Re: uri-comp draft necessary?

Miles Sabin wrote:

> Paul Prescod wrote,
>
> >Everyone agrees we don't want to go there. Let's not go there. Next
> >issue?
>
>
> If we end up in a situation where we're saying that one or other of,
>
>
>
> and,
>
>
>
> is frowned on (which one, I wonder, and why?) then we're going there
> willy nilly.
>
> That's where this alleged "best practice" leads.

I think you're trying to invent a problem where there isn't one. Have 
you ever seen two namespace declarations that differed only in case? 
Have you ever seen software that produces them? Have you ever seen 
software that was confused by them? This "best practice" is common sense 
and writing it down is at worst (and probably at best!) a waste of time.

You're using a lot of rhetorically charged words like "willy nilly" and 
"changing through the back door" and "opening a can of worms". But can 
you please outline a realistic scenario in which there is a problem 
caused by having a string-based equivalence for namespace URIs and yet 
having them be dereferencable (and thus _also_ subject to protocol 
equivalence constraints). I'm waiting to hear of a scenario where a lung 
machine explodes or a purchase order goes awry or ...

  Paul Prescod

Received on Thursday, 19 December 2002 14:03:15 UTC