Does it really still make sense to use ExtGWT for your next big project ??

[mxhn]

This is JUST UNFAIR!! I hope ExtGWT Sencha team has a way to make users like me feel better about ever thinking to include ExtGWT in their next big project.

Some time ago, I've asked on this forum about a comparison between ExtGWT and ExtJS. I've got a fairly positive response that ExtGWT is good to go and it won't have differences in performance than ExtJS. Since then, we've thought a lot about structure, how would ExtGWT fit in, etc. That was a lot of work and time consumed.
We've found this post because after including a GXT Tree and a Html Text Editor, we saw there was a huge performance issue of the Html Text Editor compared to what we did with ExtJS. We just didn't expect such a thing to happen. Combined with the issues presented by the benchmark on this thread, it was enough to make us think twice about ExtGWT.

We were planning to use ExtGWT as a base tool inside our company but now we must think twice! We were also thinking to buy ExtGWT which by the way has very nice components and it outsells any other GWT component implementation. Too bad the ExtGWT performance is so BAD and now we're thinking to either create our own components or go back to ExtJS.


In the end, I'm really sorry to place such bad remarks on such a promising framework as ExtGWT. It has so many features and it looks great. Combined with GWT it could have been the perfect tool.

I really hope there's a good reply for this by the Sencha Team so we can still feel good about using it.

Are there other users/companies out there using ExtGWT successfully in a big project that needs to be fast and without lags - which matters most for the overall user experience?

Thank you!

[Ekambos]

Hello there,
well i can understand your frustration. having working with GXT for quiet a long time now i must admit that some of their design implementations were really poor(own event classes instaed of going the gwt way, the MVC implementation is just awefull).
But the guys are working hard on GXT3 so i ll recommend to wait until there.

Regards,

Alain

[GrecoRS4]

We're currently at over 120K+ lines of just Java code using GXT, Gin, Guice, EJB3 and our own custom MVC implementation.

Performance seems to be extremely well. Our main app is a whopping 1.04Mb downloaded (using firebug to calculate) including a 260K splash image and we have about 690K in 16 deferredjs files.

I'm not worried about performance being an issue not with the right framework choices and quality of work put into your application.

[dslaycock]

I realize that the the GXT team is small and GXT3 is a big project, but some signs of life from the team would be appreciated. We are looking to greatly expand our usage of the product in our company, but I sometimes question Sencha's resource commitment to GXT.

[gishmo]

You have to do this decision by your own.

If you want to use nice widgets, you have not much alternatives. Writing (and maintaining) your own widgets is expensiv and not your business, GWT has only a poor set of widgets, SmartGWT has one developer and is a JavaScript wrapper and GWT Mosaic, mmmh .... no way.

ExtGWT has beautiful widgets, easy to use and the price is ok.

I made a proof of concept, created my own architecture. I use the layouts and widgets of GXT, mvp4g insteed od GXT MVC and my own loader API. This all works really nice and without any performance issues. If I need widgets in my next projects, I will use GXT.

[mxhn]

I'll reply here to the initial post so it doesn't mislead other people here into thinking that GXT doesn't have potential.

It is as some users pointed out in the other threads. For any gain, there has to be something that you loose. And this is some speed compared to vanilla components that don't compare even nearly with what GXT has to offer. And that is lots of out-of-the shelf powerful components that have much more features than vanilla ones. Having to write all these yourself I'm not sure you'd get much speed improvement but even if you did, the time invested in creating them will bring a heavy weight to the project development time.

I admit there can be many improvements that can be done over GXT components but as @dslaycock pointed out, GXT is a big project. And as @gishmo said, it is indeed the only framework out there that can do something with potential.

Also, the encountered issues had great support during the last period here on GXT forum from users like Colin Alworth and Stigrv. Thank you guys!

Also, it's good to see other people here that are using GXT in big projects. Hope more people could post their usage scenario. It would help a lot.

In the end, GXT remains the king of the hill - it might not be perfect, but it works.

[takayser]

But the guys are working hard on GXT3 so i ll recommend to wait until there.

Who told you so?

[Ekambos]

just a guess dough.
consedering the new features they plan to add it must quiet a lot of work.