xml-dev - Re: The Tool X horror scenario (was RE: Begging the Question (the novel)

["Clark C. Evans" <cce@clarkevans.com>]

On Sat, 30 Dec 2000 Mike.Champion@SoftwareAG-USA.com wrote:
> So, the problem is NOT with the people who thoroughly understand namespaces
> ... but with the people who think that namespaces look like URLs, and Tool X
> treats them like URLs pointing to some sort of object, and then raise hell
> with all the vendors whose tools do NOT treat namespaces like Tool X does.

Is this (above) the essence of the problem?  Since the spec is
silent on this issue, it allows those vendors with very large 
marketshare to create a de-facto interpretation of the specification
which has in fact more weight than the spec itself?

> Just as many of us, on many mailing lists, have patiently explained over and
> over and over that the XSLT language that IE5 supports is not the one that
> the W3C Recommendation defines, so one should not expect it to interoperate
> with other tools ... so I can imagine spending the next few years patiently
> explaining that the behavior of Tool X is *consistent* with the namespace
> Recommendation but not what is *required* by the namespace Recommendation,
> so that behavior should not be expected by other tools.  
> 
> I'm not sure that I agree with the argument that we can best avoid this
> scenario by deprecating URLs as namespace IDs (this *would* play havoc with
> RDF).  I can see  value in Paul Tchistopolskii's suggestion that namespace
> IDs that are NOT dereferenceable be specified as something other than URLs.
> That could entail a revision of the Namespace Recommendation ... or perhaps
> it should be promulgated as a "best practice" without formal sanction.  In
> any event, to echo a point that has been made repeatedly in this thread, we
> should take the massive confusion on this subject AMONG XML "EXPERTS" as a
> strong indication that something needs to be do before the confusion
> spreads too far among XML "customers".

Thank you Mike.

Best,

Clark Evans