Hi Ty,
for the server side of your application I think that your choice could
be good (I don't know vraptor, and I know Struts2 only a little),
because and modern framework with enough support for publishing data
in REST style could help to simplify your interaction with Pivot.
Probably also Spring MVC (+ Spring Remoting) could be a good choice.
Last year I have done a prototype webapp with Grails 1.x (and related
plugins for security, remoting, etc) and I liked it very much, very
fast in the development and very simple (but it was consuming so much
memory). Now with the 1.3.x release (and Groovy 1.7.x under) things
are also better, give it a try if you have some time ...

Then, also some Web Services framework like Apache CXF for publishing
your server side data in JSON (directly handled by Pivot) could help.

Oh, a trick, a great (and minimal) framework I used time ago for
publishing data was this: cajo ( https://cajo.dev.java.net/ ), giving
transparent usage of remote Java objects via RMI, but hiding it
completely, without the need to wrap your data in some format.


In any case, the primary usage of Pivot is for Client-side
applications (Greg has show to you a new Pivot tutorial on server-side
handling of data).

So I can suggest only few things (based on what I usually do, or try
to do :-) ):
- make it possible to call the same services (called by default by
standard webapp pages) also from Pivot Clients, to reuse the same code
where possible
- implement your server-side features in a way that it will be simple
for you to publish the data in JSON (or maybe XML) so you can consume
it simply with Pivot. But on the protection of those services I can't
say you so much, probably you can use some session authentication
trick or a cookie or other ...
- in initial part of your project, use Pivot not to replicate all your
webapp front end, but only some parts of it (where having a Rich
Client with Pivot really makes a plus for your application):
  for example those parts requiring more interactivity with the user
of maybe access to local resources (local storage, etc), or maybe
implement with Pivot some data heavy features, like administrative
features, dashboards, etc ...


On the JAAS Login, do you want to use it on the server side, or do you
want to use from the Client (in this case I don't know it can work, I
have never tried) ? We have done some experiments with Digest
authentication some months ago and this works (but the implementation
class currently is out of trunk, it wasn't tested so much, and I'll
re-insert it for the 2.0).
On the JAAS usage from client side probably we should take a look ...
and if someone had knowledge on it, is welcome to help us.


I hope to give you some help, if not tell me.

Bye,
Sandro

Reply via email to