polishing and promoting the incubator

3 views
Skip to first unread message

Arthur Kalmenson

unread,
Oct 10, 2008, 4:15:24 PM10/10/08
to Google Web Toolkit Contributors
Hello everyone,

There was some discussion on the IRC channel about this, and I figured
I'd put it up for everyone here.

I think that the incubator has a lot of useful things (FooBundle,
CssResource, DatePicker), a lot of great ideas (Declarative UI) and
some not so great things (PagingScrollTable). The incubator has a lot
of potential, it just needs some polish and promotion. I think it
needs the following:

1. Better and nicer examples of widgets and use of some nice CSS (just
use one of the GWT themes).
2. Clean up the existing widgets and make them easier to use.
3. A showcase to show off all these widgets and make them accessible
like the main GWT showcase.
4. Promoting incubator as a great resource on the regular GWT group
and on the main GWT site.

As it stands right now, few people know about the incubator and those
that do are usually scared away. Most people end up going to widget
libraries like ExtGWT and others. Just about every widget library I've
seen is poorly done and ends up giving GWT a really bad name (ExtGWT
especially). I think the incubator can offer a much better and cleaner
widget library and components, but it needs to be cleaned up and
promoted so more people contribute and give feedback.

What do you think?

Regards,
Arthur Kalmenson

Isaac Truett

unread,
Oct 10, 2008, 4:36:12 PM10/10/08
to Google-Web-Tool...@googlegroups.com
We don't want people to be afraid of the Incubator, but we do want them to be cautious. It isn't a library per se, but a workshop for ideas. Things in the Incubator may be half finished or in the middle of refurbishing at any given time. It's an environment where you may have to get your hands dirty. It's not something that all GWT users will have the patience or risk tolerance for.

Improving the contents of the Incubator is, of course, important. That's what it's there for. But I would expect that as things are "cleaned up" they'll be promoted to the main GWT project or spun off into separate libraries. What's left in the Incubator will always be works-in-progress.

Arthur Kalmenson

unread,
Oct 10, 2008, 4:57:27 PM10/10/08
to Google Web Toolkit Contributors
OK that makes sense. Maybe we need a project that's in between
incubator and GWT, something that has regular releases and uses
polished content from the incubator but content that's not so polished
that it would go into GWT. While what's in the incubator is a work in
progress, it is still a) much faster then what the other libraries
offer, b) mostly cleaner and better written then the other libraries,
c) gives people an idea of where GWT is headed.

While making custom widgets is easy in GWT, there is a lot of overlap
in what people need and there is usually a rich suite of widgets that
people would like out of the box. It's not very easy to pick up GWT
and dive in making great and interactive apps. There's a lot of group
work that needs to be done to build rich widgets. It seems a waste
that this group work is done on every GWT project by all the various
organizations that use GWT. I know that the GWT team said GWT was
meant to be pretty low level and they wanted the community to build on
top of that, but it seems that the community is not delivering. I know
Bruce mentioned that better widgets are on the agenda, so I don't
know....

Regards,
Arthur Kalmenson

On Oct 10, 4:36 pm, "Isaac Truett" <itru...@gmail.com> wrote:
> We don't want people to be afraid of the Incubator, but we do want them to
> be cautious. It isn't a library per se, but a workshop for ideas. Things in
> the Incubator may be half finished or in the middle of refurbishing at any
> given time. It's an environment where you may have to get your hands dirty.
> It's not something that all GWT users will have the patience or risk
> tolerance for.
> Improving the contents of the Incubator is, of course, important. That's
> what it's there for. But I would expect that as things are "cleaned up"
> they'll be promoted to the main GWT project or spun off into separate
> libraries. What's left in the Incubator will always be works-in-progress.
>

Emily Crutcher

unread,
Oct 10, 2008, 6:02:52 PM10/10/08
to Google-Web-Tool...@googlegroups.com
I actually tend to agree with Authur here, that in gwt-incubator we have widgets at different stages in the pipeline and while some we don't want users to touch, others are much more stable, and would benefit from wider user distribution and are at least as good, or better, then almost any other available gwt library.  

