• To: "Bullard, Claude L (Len)" <clbullar@i...>
• Subject: Re: the infoset is two infosets (or even three?) [was: Re: linking, 80/20]
• From: Paul Prescod <paul@p...>
• Date: Thu, 29 Aug 2002 11:19:43 -0700
• Cc: "'Tim Bray'" <tbray@t...>, xml-dev@l...
• References: <2C61CCE8A870D211A523080009B94E430752B8CE@HQ5>

"Bullard, Claude L (Len)" wrote:
> 
>...
> 
> The insistence by some that they are MerelyDisambiguators when
> empirical testing shows they are semantic loading devices is
> confusing.  Worse, it leads to expectations not proven correct
> in production.  Change the namespace and the code fails that
> worked before the change.  Somewhere a bit changed state. That
> is not syntax.

I'm really not understanding this complaint. Namespace names are no more
or less syntactic or semantic than element type names. If you change the
"P" element to "PARA" you expect software to break. Why wouldn't you
expect the same if you change "http://1999/WD-XSL" to
"http://2001/REC-XSL". This is one confusion I've never seen with my
customers. After all, they are taught that namespaces are compared by
string value, "like UUIDs".
-- 
 Paul Prescod

• References:
  • RE: the infoset is two infosets (or even three?) [was: Re: linking, 80/20]
    • From: "Bullard, Claude L (Len)" <clbullar@i...>

• Prev by Date: Re: the infoset is two infosets (or even three?) [was: Re: linking, 80/20]
• Next by Date: Re: the infoset is two infosets (or even three?) [was: Re: linking, 80/20]
• Previous by thread: RE: the infoset is two infosets (or even three?) [was: Re: linking, 80/20]
• Next by thread: RE: the infoset is two infosets (or even three?) [was: Re: linking, 80/20]
• Index(es):
  • Date
  • Thread