Just some additional thoughts on this front... It seems like most
frameworks are heading in very similar directions now. They are all
getting lighter, faster, and better at selecting and manipulating
elements. I'm in favor or having some AJAX functionality baked, but I
have never found it very difficult to write custom handling when
necessary. I use JQuery for everything and writing a date picker,
validation, AJAX form submits, etc was fairly easy.
-bp
Adam Peller wrote:
Musachy Barroso wrote:
I have totally changed my position about the dojo plugin. I think Struts
should have some ajax functionality for the most commom use cases, but I
think we just picked the wrong ajax framework.
Musachy, have you looked at Dojo lately? I can understand your frustration,
especially given that you're on what's now a pretty old release of Dojo.
Please take a good hard look at Dojo 1.0 (or 1.1beta) before you make any
rash decisions. I'd very much like to see your codebase upgraded to get
your users the best experience. I hope you try out Dojo 1.0 and take your
questions / issues to us at the dojotoolkit.org forums.
...What is really frustrating is that after spending time getting
everything to work on one version, all hell breaks loose on the next
"minor" release.
The 0.4->0.9 rewrite was not a typical minor release. The shift in version
numbers was intended to indicate that. And that was still relatively early
in the development of Dojo -- note the smaller decimal. Dojo has matured
significantly since then, based on this rewrite and one time, ok, massive
API shift. We know this has a lot of impact on people like you and we don't
take that lightly or have any plans to do it again.
The dojo plugin is a lot better in 2.1 that it was on 2.0.x, but by the
time that 2.1 goes GA dojo will be on version 97 or so :)
Since then, Dojo has been quite stable and has been sticking to a
deprecation cycle of one major release, which we expect to be at least a
year, and that's just for minor API changes. Our roadmap has no Dojo 97.0.
Dojo 1.0 is all about performance and stability, and the follow-on Dojo 1.x
releases continue in that direction with no radical changes to the core or
Dijit architecture. Dojo base is now very tiny (<25K, on par with other
toolkits) and the performance of the new parser in the core is order of
magnitudes better than the old one. I think you'll find the APIs are easier
to use and programmatic instantiation more intuitive. Widgets are more
easily customized and themable via CSS, there's a convenient xpath query
mechanism that is actually among the fastest of the toolkits (see
http://mootools.net/slickspeed/)
On top of that, I'll remind you that Dojo is totally open source with
extremely friendly licenses that should be fully compatible with yours.
//have you seen how many emails we get with question about dojo vs struts
itself?
We should find a way of directing those users to dojotoolkit.org for
support. Getting them off the very old release should solve the majority of
their issues.
Regards,
Adam Peller
Dojo committer
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
