java ee productivity report 2011 finalv2

Upload: dummy4mail

Post on 06-Apr-2018

226 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    1/14

    JAVA E

    What happens when overenvironment?

    Last year, ZeroTurnaround Oapplication containers in the Jsharing info about their develoresults helpful.

    This year we expanded on that, aaccurate information. The survey

    and those responses were analy

    What is more popular, Ant

    Which Java IDE is the mo

    What Java EE standards

    Which Java Framework is

    What Java Container / Ap

    How much development ti

    We asked the people we met at

    promoted the survey in the medi

    they would be notified again onc

    chance to win free JRebel licensare published.

    A total of 1027 responses were a- they were requested to provide

    including all calculations is availa

    ZeroTurnaround O

    Productivity Report 2011

    000 Java developers compare the

    published a reporti on turnaround time,va ecosystem. Over 1300 Java develop

    pment environment, and over 40,000 pe

    sking more and better questions to give yoitselfiiprovided data from over 1000+ Java

    ed to create this report. We set out to disc

    or Maven?t popular?

    re used the most?

    most prevalent?

    Server is used the most?

    e is spent redeploying Java containers?

    avaOne 2010, emailed previous survey res

    . They were sent an email with the link to t

    the results were made available. Participa

    s and one recipient will receive an iPad on

    nalyzed to create this report. Of these, 28%data from before they were purchasers. Th

    ble for downloadiii.

    2010. All rights reserved.

    ir development

    ools andrs ended upple found these

    better and moreEE developers

    ver:

    pondents and

    e survey and told

    ts were offered a

    e the final results

    are JRebel usersraw data

    http://www.zeroturnaround.com/survey2010/http://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/java-ee-productivity-report-2011/http://www.zeroturnaround.com/survey2010/http://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/survey2010/http://www.zeroturnaround.com/java-ee-productivity-report-2011/
  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    2/14

    Build Tools and IDEs Se

    What is more popular, Ant or

    We asked everyone to mark the t

    As you can, see Maven and Ant

    both. Clearly both of them are usmentioned, but none of them gat

    ZeroTurnaround O

    ting the scene

    aven?

    ools they use for building their application:

    re used almost equally and some of the re

    ful in some way. Tools like Gradle, Ivy andered even 5 votes.

    2010. All rights reserved.

    pondents use

    SBT were

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    3/14

    Which Java IDE is the most po

    The IDE results show a much wi

    In terms of user base, Eclipse ha

    respondents using IntelliJ IDEA (

    separately and not doing the sam

    idea, that's something we'll do be

    chart Oracle's [ex]-flagship IDE,

    independent player MyEclipse.

    ZeroTurnaround O

    ular?

    er distribution of preferences:

    s emerged here as the clear leader, with on

    2) and Netbeans (#3). Perhaps including

    e for other Eclipse distributions like RAD w

    tter next year. It is interesting that accordin

    Developer, is even more marginal than the

    2010. All rights reserved.

    ly one-third of

    yEclipse

    s not the best

    to this

    small

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    4/14

    Servers & Frameworks

    Which Java Containers / App S

    Compared to the last years chart

    also see that Jetty and Tomcat h

    bit. The results from these 1000+

    WebSphere have lost a total of 8last year.

    We asked people to choose onlycompanies several containers arcontainer do people spend most

    phrase it like that. So don't make

    deployment, but it's a good estim

    ZeroTurnaround O

    ho are the popular kids at school

    ervers are used the most?

    ivwe see % gains for the open source cont

    ve a bigger share than last year, while Gla

    developers shows that Oracle WebLogic a

    of the market to the open source contain

    one primary container in this question, evein use. Since we were interested first of allf their time in development it just made mo

    any assumptions as to how this applies to t

    ate for the situation in development.

    2010. All rights reserved.

    iners. We can

    ssfish is sliding a

    nd IBM

    rs compared to

    though in manyl, in whatre sense to

    he production

    http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/
  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    5/14

    What Java EE standards are u

    As far as Java EE standards are

    Here the market penetration is malmost as much as the venerable

    altogether have 39%, which woul

    some overlap between EJB2 and

    coming CDI standard has gainedchange next year. We'll cover JS

    ZeroTurnaround O

    ed the most?

    concerned we have the following picture:

    ore important than the comparison aspect.JSP, with 37% of market penetration. EJB

    d make them the most popular standard, b

    EJB3 users, so the actual total is likely a bi

    6% so far, will be very interesting to see hoin the next section.

    2010. All rights reserved.

    PA is usedversions

    t there's likely

    t smaller. Up-and-

    w much this will

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    6/14

    Which Java Framework is mos

    Let's take a look at the framewor

    Spring and Hibernate are by far tstandards. However, as far as w

    of answers.

