- From: Dave Beckett <dave.beckett@bristol.ac.uk>
- Date: Mon, 15 Oct 2001 15:32:37 +0100
- To: Louis Theran <theran@cs.umass.edu>
- Cc: www-rdf-comments@w3.org, barstow@w3.org
I have been editing a new version of the RDF test cases doc and updated the N-Triples section. With respect to the comments you raise in http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0288.html You raised these points: * Why not have a canonical N-Triples form and allow comparison using cmp As Art Barstow notes in http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0293.html this is actually a graph comparison problem between the two (or more) graphs created when the N-Triples files are processed into RDF graph models. Since each of these RDF graphs can be arbitrary, and there is no good way to provide a canonical ordering of arbitrary graphs, and even if there were, this format is meant to be easy for people to read and write, so we make the machines do a little more work instead. The syntax details such as comments don't really interfere but rather are useful for people to indicate the usual kind of stuff such as copyright, source code version information, reasons for the test, pointers to references etc. We have now had several implementations of this standard graph comparison algorithm for N-Triples and it seems relatively easy and straightforward for the small test cases that we are using. We appreciate there are issues with doing this for larger graphs in which case more sophisticated algorithms may be needed. * Strange escaping for URIs I have replaced this with the recommended %HH hexadecimal escaping These changes have been folded into http://www.w3.org/2001/08/rdf-test/ (the list of the changes is at the end of that doc) I hope this is satisfactory to you. Thanks Dave
Received on Monday, 15 October 2001 10:33:01 UTC