- From: Mukund Balasubramanian <mukund@cs.stanford.edu>
 - Date: Mon, 25 Sep 2000 13:49:59 -0700
 - To: Simon Brooke <simon@jasmine.org.uk>
 - CC: www-forms@w3.org, "xmlschema-dev@w3.org" <xmlschema-dev@w3.org>
 
I agree,
    It is practically impossible to do the same with the current spec. On the
other hand, I am also exploring the idea of generating the same from instances of
XML schema.
    The way I see it, it is more likely that XML schema instances will be
representable by forms rather than have a seperate XForms data model.
    XML Schema -> Types -> Instances (with data) -> XML Reresentation (easy)
                                                        |
                                                        -----------------> Form
representation (my own) --> XSL/T --> Legacy HTML forms
                                                          (programmatic)
    This works very well for me since my basic programming model is based on XML
schema types. If anybody is interested, I would like some feedback on the same
too.
Mukund Balu
Simon Brooke wrote:
> I've just been studying the XForm data model with a view to writing a general
> XSL-T transform to convert an XForm for delivery on legacy HTML devices, and
> I come to the conclusion that the separation of model and instance data makes
> this impossible. I cannot, in a single transformation, create a pre-populated
> HTML form from an XForm document.
>
> I'm not arguing that separation of model and instance is a bad thing -
> clearly it's not - but if I'm right this does mean that writing code which
> will address both XForm-capable and legacy devices is extremely difficult and
> requires great duplication of code. Legacy devices are inevitably going to be
> with us for a considerable time and so must be supported, There is, in
> effect, no migration path.
>
> Am I right? If so, has some thought been given to this?
>
> --
> simon@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
>
>                 ;; Semper in faecibus sumus, sole profundum variat.
Received on Monday, 25 September 2000 04:17:15 UTC