Future of GWT-ext

[michaelneale]

Due to the unfortunate-ness of ext future versions, Sanjiv, the author of gwt-ext has been looking into a new JS based framework: http://www.jroller.com/sjivan/entry/smartgwt_is_arriving_soon (SMartGWT) - ideally providing a migration path for gwt-ext.

Ideally we would use native GWT as much as possible though.

[heiko.braun]

Ideally we would use native GWT as much as possible though

Why? The extension are a huge time saver and add improved user experience.

[michaelneale]

Yes the extensions are a timesaver, but they are not necessarily as well supported or as performant as GWT, or as compatible across browsers. GWT is very well supported and developed thanks to google, the other extensions may not be.

There is also the whole ext licence debacle which went on, which kind of left a bad taste in my mouth.

Note that I am mainly suggesting when there is a GWT widget that is good enough, use it, as generally they work together well (eg text box).

[cbrock]

You won't think Ext is a timesaver if ExtJS Inc. slaps you or a customer with a copyright infringement suit.

[heiko.braun]

Both gwt-ext 2.0.5 and ext 2.0.2 use LGPL 3.0. What's the problem with that?
I think you refer to later versions which has change in their license model.

[michaelneale]

Interesting - I like that use of REST. If it works out for you, I would consider ditching GWT RPC as like you say, it encourages re-use.

The GWT RPC style interface is pretty close to rest in semantics already - so its probably not terribly hard to switch over.

I think there is a performance cost - however its probably not an issue other then for complex data structures or circular references. But in any case REST can encapsulate DTO serialization using some other mechanism. Not a huge deal.

[michaelneale]

Whoops I posted the wrong thread (annoying forums interface, another reason I prefer mailing lists).

There is some dispute over the licensing of all the elements in that version. Mainly the company behind it makes angry noises - generally not pleasant for all involved.

Mike can point to some other prior articles on the web. The Gwt-ext community is alive and well though.

[kukeltje]

Both gwt-ext 2.0.5 and ext 2.0.2 use LGPL 3.0. What's the problem with that?
I think you refer to later versions which has change in their license model.

He probably does. 2.0.x (x>2) use GPL, a commercial license or an OEM license. But what does that give us for the future? Are we stuck with 2.0.2?

[marklittle]

I think our future is literally in our own hands with vanilla GWT as the basis, which is what Mic and Mike have been saying. There is far too much legal complexity/haziness around GWT-ext at this point for us to to consider going further with it. It's a shame, but that's the direction those guys have decided to take.

[michaelneale]

Well the nice thing about GWT is people can choose to use what they want - you can mix and match without js namespace hassles (as the GWT compiler takes care of it) - so if there is a widget that someone really wants in a toolkit, they can have it.

[heiko.braun]

He probably does. 2.0.x (x>2) use GPL

You are talking about extJs. For GWT-Ext it's 2.0.5.
Yes that's right, we have to stick with that.

At least until we have an alternative. But I'd more likely stick to that one for now, because it saves my time. However it's on my roadmap, that after GA I'll start replacing it with whatever we chose and help pushing an open source implementation like the one Mike chose.

[heiko.braun]

Just for the records: The GWT console has been ported to gwt-mosaic.
See

http://relative-order.blogspot.com/2009/04/bpm-console-springtime-cleaning.html