    Unfortunately, we havent separayou must) that it is at least as po

    or less (GWT is barely over that

    The third most popular frameworasked Matt Raiblev if these result

    presentation Comparing JVM We

    what he had to say:

    "It's surprising to see that

    Framework Matrix results

    the top JVM Frameworks,

    ZeroTurnaround O

    t prevalent?

    chart:

    e most popular and in fact are still used mb frameworks go, JSF looks like a popular

    ted Spring MVC from Spring, but we assumular as JSF. The rest of the frameworks ho

    nd is the third most popular).

    in our survey was GWT, followed closelys aligned with his recent controversial

    b Frameworks presentation from Devoxx 2

    ramework popularity closely aligns with the

    calculated at Devoxx. I had Spring MVC a

    along with Ruby on Rails (running on JRub

    2010. All rights reserved.

    re thanchoice with 24%

    e (interpret this asld a share of 10%

    y Struts 1. We

    10vi and here's

    JVM Web

    d GWT listed as

    of course).

    http://raibledesigns.com/http://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/
  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    7/14

    "While this matrix was the

    decent technique for choo

    web framework is to picksatisfies your needs to be

    "More than anything, makthat's likely to be key in th

    "I'm surprised to see so m

    companies choose it beca

    say that Struts 1 and Spri

    However, just because fra

    use them. There's other gTapestry come to mind) th

    "I think the biggest thing thREST support. I think mor

    (with frameworks like jQue

    embrace that. I especially

    same classes. This allows

    ZeroTurnaround O

    result ofmuch controversyvii

    , I think it gives

    ing a web framework to use. Of course, th

    few you like, prototype with them and seet.

    sure the developers on your team like devir productivity using it.

    ny folks using JSF, but I do understand w

    use it's such a "safe" choice as a standard.

    g are pseudo-standards based on their po

    meworks are popular doesn't mean you sh

    eat component-based frameworks in Javaat are easier to use than JSF."

    at component-based frameworks need toand more components are being develop

    ry, ExtJS and SproutCore) and web frame

    like frameworks that allow emitting JSON a

    for easy web page and API development a

    2010. All rights reserved.

    developers a

    way to choose a

    which one

    eloping with it, as

    y many

    In fact, you could

    ularity.

    uld automatically

    GWT, Wicket and

    ork on is theird on the client

    orks should try to

    d HTML from the

    t the same time."

    http://raibledesigns.com/rd/entry/how_i_calculated_ratings_forhttp://raibledesigns.com/rd/entry/how_i_calculated_ratings_for
  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    8/14

    Turnaround Time - How di

    This year we asked questions ab

    redeploys per hour. We also askhappens. We specifically asked e

    were glad to see the developmen

    "I wait till I have enough c

    like this, it's definitely one

    artificial way of coding with

    "We use tests (unit tests, iespecially if you are not d

    that the delivered system i

    "We use Eclipse/Sysdeo/I

    quickly this is a great wayevery time they occur. Unf

    minutes. "We use HotSwap/Eclipse

    mechanism is a great wayAlso, if you can't use explo

    JSP/HTML/etc.

    "We use OSGi" -- only a c

    application redeploy can breinitialized.

    "We use Grails, Play! or T

    that these extremely prod

    "We use JRebel" -- the morespondents were JRebel

    We originally did not include a 3

    your application? However, we hhalfway through the responses. T

    for the 1 minute answer.

    ZeroTurnaround O

    ferent J2EE environments compar

    out the time it takes to redeploy the applicat

    d those who don't have to redeploy to comveryone to provide the answers without u

    t in this area since last year:

    anges to warrant a redeploy" -- we've got a

    f the ways around the problem, but it can l

    little incremental feedback.

    tegration tests, etc)" -- if you have enoughing UI), it's a great way to both avoid redep

    s stable and correct. Kudos to those who d

    tellij IDEA/etc HotDeploy" -- for application

    to improve turnaround, by making the IDE rortunately for bigger applications it saves cl

    Debugger Session/etc" -- The built-in class

    to make small changes, but limited only toded deployment, you still have to redeploy

    uple of answers like that, but in a well desi

    e very quick, as only the changed bundle n

    pestry 5" -- a few answers for each of thos

    ctive frameworks are being used.

    st popular answer, but not surprising consiusers :)

    sec answer to the question How long do

    ave since received a few comments abouto compensate for that, we will use a 48se

    2010. All rights reserved.

    e

    ion and number of

    ment on why thising JRebel. We

    few comments

    ad to a very

    tests (andloys and ensure

    this!

    s that deploy

    epublish changesicks, but not

    replace

    method bodies.for any change in

    gned OSGi

    eds to be

    e, great to see

    ering that 28% of

    s take to redeploy

    hat, and fixed it (0.8 min) value

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    9/14

    How long does it take to restar

    This was the first question we as

    We didn't ask with such accuracytrends have stayed the same. Th

    deviation is 2.8, which means tha

    It can be noted that a statisticallydevelopers) responded that it tak

    ZeroTurnaround O

    t your container and redeploy your app?

    ed. The answers are distributed below:

    last year, which makes it harder to compare average redeploy time is 3.1 minutes, but the redeploy time varies greatly.

    significant segment of respondents (just oves over 10 minutes to fully redeploy.

