Lists Home |
Date Index |
Bob Foster wrote:
> I sent a formal comment on that "feature" and was advised it was an
> error that would be corrected; the intention was to _allow_ the
> schemaLocation attribute.
That's good to hear.
> That reduces the shock value, but it's still pretty silly. Why don't
> they just allow attributes from any foreign namespace on every element
> like schema languages do and let processors ignore the attributes they
> don't recognize?
Yeah it's quite silly indeed. I guess that needs to be another formal
> (I didn't fare so well with my accompanying suggestion that it was
> absurd to use XML Schema for XHTML validation, as there are so many
> constraints it can't model. The response to that was more like "Harumph!
> XML Schema is an official W3 recommendation!". ;)
I wonder how long that's likely to last as the HTML and Forms WGs (ie,
basically the same people) are pretty much the only Interaction WGs to
still use WXS. Last week Mark Birbeck came to the CDF WG meeting to give
a presentation (which was excellent) on XML Schema best practices so
that WGs could produce WXSs that would be easy to combine. While I can't
speak for the WG, and while Mark's ideas are very interesting, I got the
impression that the presentation finished up convincing the pro-WXS on
the WG that one just had to jump through too many hoops to make it work,
especially when you have NVDL handy that's just begging to be used.
So discussion is still ongoing, but I think there's hope.
Senior Research Scientist