You know, I've played with GWT a bit and have been thinking about it a
lot, and for me, I've come to these conclusions...
* As a technology demonstration, it's cool as hell. If your a coder,
you've gotta love it.
* You have to begin from the mindset that you actually like the Java GUI
way of writing UI's. I don't, never have... it has always struck me as
too much work for what you get. Describing a GUI layout in code is a
bad idea in my opinion, and I'll get back to this in a moment.
* For some applications, a very limited scope in my opinion, it will be
useful. For anything else, it'll be more trouble than its worth.
Here's my overall take on it... many people work in shops where a core
group of developers do it all. They code the back-end business logic,
they code the front-end UI, and everything in between. In shops like
that, GWT might offer something interesting (although I still don't
think it's a great idea there, as I'll explain shortly).
In shops where you have a clear separation between page developers,
back-end developers, DBA's, and so on, it won't work in my estimation.
Here's the core problem in my opinion: should layout be described in
code by programmers, or by graphic designers and layout experts in
simple markup? I'm in the later camp (even though, ironically, I've
never had to personally deal with that situation... conceptually though,
that's where I am). Using GWT, every change to the UI has to go through
the programming staff, not the graphic design staff. That's not the way
it should work.
Also consider that every layer of generated code you add is another
layer of your possible mission-critical application you don't truly
understand and can't properly maintain. I understand that to some
degree we always have to trust code generation... javac is a code
generator in a sense, albeit not quite in the same vein. We all trust
that just fine, so why not GWT? You could make that argument, but I
would argue that it's one more layer you have to get through to solve
problems. It may be worth it, it may not.
On the question of how it might fit into SAF2, my own opinion is it
doesn't, plain and simple. I view it as something that would take the
place of SAF2, not something you would add to SAF2. The one exception
is if the code it generates could be made to talk to SAF2 instead of
what it has running on the server... that's a possibility. I suppose
you could have some sort of proxy layer written in GWT code that calls
SAF2, and I think someone around here may even have suggested that
recently, but I'd have to see that to judge it's merit.
So, from a purely technical "cool" point of view, I very much like GWT.
From a "going to use this for the very next project I can" point of
view, I'm far less convinced. As a developer I think it's neat, but I
think actually using it would basically be a reversion to approaches to
development that are not optimal in the Internet world.
Frank
netsql wrote:
http://www-128.ibm.com/developerworks/library/j-ajax4/index.html?ca=dgr-lnxw07GWT4Ajax
What do people think relative to Struts 2's built in Ajax?
.V
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
.
--
Frank W. Zammetti
Founder and Chief Software Architect
Omnytex Technologies
http://www.omnytex.com
AIM: fzammetti
Yahoo: fzammetti
MSN: [EMAIL PROTECTED]
Java Web Parts -
http://javawebparts.sourceforge.net
Supplying the wheel, so you don't have to reinvent it!
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