    2010. All rights reserved.

    e, but it seems thet the standard

    er 1 in every 10

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    10/14

    In an hour of coding, how man

    The average frequency is 4 time

    With these two questions answerredeploying in an hour. We bega

    reported redeploying more than 6

    ZeroTurnaround O

    times do you redeploy?

    an hourwith the standard deviation of3.2

    ed, we can make a reliable estimate for theby removing those who don't redeploy at

    0 minutes an hour (ummm.what?)

    2010. All rights reserved.

    .

    total time spentll and those that

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    11/14

    How much time per hour do yo

    The average respondent spends

    deviation of9.8. This is about 17.day as "coding time", and assum

    average 5.3 full, 40-hour workwe

    ZeroTurnaround O

    u spend redeploying?

    about 10.5 minutes an hourredeploying w

    5% of total coding time. If you consider 5 h4 weeks out of each year are vacation tim

    eks per year are spent exclusively redeplo

    2010. All rights reserved.

    ith a standard

    urs out of eache, then on

    ing and restarting.

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    12/14

    Redeploy time per Java EE Co

    Finally, we can also see how the

    Note: This doesn't mean that big

    influence the choice of container:

    Not many changes from last year

    v3.

    ZeroTurnaround O

    tainer (as a % of overall coding time)

    choice of container correlates to the time s

    er containers are that much slower, rather

    here, although it's nice to see the improve

    2010. All rights reserved.

    ent redeploying.

    bigger apps

    ents in Glassfish

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    13/14

    Conclusions?

    Interpreting data is always a hard

    reasonably large. But some bias i

    to answer.

    The data gives a fair overview of

    the most interesting thing is that,

    commercial software manage to

    Of course with companies like JB

    distinguish the two, but commerci

    Sonatype backing Maven). It will

    upcoming years:

    Whether it will be the vend

    will find the most use. How much the Open Core

    and their commercial back

    Will the closed-source con

    diminish or increase?

    Come back next year, let's see w

    ZeroTurnaround O

    and dangerous task. A sample of 1027 de

    is possible, as it's likely the more active dev

    what the Java development ecosystem loo

    despite what some say, the world of open s

    oexist quite well.

    oss and SpringSource it's becoming increa

    ial interests are present on every single gra

    be interesting to see how this balance deve

    or-backed standards or community-backed

    will be adopted by projects like Jetty, Tomc

    ers.

    tainers and tools survive in the ecosystem?

    hat happens!

    2010. All rights reserved.

    elopers is

    elopers bothered

    s like. Probably

    ource and

    singly difficult to

    ph (if you include

    lops in the

    frameworks that

    at and GlassFish

    Will their role

  • 8/2/2019 Java EE Productivity Report 2011 Finalv2

    14/14

    Footnotes and URLs

    ihttp://www.zeroturnaround.com/blog/jaii

    http://www.zeroturnaround.com/surve

    iiihttp://www.zeroturnaround.com/docs/ivhttp://www.zeroturnaround.com/blog/jvhttp://raibledesigns.com/vihttp://raibledesigns.com/rd/entry/my_vii

    http://raibledesigns.com/rd/entry/how

    ZeroTurnaround O

    va-ee-container-redeploy-restart-turnaround-report/

    2010/

    ava_ee_productivity_report_2011_rawdata.zip

    ava-ee-container-redeploy-restart-turnaround-report

    omparing_jvm_web_frameworks

    i_calculated_ratings_for

    2010. All rights reserved.

    /

    http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://www.zeroturnaround.com/survey2010/http://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://raibledesigns.com/http://raibledesigns.com/http://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/rd/entry/how_i_calculated_ratings_forhttp://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://www.zeroturnaround.com/survey2010/http://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/rd/entry/how_i_calculated_ratings_forhttp://raibledesigns.com/rd/entry/how_i_calculated_ratings_forhttp://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://raibledesigns.com/rd/entry/how_i_calculated_ratings_forhttp://raibledesigns.com/rd/entry/my_comparing_jvm_web_frameworkshttp://raibledesigns.com/http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/http://www.zeroturnaround.com/docs/java_ee_productivity_report_2011_rawdata.ziphttp://www.zeroturnaround.com/survey2010/http://www.zeroturnaround.com/blog/java-ee-container-redeploy-restart-turnaround-report/