W3C home > Mailing lists > Public > www-forms@w3.org > September 2002

RE: A silly idea?

From: Danny Ayers <danny666@virgilio.it>
Date: Sun, 8 Sep 2002 01:25:48 +0200
To: "Robert Trybis" <rt@cygnets.co.uk>, <www-forms@w3.org>
Message-ID: <EBEPLGMHCDOJJJPCFHEFIEGJHLAA.danny666@virgilio.it>

Hi Robert,

I think the other responses have covered the silliness angle (i.e. it isn't)
but I reckon there are potential practical problems with your 'upgrade
stylesheet' suggestion. There may be some issues like DOCTYPE (which turned
out not to be a problem) lurking around the infoset, but what I have in mind
here is something where a mapping just isn't possible.

To be able to transform xx version 1 into xx version 2 using a stylesheet it
must be possible to transform v1 into v2 - which isn't nearly as silly as it
sounds, so (knowing some of Andrew's recent history), here's an example from
SVG (Structured Vegetable Gardening)

version 1.0 has the element <veg>
version 2.0 replaces this with the elements <largeVeg> and <smallVeg>

in v1.0 my markup looks like this

in v2.0 it *should* look like this

how does the stylesheet know to use <smallVeg> rather than <largeVeg> in
this case?

given <veg>aubergine</veg>
where does the brinjal go?

When using an XML-based description, there is also the fact that it is
fundamentally tree-based. Quite a few specs effectively break out of the
limitation by layering on top (e.g. href links), and anything that slips
outside of the tree model could be potentially untransformable using XSLT.

So although the idea might well be very useful for syntax-only changes, and
a subset of structural changes, it has its limitations for the market


Danny Ayers
<stuff> http://www.isacat.net </stuff>

Idea maps for the Semantic Web

>-----Original Message-----
>From: www-forms-request@w3.org [mailto:www-forms-request@w3.org]On
>Behalf Of Robert Trybis
>Sent: 05 September 2002 10:34
>To: www-forms@w3.org
>Subject: A silly idea?
>I said that I would like to "future proof" my forms by starting to write
>them in xforms.
>Responses pointed out that xforms is only at working draft 1.0 so any
>xforms generated now might require to be revised several times.
>This triggered the, probably silly, idea that all XML specifications
>(XFROMS, XHTML etc. etc.) should include XSLT stylesheets that would
>take a document conforming to that revision of the specification and
>spit it straight back out.
>This sounds a bit pointless. However such a specification would mean
>people instantly had working stylesheets, conforming to the
>specification, which they could begin to modify. It would also mean
>that, with a bit of cutting and pasting between two specifications,
>stylesheets could be generated that would update xforms compliant with
>earlier specifications to the latest revision. It might even help with
>the generation of stylesheets that translate between say Xforms and
>XHTML, if only for UI markup purposes. Having working stylesheets would
>be a big boost to people like myself who are trying to quickly get to
>grips with several new technologies and it would help standardisation
>and long term maintenance.
>Is this a silly idea, if so why? - fire away.
Received on Saturday, 7 September 2002 19:35:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:36:07 UTC