[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: more QName madness

• To: "'Rich Salz'" <rsalz@d...>,"'Simon St.Laurent'" <simonstl@s...>
• Subject: RE: more QName madness
• From: "Michael Kay" <michael.h.kay@n...>
• Date: Fri, 15 Nov 2002 13:43:18 -0000
• Cc: <xml-dev@l...>
• Importance: Normal
• In-reply-to: <Pine.LNX.4.44L0.0211141846120.28409-100000@s...>
• Reply-to: <michael.h.kay@n...>

> > >Sadly, my prediction that no-one would ever use XML Namespaces 
> > >because the whole idea was so unworkable is one of many 
> occasions on 
> > >which I
> 
> I don't think I understand what the alternative is.  Is it this?
>     Element names don't matter but it can have multiple xmlns:type
>     attributes that have URI(s) describing the element?

It's too late for an alternative, unfortunately.

An alternative might have been:
  - encourage element/attribute names following the Java convention
"com.bigco.package.somename"
  - allow definition of a default prefix, so that the above could be
abbreviated to "somename"
  - require XML parsers to report the expanded name to the application,
not the abbreviated name or the prefix declaration.

It would have saved millions of lines of code and countless hours of W3C
WG meeting time...


Michael Kay
Software AG
home: Michael.H.Kay@n...
work: Michael.Kay@s... 

• References:
  • RE: more QName madness
    • From: Rich Salz <rsalz@d...>

• Prev by Date: RE: XML/RDF
• Next by Date: Re: Malicious XML
• Previous by thread: RE: more QName madness
• Next by thread: Article: Proper XML Output in Python (with broader appeal)
• Index(es):
  • Date
  • Thread