the ree-riding thatendangers the existence or the quality

o the creator database. Unlike in HR

3261, injury in the orm o decreased

revenue alone is not an oence.On Dec. 12, 2005, the Commission

o European Communities2 issued

gFull Federal Court o Australia, 2002.

h Supreme Court o Canada, CCH Canadian Ltd. V. LawSociety of Upper Canada, 2004.

i Due to limitations on length, we will not discuss all thetechnical methods that have been used, such as blockingrequests rom IP addresses that appear to be extractinglarge quantities o data, etc. In general, or each technical

approach to prevent data extraction, there is a possibletechnical counter-measure to overcome it.

k 105 F.3d 841 (2nd Circuit, 1997). Motorola transcribedNBA playo scores rom broadcast and sent them to its

pager subscribers. The misappropriation claim by NBAwas dismissed.j In Latin, meaning o its own kind, unique.


4/6



allowed to the extent that the creators

still have enough incentives to create

the databases. Consensus can develop

or international harmonization i wecan determine the policy choices that

eectively balance these actors; a da-

tabase policy so ormulated shouldsurvive the scrutiny o constitutionality

and other ineciencies can be avoidedor mitigated.

av bl

d Ll

We approach the challenge with an eco-

nomic model12 that considers the com-mercial value o databases. Based on

dierentiated competition theory, the

model considers a database creator,which incurs a cost to create the initial

database, and a data reuser, which ex-

tracts a certain amount o data rom thecreator database to create the reuser da-tabase. The reuser database can be di-

erentiated rom the creator database

in terms o scope (such as, extracting araction o the creators data, combin-

ing it with data rom other sources) and

unctionality (such as, dierent kindo search algorithm). The reuser uses

technology to allow it to easily extract

and combine data rom existing data-bases so that the cost o creating the re-

user database can be negligible.

The competition rom the reuser da-tabase can reduce the creators revenue.

When the reduction is such that the cre-

ators revenue cannot oset its cost o

creating the database, the market ails.o

From an economic point view, regula-tion or data reutilization is needed to

prevent or correct market ailure.

A regulation potentially can restrictcertain stakeholders and benet cer-

tain other stakeholders, but the society

as a whole should better o with theregulation. Our analysis shows that

such choices depend on the relation-ship among several actors. The mostimportant two are: the cost o creat-

ing the initial database, and the level

o dierentiation between the creator

database and the reuser database. Thechoicesp in relation to these two actors

its rst evaluation o the Database Di-rective. The evaluation shows that al-

though the Directive helped harmonize

copyright laws within the EU, the eco-nomic impact o the sui generis right

on database production within the EU

is unproven. In addition, the scope othe sui generis right has proved to be

dicult to interpret and its relatedprovisions have caused considerablelegal uncertainty, both at the EU and

national level.

These worldwide legislative initia-

tives demonstrate the substantial di-culties in ormulating a database

protection law that balances creator

incentives and the values added bydata reuses. Some o the challenges are

briefy discussed here.

c Pv Ll

P d c

Data monopoly. There are situationswhere data can only come rom a sole

source due to economy o scale in data-

base creation or impossibility o dupli-cating the event that generates the data

set. For example, no one else but eBay

can generate the bidding data o items

auctioned on eBay. A law that preventsothers rom using the actual data rom

a sole source in eect legalizes a data

monopoly which would endanger any

downstream value-creating reutiliza-tions o the data. The European Court

o Justice (ECJ) partially addressed thisissue by trying to distinguish data cre-

ated rom data obtained, and by pro-

tecting only databases whose data isobtained by collecting existing inde-

pendent materials.l

Cost distortion. Both the EU databasedirective and the latest U.S. proposals

require substantial expenditure in cre-

ating the database or it to be qualiedor protection. Database creators thus

may over invest at an inecient level toqualiy;10 see12 or an economic model

that explains such cost distortion.

Update distortion and eternal protec-

tion. This is an issue in EU law, which

allows or automatic renewal osui ge-neris right when the database has been

substantially updated. Such a provision

can induce socially inecient updatessolely to attain eternal rights.6

Constitutionality. Although the U.S.Congress is empowered by the Consti-

tution to regulate interstate commerce

under the Commerce Clausem and themisappropriation approach oten gives

a database law a commercial guise, this

must be balanced against the Intellec-tual Property Clausen which restricts

the grant o exclusive rights in intan-

gibles that diminishes access to publicdomain and imposes signicant costs

on consumers.4 Certain database con-

tents are actual data in the public do-main; disallowing mere extraction o

such data or value-creating activities

runs aoul o the very purpose o the In-

tellectual Property Clause to promotethe Progress o Science and useul

Arts. Excessive restrictions on reuse oactual data (a orm o speech or press)may also violate the Constitutions

First Amendment,3 which protects thereedom o speech and press. Since lit-

tle extra value or the society as a whole

is created by simply duplicating a data-base in its entirety, preventing verba-

tim copying o a database is likely to be

constitutional. A constitutional data-base law needs to determine how much

one is allowed to extract database con-

tents. The constitutional line-drawing

between extraction and duplication indata reuse is very dicult.4

International harmonization. Giventhe global reach o the Web and in-

creasing international trade, it is desir-

able to have a harmonized data reusepolicy across jurisdictions worldwide.

