The future of GWT?

The future of GWT?

Last month I was researching options for writing multi-platform mobile apps in Java. Beside our own framework Tabris the Google Web Toolkit is an obvious option. When I started to dig a little deeper I stumbled upon Google’s announcement at Google IO 2012 to go from “gatekeeper to peer” regarding the development of GWT. It was not really a formal announcement, rather it had been mentioned as part of a presentation on “The history and future of GWT“. The announcement did not contain a lot of tangible information beside the introduction of a steering committee and its members. Now six weeks have passed since the announcement and there is still only very little information available. The steering committee website is still not live (as of Aug 14, 2012) and the associated mailing list is exposing pieces of information without providing enough context. So what is going on?

GWT logoSome people compared Google’s decision to release GWT into the open source community to IBM’s decision regarding Eclipse in 2001. Eclipse transition into Open Source was extremely well managed and a clear vision and a dedicated management team was established at the very beginning. When I became member of the Eclipse Board of Stewards in 2002 I was impressed by the breadth of companies and the resourcefulness of its members. And its vibrant community of committers, even if most of them were employed by a single company – IBM. Comparing this to what’s happening now with GWT it looks more like Google is dumping GWT. Or the transition has been managed very poorly, but that is somewhat hard to believe. There is no visible leadership in the steering committee and everybody seems to be busy with other more important things. While there is nothing wrong with this in general it makes it hard to assess the viability of GWT. At Eclipse the “Open Source Rules of Engagement” have proved to work really well when you are trying to establish an open source community (openness, transparency, meritocracy, permeability and vendor neutrality). Especially when uncertainty is starting to spread it is a good idea to have open communication.

So what does that mean for GWT’s future? From the information that has been made available so far it seems that Google has quite a few applications that rely on GWT so they will have to support it for the next little while. There is no reason to worry if you have GWT based applications in production. IMHO it is a different topic to start new projects based on GWT. Especially the discussion on the introduction of “vendor branches” is a bit disturbing. Google will have its own branch of GWT on which they enforce rigorous testing procedures based on internal applications (that they can’t make available). And they are not sure if they have the resources to validate all changes on trunk. This hints more into the direction of maintenance than active development. But maybe some of the other members of the steering committee will step up and lead GWT into the wild. In the meantime we will have to wait and hope that it is only a lack of communication and not a lack of activity.

Resources:
[1] http://developer.eclipsesource.com/tabris/ – Tabris – multi-platform mobile app framework
[2] https://docs.google.com/presentation/d/1pC… – History and future of GWT
[3] https://docs.google.com/document/d/1oG… – Steering Committee meeting #3 minutes
[4] http://www.eclipse.org/projects/dev_process/development_process_2011.php – Eclipse development process
[5] http://sites.google.com/gwt-steering – GWT steering committee website
[6] https://groups.google.com/forum/?fromgroups#!forum/gwt-steering – GWT steering committee mailing list
[7] https://groups.google.com/forum/?fromgroups#!topic/gwt-steering/O3Yl25n16Xs%5B1-25%5D – Discussion on vendor branches for GWT

2 Comments
  • Zen Balancer
    Posted at 5:23 pm, August 26, 2012

    To clarify and update as an enthusiast: 2 days later after this well-put article –whether any direct connection or as a follow-up to the gwt-public-mood online– the GWT steering committee mailing-list showed their energy with a clear pulse.

    “I am actually getting tired of this GWT is going to die nonsense. It is very clear from Ray’s Future of GWT talk from Google IO, that GWT is a very viable project inside of Google as well as outside.
    I think the best think the committee can do right now is to pick up pace and prove those doubts wrong.

    I think the one thing we can improve from the committee is the visibility of our meetings. The next meeting is scheduled to take place on google+ hangouts maybe we can do those meetings with a live hangout as well. I will bring that up on the next meeting.”

    Said Daniel Kurka (mgwt & gwt-phonegap), another appreciated and influential GWT steering committee member.
    http://www.m-gwt.com/
    https://code.google.com/p/gwt-phonegap/
    http://www.daniel-kurka.de/

    “I commented on it in another G+ post. I said in my Google I/O
    presentation it would take several weeks, not the least of which was
    the fact that people were on vacation (I was gone for 3 weeks after
    I/O) and there are other things we are working on (GWT 2.5 final,
    recent Java7 fixes, etc) plus decisions on hosting and continuous
    builds. I realize it’s a frustrating process and wish it could go
    quicker, but nothing has changed in terms of GWT and Google’s use of
    it.”

    Said the GWT Lead Ray Cromwell.

    https://groups.google.com/forum/?fromgroups=#!topic/gwt-steering/qO9MW9lSL5Y

    Although the pace is not that clear at first glance, people inside are very serious about the health & direction of GWT. Google’s disengaged look regarding GWT before GIO12 to the outside world was because their focus was on GWT for their own business first, where the cash flows to keep the inside GET-oriented developers among others well paid. The users and the public can count on these people. Will have to wait and see to be sure, perhaps.

  • Zen Balancer
    Posted at 5:52 pm, August 26, 2012

    That said, I wonder about Tabris & RAP (Eclipse’s “Open Source browser client”), looks also very promising.

    “Currently Tabris provides native clients for iOS and Android plus an Open Source browser client which runs on all major web browsers, without requiring any add-ons. ”

    Will take a look there too.
    http://eclipse.org/rap/

    “Tabris & RAP”

    ~’vs’

    “mgwt/gwt-phonegap & PhoneGap”?

    It seems if one is comfortable with SWT, then Tabris is the way to go. It’s interesting to look into both further in-depth, and decide based on the project and the team.