Wednesday, August 24, 2016

Re: GWT 2.8.0 RC2 is here!

Tnx Thomas for this detailed and valuable answer.

For software creators of my generation (born 1956), the immensely attractive argument of GWT was : "you can create WEB apps (complex and powerfull web apps) by just using the best langage you ever used : Java, and all the formal benefits behind Java (OO thinking was the key turn in software history, and Java is the most usable effect of that). 

THAT ("Only Java") made GWT a brillant approach. Developers like me are not interested in technologies, langages, etc... per se: we just want the most powerful tools to turn conceptualization process into running software, with minimal technology concerns. Learning new technogies is "wasted time" if it does not help on that.

Hearing that 3.0 will be a "bundle", our key concern is "How many technologies should we (learn?) mix and organize to work together?". Knowing from my experience that technology mixing => unpredicable inconsistencies incompatibilities, probable interferences, or unclear bug responsability, I am rather skeptical about this evolution.

I see GWT as a promising path facing now a dead end road. Disappointing. Just hoping that the 2.8.x branch will survive  10 years....

Anyway thanks again for your explanations Thomas.


On 08/23/2016 04:15 PM, Thomas Broyer wrote:

On Tuesday, August 23, 2016 at 3:39:26 PM UTC+2, Philippe Gonze wrote:
"the fact that 2.8 will be maintained in parallel. (Jens)  " !?!

We would certainly appreciate to know more about this fact (announced? published? where?).

(and you can read echoes of it many times here since then)
As a matter of fact, if version 3.0 is in line with various statements read here and there on the web (disparition of widget library, disparition of RPC), we would certainly prefer version 2.8.x to any 3.dead versions !
We think the 2.8 branches will continue to attract more developers than the 3.x branches.

But in any case, the first need is a fair vision of the future of GWT. Something like an 'official' statement, or at least a target roadmap.

There's no such thing yet.
If both branches are maintained (hopefully), we suggest a better version naming. Version 3.0 should receive a radically different name, and version 3.0 should succeed to 2.8 on the 'classic' branch...

Reality is that most of the work is done by Google, and Google wants to move on to a new compiler (J2Cl). If Google no longer works on the GWT 2.x compiler, GWT 2.x is going to die, as I can't see anyone putting enough energy to make it live.
Because "GWT 3.0" is going to be based on J2Cl, which removes GWT.create(), this is going to be a breaking change no matter what. This is why it'd been decided that the 2.x branch would be maintained in parallel to the 3.x branch, but it'll likely only remain a "maintenance branch" (understand: bug fixes, but I believe you can kiss goodbye to any JDT upgrade, so no Java 9 for you; and probably no new Java 8 emulated APIs –think java.time et al. unless they come before real work on 3.x has begun), mostly there so that people have time (2 years? maybe more) before switching to GWT 3.0.
That being said, "GWT 3.0" will likely be a "bundle" of various projects (similar to the Eclipse bundles released every year): J2Cl, the Java Runtime Emulation library (unless it comes with J2Cl), Elemental, etc. see also!topic/google-web-toolkit-contributors/uhSgR5CWBK8 (as you can see, nothing's clear yet, so no "official statement" or "target roadmap"), and as such Google likely will never use "GWT 3.0" per se, that one really being a community-lead project.
For now, Google needs to make an MVP of their new compiler and make its sources public; then only we can start talking about what GWT 3.0 might look like, built around that new compiler, and start really testing things against it; in the meantime all we can do is handwave, throw FUD, or more constructively prepare for the inevitable demise of GWT.create(), particularly for those features that Google does not use themselves (they will "port" the ones they use), and all third-party projects. Work has begun already for a few of these things.
You received this message because you are subscribed to a topic in the Google Groups "GWT Users" group.
To unsubscribe from this topic, visit
To unsubscribe from this group and all its topics, send an email to
To post to this group, send email to
Visit this group at
For more options, visit


Philippe Gonze
Address: Rue de la bruyère 12 / 1428 Lillois / BE
Phone : 00.32.473580758
Skype : matscape

No comments:

Post a Comment