Re: A simpler solution to %7e vs %7E vs ~ in namespace comparison

Larry Masinter writes:

>> Don't allow hex-escapes in namespace names.

Hmm.  I don't object to the spirit of this, but I think we need to 
distinguish two possible interpretations:

--------------
1.  Implementations of namespaces must check for and enforce this rule.

vs.

2. Hex-escapes aren't allowed.  Don't use them.  Implementations won't 
check, but results may be unpredictable when namespace names are then used 
to access Web resources.
--------------

#1 is cleaner architecturally BUT, I think one of the reasons namespaces 
went with character-by-character is so that implementations could avoid 
the overhead of rummaging around in the internals of the string.  #1 
breaks that, and I think I'm against it on those grounds.  #2 sort of says 
don't go there, but might lead to (probably manageable) interop problems 
if some tools reject namespaces with escapes and others don't check. 
Bottom line:  I can live with #2, I'm not sure I can live with #1. 
Speaking from experience, namespaces are already a significant source of 
implementation overhead in high performance implementations.  I'd rather 
not make it worse.  Thank you!

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------

Received on Tuesday, 11 February 2003 10:10:48 UTC