Hi Laird, see below.

> I appreciate the efficiency of this from a DRY standpoint--the GWT
> internals are downloaded once--but I don't appreciate the
> *in*efficiency of this from a load-the-whole-world perspective.  That
> is, I don't particularly want 27 sub-apps living in my browser unless
> they are absolutely needed; I'd rather only reference an HTML page
> that does a <script src=""> import when that portion of the overall
> app is needed.

Above you said 37, here you say 27 sub-apps, I assume neither number
is real. But I think a more important estimate is how many different
configurations of these sub-apps do you envision? You mentioned two
options, but I think there is a third option, and I think it might
work better for you. But it depends on how many different
configurations of your sub-apps are to be expected.

Option #3:
You have many (ca. 30?) sub-apps. These will be combined in various
ways depending on user needs. How many such configurations are there?
Let's assume there are five. I suggest you take your "option 2"
approach but apply it to each of these five configurations, rather
than the single monolithic approach.

a. So if you had 30 sub-apps, you'd have 30 sub-app modules (each
inherits from "Application").
b. You'd still have the "housing" module, to maintain a common feel
(it houses multiple "Application" instances).
c. If you had 5 required configurations, you'd have 5 more modules.
Each of these modules instantiates a Housing instance, instantiates
the necessary Application instances, and adds these to the Housing.
You'd probably define a "Configuration" class.

I think this will give you the benefit of delivering to your users
only just what they need, without duplication and without extraneous
sub-apps.


> (Lastly, it looks like option #2 could become a valid choice for me
> again oncehttp://code.google.com/p/google-web-toolkit/wiki/CodeSplitting
> goes in to an official release.  That is, it looks like code splitting
> will allow me to compile a giant GWT module--consisting of the finance
> app, the order app, the housing app, the kitchen sink app--but
> deferring much of the loading of hunks of the Javascript until they're
> actually needed.  Should I begin designing assuming that code
> splitting will appear soon, or is that a mechanism that is not
> designed to be used in such ordinary business-application ways?)
No idea on the timing. My suggestion would be to go with my "option 3"
for now and, as dynamic code slurping becomes available I think you
should have a pretty smooth transition. This announcement says it will
be post-1.6 but that's all I've seen:
http://googlewebtoolkit.blogspot.com/2008/12/whats-ahead-for-google-web-toolkit_10.html

>
> Thanks again for your time.
>
> Best,
> Laird
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Google Web Toolkit" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/Google-Web-Toolkit?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to