>Francois-Xavier Prunayre wrote:
>>>>Simon Pigot wrote:
>>>> I have a strong feeling that a shift to XFORMs for XML editing is becoming
>>>> more and more attractive.^M
>>
>> Does XFORM can facilitate creation of inheritance beetween metadata ?
>> >> > Not really, xForm could help on the client side to have a more interactive editor:
> - using xForms, the client side deals with an XML tree which is created
> by the browser as the user add elements and not with key value pair (KVP) as we have
> actually.^M
> - xforms engine is usually based on ajax
> - xforms could support validation against schema types
> - xforms allows defining rules between elements (eg. element a is
> mandatory if b is greater than 10, ...) : close to what schematron does
> i think
>
> Examples :
> http://www.orbeon.com/ops/xforms-todo/
>
> But I've never tried to deals with ISO xml tree on the client side ...
> maybe too big ?
>
>
Yep - XForms has definitely got potential and (as Francois noted) it could be a shortcut to employing Ajax in the editor too....
I think the problem of parsing a schema to drive an (interactive) editor is hard/tedious - I expect that the XForms stuff won't be able to handle all the complexities of the ISO schema tree without changes and since we need the current schema parser and html forms editor to work in the meantime, I'm spending some time on making sure we can parse the schema constructs now! 
Fabien's problem with the uom attribute on gco:distance in gmd:distance was arising from the parsing code not following inheritance links in simpleContent in complexTypes and thus it wasn't picking up the uom attribute - I've fixed it in my test code now - so thanks Fabien!
I forgot to mention that for optional attributes you will be able to force geonetwork to show the attribute (if you take care with the xsl templates) using the schema suggestions file (one for each schema in its directory) - Pierre Lagarde came up with the fix for this at the GN workshop.
Cheers,
Simon