As widgets are moved into gen2, they are being all given default style sheets though improvements are always appreciated.

Internally, we've started to toy with the idea of labeling widgets/libraries: Here is the initial proposal
  1. @ReleaseCandidate: This widget has been slotted for a GWT release. Its bug reports are taken as seriously as any normal gwt widgets and its API is less likely to change. 
  2. @Beta:  This is a widget intended for wide distribution and is being actively supported by at least one developer. The code should be reasonably bug free as well. 
  3. @DoNotUse or no tag: Don't use, is either in the process of being dropped from incubator or is still in alpha. 
 
Also, here is my wish list of things I'd love to get volunteers for:
  • A maven repository to pick up the most recent gwt-incubator + gwt-trunk milestone build. If possible, would also create a nightly build as well. 
  • Some sort of  ant/python/shell/ script that crawls through the new src-demo and automatically builds all the  demos there.  We could then use that  to easily publish on google app engine up-to-date demos for all incubator widgets before each public drop.
  • Someone to go through the documentation, fix what can be easily fixed, and then e-mail out those widgets/libraries whose documentation need serious attention.
  • Also, someone who would be willing to go through the current demos and improve them/flag developers if they are missing and/or just suck.
  • Finally,  if people could nominate widgets they would like to see leave incubator, that would be very helpful as well!
--
"There are only 10 types of people in the world: Those who understand binary, and those who don't"

Emily Crutcher

unread,
Oct 10, 2008, 6:43:42 PM10/10/08
to Google-Web-Tool...@googlegroups.com
  • Also, someone who would be willing to go through the current demos and improve them/flag developers if they are missing and/or just suck.
I forgot to mention that all demos must still only contain client side code, as the plan is to make it as automatic a process as possible, and trying to deploy to n python back-ends as an automated process is not a challenge I want to take on!


 

Isaac Truett

unread,
Oct 11, 2008, 11:50:17 AM10/11/08
to Google-Web-Tool...@googlegroups.com
> Some sort of  ant/python/shell/ script that crawls through the new src-demo and automatically
> builds all the  demos there.  We could then use that  to easily publish on google app engine 
> up-to-date demos for all incubator widgets before each public drop.

I'll volunteer for that. This is pretty easy to do in an Ant script. Building the demos will be trivial if issue #2970 is accepted and implemented. I'm not familiar with deploying to GAE. Would a zip file of all the compiled demos, plus a simple index.html, suffice?

Emily Crutcher

unread,
Oct 11, 2008, 12:17:00 PM10/11/08
to Google-Web-Tool...@googlegroups.com
That would be awesome!  The zip isn't even necessary, as long as the output represents a predictable static file structure.  

My guess is that 2970 will probably not be implemented right away.  You're right about ant,  so if you like python or want to create a specialized ant task, that would work.

In fact, the script I'm using to upload static content to the Google app engine is already in python, so I'm going to have to deal with python/ant integration anyway.

Also, in an ideal world we'd some day also generate a master index page as well,  so having the base demo-generator be in an expressive language definitely has some advantages.

Thanks a million!

               Emily

Isaac Truett

unread,
Oct 11, 2008, 4:26:11 PM10/11/08
to Google-Web-Tool...@googlegroups.com
So, the compile-with-file-name issue turned out to be a little more
complicated than my first assessment. I've added a GwtCompilerTask for
Ant that accepts a module file path, but it also needs to know the
source root. I think I could get around that by starting with the
module file name and prepending elements from the absolute path until
I find something that exists on the classpath. That'll wait for
another day, since what I have works for now.

The build.demos target and GwtCompilerTask are available in r1130.

Emily Crutcher

unread,
Oct 11, 2008, 7:24:42 PM10/11/08
to Google-Web-Tool...@googlegroups.com
Terrific!

Isaac Truett

unread,
Oct 12, 2008, 10:13:35 AM10/12/08
to Google-Web-Tool...@googlegroups.com
Oh, I forgot to mention one thing. I found yesterday that compiling
the PagingScrollTableDemo caused a stack overflow. The other demos
were fine. You might need to bump up your stack (-Xss) or rename
PagingScrollTableDemo.gwt.xml so that it won't compile.

