[ALL] review version of XSCH

While I have not allowed Jeff time to make his updates (apologies Jeff), 
I think todaty is the deadline for the review versions for the F2F.

At the F2F I hope we can decide either to go with the xpath equality 
solution or the all primitives types differ solution.

Thus I have prepared two versions of the document, one for each:

http://www.w3.org/2001/sw/BestPractices/XSCH/xsch-sw-20051027/xpath

http://www.w3.org/2001/sw/BestPractices/XSCH/xsch-sw-20051027/derive

if you want to see both lots of text together, try:

http://www.w3.org/2001/sw/BestPractices/XSCH/xsch-sw-20051027/both

if you want to see the old text as well, (struck out), see

http://www.w3.org/2001/sw/BestPractices/XSCH/xsch-sw-20051027/


Jeff has indicated that he wishes to check some of the references.

A brief summary of the choice between the two versions.

1) The XPath version is significantly harder to implement. Dave Reynolds 
has reported that it will cause difficulty to the HP Jena team (note: 
hence I represent this position)

2) The XPath version is mathematically less well defined, since the 
underlying operator is not an equivalence operator; while this is 
fixable in practice, it is not without significant cost, both in terms 
of software complexity, runtime, and potentially surprising behaviour in 
corner cases.
(The solutions discussed depend on the vocabulary being considered. 
Adding items to the vocabulary without adding any additional triples may 
result in additional entailments which make no explicit use of the new 
vocabulary)

3) The entailments resulting with the XPath solution may be more 
intuitive and easier for the end users.


Jeremy

Received on Thursday, 27 October 2005 15:53:43 UTC