struts-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Martin Cooper" <>
Subject Re: [action2] Upgrade Dojo toolkit
Date Tue, 20 Jun 2006 16:21:45 GMT
On 6/20/06, Jason Carreira <> wrote:
> >
> > 2. The current Dojo based widgets are simple enough
> > to be used in some
> > sample/tutorial like pages but for real life apps
> > they don't work (or
> > put it in another more realistic way: I haven't been
> > able to have them
> > working within a normal timeframe).
> > I have to mention, that  I haven't used the
> > Dojo-based tree, but I
> > tried all the others.
> >
> Well, I don't use the ww:a or ww:div tags that much, but we do a lot of
> custom tags based on Dojo, the tree tag being the one I spent the time to
> port back to WW (although I didn't doc it, sorry). From our perspective, WW
> tags + Freemarker templates + Dojo widgets works great. It's not really
> rocket science.
> I talked with Patrick, and the understanding we came to was that the Dojo
> stuff would still be there in a different theme, and there would be a
> default AJAX theme that would be simpler to use (I guess).
> My main concern is what you lose:
> 1) The publish / subscribe topics -> this is the biggest one for me. It
> only really makes sense to make a lot of simple little AJAX widgets if you
> can loosely couple them and have them interact based on runtime wiring,
> rather than having to have them all pre-configured together. Otherwise
> you're back to custom Javascript coding everything, which you get to if you
> get very complicated anyway...

I'd add to this the dojo.event.connect* family, which provides another
hands-off mechanism for conencting up the pieces of your app, as well as
adding awesome AOP-like functionality.

2) The widget framework -> As Martin pointed out, the widget framework is
> really excellent and gives you a nice structure to build widgets in. No,
> it's not super-duper-easy, but we're still in the cowboy phase of AJAX
> development...
> Let's face it, you're going to have to get your hands dirty and learn a
> little bit if you want to do anything real. Patrick and Ian say Prototype,
> etc. are easier to get your hands dirty with. Ok, I'll accept that. Writing
> my first Dojo widget wasn't that quick or painless. Makes sense to give
> users something they can get going with quicker.

My fear with Prototype is that it may be easy to get going with, but it's
also way easier to get yourself in deep trouble with if you mix toolkits and
/ or you're building portlets, since it's just not a polite browser citizen.

Personally, I'll stick with Dojo. For me, it's the same as the
> script-language vs. Java debate. Script languages are undoubtedly quicker to
> get something going with, but if you're engineering an enterprise app you
> use Java. I see the same level of engineering commitment and corporate
> backing behind Dojo as I see behind Java.

Yup, and the corporate backing continues to grow.

Martin Cooper

Who knows? With IBM and Sun behind Dojo maybe we'll even see some usability
> experts and technical writers assigned and Dojo will end up being easy to
> use. We're not shutting the door on Dojo, and maybe it will be the default
> again in the future.
> ---------------------------------------------------------------------
> Posted via Jive Forums
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message