Emily Crutcher

unread,
Oct 12, 2008, 2:27:44 PM10/12/08
to Google-Web-Tool...@googlegroups.com, John LaBanca
We'll look into that, as compiling seems like something we really want each demo to do correctly!

Arthur Kalmenson

unread,
Oct 12, 2008, 8:54:08 PM10/12/08
to Google Web Toolkit Contributors
Hello Emily,

> Internally, we've started to toy with the idea of labeling
> widgets/libraries: Here is the initial proposal
>
> 1. @ReleaseCandidate: This widget has been slotted for a GWT release. Its
> bug reports are taken as seriously as any normal gwt widgets and its API is
> less likely to change.
> 2. @Beta: This is a widget intended for wide distribution and is being
> actively supported by at least one developer. The code should be reasonably
> bug free as well.
> 3. @DoNotUse or no tag: Don't use, is either in the process of being
> dropped from incubator or is still in alpha.

That sounds like a great idea, we could probably make at least
@ReleaseCandidate widgets available to a wider audience.

> - A maven repository to pick up the most recent gwt-incubator + gwt-trunk
> milestone build. If possible, would also create a nightly build as well.

Yes please! I'm a Maven person and I know there are a lot of Maven
users out there. A nightly gwt-incubator for Maven would be very
welcome!

> - Also, someone who would be willing to go through the current demos and
> improve them/flag developers if they are missing and/or just suck.

I can try to improve some of them, but I can definitely say which ones
need to be improved and which aspects. Should I start another post so
people can comment on which demos need improvement and what needs to
be improved?

> - Finally, if people could nominate widgets they would like to see leave
> incubator, that would be very helpful as well!

This should be very easy to do with a Google Docs Form. Do you want me
to create one?

Regards,
Arthur Kalmenson

On Oct 10, 6:02 pm, "Emily Crutcher" <e...@google.com> wrote:
> I actually tend to agree with Authur here, that in gwt-incubator we have
> widgets at different stages in the pipeline and while some we don't want
> users to touch, others are much more stable, and would benefit from wider
> user distribution and are at least as good, or better, then almost any other
> available gwt library.
>
> As widgets are moved into gen2, they are being all given default style
> sheets though improvements are always appreciated.
>
> Internally, we've started to toy with the idea of labeling
> widgets/libraries: Here is the initial proposal
>
>    1. @ReleaseCandidate: This widget has been slotted for a GWT release. Its
>    bug reports are taken as seriously as any normal gwt widgets and its API is
>    less likely to change.
>    2. @Beta:  This is a widget intended for wide distribution and is being
>    actively supported by at least one developer. The code should be reasonably
>    bug free as well.
>    3. @DoNotUse or no tag: Don't use, is either in the process of being
>    dropped from incubator or is still in alpha.
>
> Also, here is my wish list of things I'd love to get volunteers for:
>
>    - A maven repository to pick up the most recent gwt-incubator + gwt-trunk
>    milestone build. If possible, would also create a nightly build as well.
>    - Some sort of  ant/python/shell/ script that crawls through the new
>    src-demo and automatically builds all the  demos there.  We could then use
>    that  to easily publish on google app engine up-to-date demos for all
>    incubator widgets before each public drop.
>    - Someone to go through the documentation, fix what can be easily fixed,
>    and then e-mail out those widgets/libraries whose documentation need serious
>    attention.
>    - Also, someone who would be willing to go through the current demos and
>    improve them/flag developers if they are missing and/or just suck.
>    - Finally,  if people could nominate widgets they would like to see leave
>    incubator, that would be very helpful as well!
>

Isaac Truett

unread,
Oct 12, 2008, 9:09:53 PM10/12/08
to Google-Web-Tool...@googlegroups.com
I would suggest opening issues for demo todo items and graduate nominations.
Reply all
Reply to author
Forward
0 new messages