Re: Request for Feedback: Content Models in XForms

>...
>OPTION 1:
><xsd:sequence>
>    <xsd:element ref="xform:submitInfo" minOccurs="0"/>
>    <xsd:element ref="xform:model" minOccurs="0"/>
>    <xsd:element ref="xform:instance" minOccurs="0"/>
>    <xsd:element ref="xform:bind" minOccurs="0" maxOccurs="unbounded"/>
></xsd:sequence>

>This means authors MUST write the elements in exactly the specified
>sequence.

>Changing the order of the sequence in any way results in an invalid
>form.

Thatīs is unacceptable. Even if we count on automatic tools (Page Editors, 
XForms Back-End Preprocessors, ...) to help us, it makes "open", meaning 
"extensible", forms much harder to develop.

Just to say how important it is: I have hundreds of extensible forms in my 
web application today, 'extended' diferently for each of my thousand 
customers, whose have today just one choice of browser (IE5.x) to use 
because I couldnīt make my javascript/XML solution work on anything else. 
XForms, if kept easily extensible and then broadly deployed, would 
marvelously improve the situation.

>Unfortunately XML Schema, with which we are mainly concerned, doesn't
>support this. Schema only supports <sequence>, as above, or <choice>.

I think itīs time for version 1.1 of Schemas. They probably dropped the 
initial support for unordered sequences of elements, to easy the job on 
"verifiers", so that they can be smaller and faster (obviously a requirement 
on mobile platforms). Itīs harder to verify out-of-sequence itens with 
multiplicity restrictions, because you have to keep separate
counters, but I think itīs better to have a <xsd:unorderedSequence> than 
having to code all the permutations and have the verifier check all of then, 
what is possibly worst-performing.

>Another option in Schema, <all>, works only with single elements. If we 
>restrict ourselves to maxOccurs=1 on every element, then we
>could use the <xsd:all> connector...

If changing Schemas direction/evolution is out of question, or not a fast 
enough route, using <xsd:all> looks not terribly bad for the <xform> 
definition, but for form controls:... They need "freedom" to mix easily with 
xhtml, xmath, ..., and even with non-standard things like "Flash".

Just a question, xhtml is "schematable"? More properly, is there a XML 
Schema to validate properly a xhtml document/page? If it exists and copes 
with the highly nestable and unordered sequences of div, p, span ... we can 
follow it as a guideline.

Rafael Teixeira
Brazilian Developer

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

Received on Monday, 23 July 2001 13:47:06 UTC