Re: Non-deterministic or wrong partition?

I agee with Henry:  we should proceed carefully before making changes in 
this area.  The current state of the Rec is clear, and these are more than 
clarifications.  I certainly have been aware of the fact that the REC 
works this way, and indeed have wrestled with writing implementations that 
deal with the allowed "ambiguities".

I wouldn't necessarily be against making the changes you suggest, Ashok. 
It would in certain respects make implementations simpler.  I do think it 
is a significant and clear incompatibility with the current Rec, and I 
think we need to look carefully for useful content models that might be 
ruled out  I do not think this fits as a clarification or erratum.  Thank 

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

Received on Monday, 10 June 2002 17:05:18 UTC