We have discussed some o the dier-

ences in the U.S., the EU, Australia, andCanada. A World Intellectual Property

Organization (WIPO) study11 also re-veals dierent opinions rom other

countries and regions.

A key element to solving these chal-lenges hinges upon nding the rightactors or a reasonable balance be-

tween protection o incentives and pro-

motion o value creation through datareuse. With this balance, value cre-

ation through data reuse is maximally

m Constitution 1.8.3, To regulate Commerce with oreignNations, and among the several States, and with theIndian Tribes.

n Constitution 1.8.8, To promote the Progress o Scienceand useul Arts, by securing or limited Times to Authors

and Inventors the exclusive Right to their respectiveWritings and Discoveries.

o Market ailure is an economic term or the situation wheregoods or services cannot be provided to consumers(such as it is not proftable or creator to produce thedatabase.) Policy intervention can sometimes restore aailed market.

p There are actually more than three regions in,12

we havesimplifed the situation slightly to shorten this paper.

l European Court o Justice, Grand Chamber, The BritishHorseracing Board Ltd and Others v. William HillOrganization Ltd., 2004. A database creator that createsthe data, such as, BHB, which created the fxture list,would be a natural monopoly i legal protection was

granted. Data that is obtained presumably could beobtained by anyone willing to make the eort.


5/6



are depicted in Figure 2, which, as we

mentioned earlier one size does nott

all, illustrates that the policy choices

are not just binary.

No reuse region. When the level o

dierentiation is low, not allowing re-

use is a reasonable policy choice sincesuch reuse adds little value, and, at the

same time, the intense competitioncan drive the price so low that the cre-ator cannot have enough revenue to

oset the cost. Verbatim copying o an

entire database is a typical example othis scenario.

Free reuse region. When the level o

dierentiation is moderate or high,

there are two scenarios where ree reuseshould be allowed: creation cost is low,

or dierentiation is high regardless o

creation cost. With moderate dieren-

tiation, competition is not as intenseas that in the case o low dierentia-

tion. The sotened competition allowsthe creator to make enough revenue

to oset its cost. With high dieren-

tiation, there will be little competitionbetween the creator and the reuser. In

other words, the data reutilization has

little impact on the creator.

Although in both cases the reusercould be required to pay the creator a

ee, this is not needed to prevent mar-

ket ailure nor is it desirable because

there is always an ineciency asso-ciated with money transer, which is

known as transaction cost. The ee canbenet the creator, but it does not cre-

ate any extra value and the society as a

whole incurs a transaction cost.

Fee-paying reuse region. When thelevel o dierentiation is moderate but

the cost o creation is high, the reuser

should pay a ee to the creator. This isthe case where without a ee the reuse

would cause market ailure, but with

a ee the creator can sustain. Since the

creator may not be willing to license itsdata to the reuser, a compulsory licens-

ing provision should be in place.

s expl ill

appl Ppl

The economic model provides a useul

ramework or acilitating the ongoing

debate o database legislation, analyz-

ing data reuse cases, and interpretingcourt decisions. We will illustrate the

applications o the model by revisiting

the two cases mentioned earlier.

eBay v. Bidders Edge. According to

our analysis, we need to at least exam-

ine the level o dierentiation o the

database developed by the reuser Bid-

ders Edge. In terms o searching obidding data, the reuser database has

a much broader coverage; thus, there is

competition rom the reuser database.In terms o unctionality, eBays data-

base allows one to buy and sell items;the reuser database does not provideany actual auction service. Thus the two

databases exhibit signicant dieren-

tiation. Searching alone does not, ingeneral, reduce eBays revenue rom its

auction service. eBay can still compete

in the search space, but according to

the model eBay should not be given theright to prevent innovative rms such

as Bidders Edge rom oering search

unction beore eBay acquires the neces-

sary technical and business skills. Fur-thermore, i we subscribe to the spin-o

theory,5 the eBay database will not meet

the cost criterion. Thereore, ree reuseby Bidders Edge should be allowed.

