• To: "'Rick Jelliffe'" <ricko@a...>,<xml-dev@l...>
• Subject: RE: Unicode normalization in XML 1.1
• From: "Michael Kay" <michael.h.kay@n...>
• Date: Fri, 4 Apr 2003 10:39:24 +0100
• Importance: Normal
• In-reply-to: <008701c2f9ff$2a4be9b0$4bc8a8c0@A...>
• Reply-to: <michael.h.kay@n...>

> To implement it, you might put a switch in SAXON for
> "server mode" which switches off any normalization.

Actually, the XSLT 2.0 spec provides a switch, and its default is that
it doesn't do normalization.

But if this is what the authors of charmod meant, they should have said
so. If an XSLT processor is a text processing component, then it doesn't
conform to charmod as written. You are saying that it's the
responsibility of the user of the text processing component to decide
whether to do normalization. That is eminently reasonable, but it's not
what charmod says.

Incidentally, the whole tone of the Character Model spec reminds me very
strongly of the rules and regulations produced by QA departments who
don't have any responsibility for shipping real software to real users.
It's full of admirable good intentions and steers well clear of the real
practical problems, such as handling of legacy data. Power without
responsibility.

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

• Follow-Ups:
  • Re: Unicode normalization in XML 1.1
    • From: "Rick Jelliffe" <ricko@a...>
  • Summary of draft W3C Character Model for programmers
    • From: "Rick Jelliffe" <ricko@a...>

• References:
  • Re: Unicode normalization in XML 1.1
    • From: "Rick Jelliffe" <ricko@a...>

• Prev by Date: help wanted
• Next by Date: RE: BASE64 (was Re: CDATA)
• Previous by thread: Re: Unicode normalization in XML 1.1
• Next by thread: Summary of draft W3C Character Model for programmers
• Index(es):
  • Date
  • Thread