- From: Tim Berners-Lee <timbl@w3.org>
- Date: Sun, 29 Feb 2004 16:26:06 -0500
- To: Dave Beckett <dave.beckett@bristol.ac.uk>
- Cc: public-cwm-bugs@w3.org
> cwm (and jena's n3 parser) seem to implement something different than > specified in the Notation3 grammar. > > Specifically, the grammar says this: > [[ > objectlist > object > object , objectlist > ]] -- http://www.w3.org/DesignIssues/Notation3.html > > and in the RDF/n3 one: > [[ > objecttail bnf:mustBeOneSequence ( > ( ) > ( "," object objecttail ) > ). > ]] -- http://www.w3.org/2000/10/swap/grammar/n3.n3 > > > So a comma is required between entries. Yes, this is an objectlist as in a sentence with several triples of the same subject and predicate, like :Joe :child :Egbert, :Marmaduke. objecttail is used in propertlist, which is used in [ propertylist ] as an anonymous node, and in statement: statement bnf:mustBeOneSequence(( subject propertylist )). > cwm implements something else - commas are forbidden. > > Failing test with cwm: > @prefix : <http://example.org/stuff/1.0/> . > :a :b ( "apple", "banana" ) . That is a list in the sense of a collection. In the grammar, ( pathlist ). Pathlist has no commas. > Passing test with cwm: > @prefix : <http://example.org/stuff/1.0/> . > :a :b ( "apple" "banana" ) . > > I need to change turtle one way or the other to match. I assume > the grammar is in error. I think you just got the productions crossed. If I have missed something, let me know. Thanks for implementing it. Maybe we should make various levels of common test suite. Tim > Dave
Received on Sunday, 29 February 2004 16:26:07 UTC