BHB v. William Hill. The ECJ deter-

mined that although William Hill didsystematically extract and reuse an

insubstantial part o BHBs database,

the cumulative eect has no possibil-

ity or William Hill to reconstitute andmake available to the public the whole

and substantial part o the contents o

the BHB database and thereore seri-ously prejudice the investment in the

creation o the database. The criterion

o reconstitution eect can be ex-plained using the economic model as

the reuser database having little di-

erentiation. The ECJ also stressed thatthe injury needs to be serious, which

can be understood rom the market

ailure perspective in the model.

BHB spent 4 million annually tomaintain the database. The ECJ judg-

ment provides a guideline or deter-

mining i this cost is protected by theDatabase Directive. Ater making the

distinction between creating and ob-

taining data, the ECJ determined thatthe investment protected by the sui gen-

eris right does not cover the resources

used or creating the materials whichmake up the contents o a database.

To create the racing list, BHB had to

veriy inormation o participants, such

as, a horses age and pedigree, and suchinormation was obtainedby BHB. The

ECJ urther ruled that The resources

used or verication during the stage

o creation o materials are not part

o protected investment. These cost ac-

counting rules used by the ECJ consti-

tute a particular standard o determin-ing the cost actor in the model.

cl

Although the legislative eorts may

seem to have stalled in the U.S. duringthe past ew years, the issues relatedto technology-enabled data reuse have

not been resolved. We discussed these

issues and presented the preliminaryresults o an economic analysis on how

to balance the benets o data reuse to

society and the interests o protingrom creating the initial databases.q

The results show there is not a one-size-ts-all ormula or data reuse regula-

tion. Rather, depending on several ac-

tors, no reuse, ree reuse, or ee-payingreuse are welare-enhancing choices.

As technologies or reusing data

rom various sources continue to

emerge and improve, the need or un-derstanding the legal implications o

applying these technologies will be-

come increasingly acute. We are con-tinuing to develop urther understand-

ing o the issues related to data reuse

technologies. The research will bringus closer to nding the right balance

with which serendipitous and innova-

tive data reutilization can be maximallyallowed to provide value-added services

without diminishing the incentives o

compiling databases and making them

available on the web.

References1. Colsten, C. Sui Generis database right: Ripe or

review? The Journal of Information, Law andTechnology 3 (2001).

2. Commission o the European Communities (CEC).First Evaluation o Directive 96/9/EC on the LegalProtection o Databases. (Dec. 12, 2005), Brussels.

3. Grove, J. Wanted: Public policies that oster creationo knowledge. Comm. ACM 47, 5 (2004), 23-25.

4. Heald, P.J. The extraction/duplication dichotomy:

Constitutional line drawing in the database debate.Ohio State Law Journal 62, 2 (2001) 933-944.

5. Hugenholtz, P.B. Program schedules, event data andtelephone subscriber listings under the databasedirective: The spin-o doctrine in the Netherlandsand elsewhere in Europe. 11th Annual Conference onInternational Law & Policy(2003), New York.

6. Koboldt, C. The EU-Directive on the legal protection odatabases and the incentives to update: An economicanalysis. International Review of Law and Economics17, 1 (1997) 127-138.

7. Lipton, J. Private rights and public policies:Reconceptualizing property in databases. BerkeleyTechnology Law Journal 18, 3 (2003) 773-852.

q There are many other actors, such as the political, legal,and enorcement processes in dierent jurisdictions, thatare beyond the scope o this paper. The intention o this

paper is to establish some basic principles that couldacilitate these other processes.


6/6



8. Madnick, S.E., Siegel, M.D. Seize the opportunity:Exploiting Web aggregation. MISQ Executive 1, 1(2002) 35-46.

9. ORourke, M.A. Is virtual trespass an apt analogy?Comm. ACM 44, 2 (2001), 98-103.

10. Samuelson, P. Legal protection o database contents.Comm. ACM 39, 12 (1996), 17-23.

11. Tabuchi, H. International Protection o Non-OriginalDatabases: Studies on the Economic Impact o theIntellectual Property Protection o Non-OriginalDatabases. CODATA (2002), Montreal, Canada;http://www.codata.org/codata02/03invited/Tabuchi/Tabuchi_CODATA_ejournal.pd.

12. Zhu, H., Madnick, S.E., Siegel, M.D. An economicanalysis o policies or the protection and reuse o non-copyrightable database contents. J. of ManagementInformation Systems 25, 1 (2008), 199-232.

Hongwei Zhu ([email protected]) is an assistant proessoro Inormation Technology in the College o Business &Public Administration at Old Dominion University, Norolk,VA.

Stuart E. Madnick ([email protected]) is the JohnNorris Maguire Proessor o Inormation Technologiesin the Sloan School o Management and Proessor oEngineering Systems in the School o Engineering at theMassachusetts Institute o Technology, Cambridge, MA.

2009 ACM 0001-0782/09/0900 $10.00

02-12 one size does not fit all_legal protection for non-copyrightable data

Documents