RE: xforms processors

Joel,

> IMO only option 3 is really flexible enough for real applications ...

Absolutely!

I was just trying to establish the point that there is only really one model
for a processor (from 50,000 feet), and that the issue is where you place
the modules. I feel current implementations that take the 'option 2'
approach in order to achieve 'no install', are tending to dump crucial parts
of the XForms spec by simply missing out some of these modules. The goal
should surely be forms that will run on any XForms processor, regardless of
whether that is a full client-side implementation, or a full server-side
one.

So, a good XForms processor will sit on a server, and be clever enough to
decide how much of its 'controller' to put on the client on a per-connection
basis. A connection to a WAP phone might be best implemented by giving up
only one form control at a time, and then every single UI event on the phone
results in a trip straight back to the server, and a new page. All the
validation, model item properties, and so on, are server-side.

An HTML 4 implementation, on the other hand, may be given a snapshot of all
instance data that is referred to in bind statements or forms controls, and
so could happily sit interacting with the user until they do anything that
requires different instance data. At this point your return to the server
with your changes to the instance data, the server incorporates them, and
then gives out some new instance data with a new form.
 
Finally, a dialog between the same server and a fully-fledged XForms browser
would 'push' almost all of the controlling onto the client.

So, as you say, option 3 will include a 'no install' option, as well as a
'full install' option - all completely dynamic, and requiring only one
version of the forms.

Regards,

Mark


Mark Birbeck
Co-author Professional XML and
Professional XML Meta Data,
both by Wrox Press

Managing Director
x-port.net Ltd.
4 Pear Tree Court
London
EC1R 0DS

E: Mark.Birbeck@x-port.net
W: www.x-port.net
T: +44 (20) 7689 9232

Received on Tuesday, 1 July 2003 11:29:15 UTC