W3C home > Mailing lists > Public > public-cwm-bugs@w3.org > May 2004

qnames with - in cwm

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Tue, 4 May 2004 10:39:34 +0100
To: public-cwm-bugs@w3.org
Message-Id: <20040504103934.1f962a3b@hoth.ilrt.bris.ac.uk>

I was not sure where to send this, since I consider it more a Notation3
bug than a CWM bug. http://www.w3.org/DesignIssues/Notation3 doesn't say
to send to any particular list.  Maybe not here but public-cwm-talk?

I noticed you've switched the definition of Notation3's specification in
April, more about this in other messages.

http://www.w3.org/DesignIssues/Notation3 says:
Identifier munging

This syntax does not allow minus signs in identifiers, whereas the
XML encoding for RDF does.

The reference implementation of N3, cwm has been more generous
(in a checked out dev.w3.org/2000/10/swap):

$ grep -i A-Z *|grep -i qname
rdfn3.g:    token QNAME:    r'([a-zA-Z][a-zA-Z0-9_-]*)?:[a-zA-Z0-9_-]+'
rdfn3_yapps.py:            ('QNAME', '([a-zA-Z][a-zA-Z0-9_-]*)?:[a-zA-Z0-9_-]+'),
relaxNG.g:    token QName : r'[a-zA-Z0-9_-]+:[a-zA-Z0-9_-]+' # @@???

and in notation3.py:
  _notQNameChars = "\t\r\n !\"#$%&'()*.,+/;<=>?@[\\]^`{|}~"  # Assume anything else valid qname :-/
(as used in 'def qname' later)

All of these would allow foo:bar-baz

Trivial test:

$ cat > t.n3
@prefix foo: <http://example.org/> .
foo:bar foo:baz foo:bar-baz .
$ cwm t.n3
#Processed by Id: cwm.py,v 1.148 2004/03/21 04:24:32 timbl Exp
        #    using base file:/home/cmdjb/w3c/2000/10/swap/t.n3
#  Notation3 generation by
#       notation3.py,v 1.153 2004/03/21 04:24:35 timbl Exp
#   Base was: file:/home/cmdjb/w3c/2000/10/swap/t.n3
    <http://example.org/bar>     <http://example.org/baz> <http://example.org/bar-baz> .

The newer http://www.w3.org/2000/10/swap/grammar/n3.n3
CVS 1.9 2003/10/30 22:41:34 says
  qname bnf:matches "(([a-zA-Z_][a-zA-Z0-9_]*)?:)?([a-zA-Z_][a-zA-Z0-9_]*)?";

I think it would be a good idea to allow - inside qnames.  You can other
uses by whitespace and let people avoid CamelCase inside names if they
don't like that.

I think some people have already been using - in names, relying on cwm
accepting it.  (I had a quick look and, for example, Jena's N3 parser
accepts it).  I'm tempted to add it to Turtle to match what people are
using and want to use.

Received on Tuesday, 4 May 2004 05:40:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:51:59 UTC