- From: Sandro Hawke <sandro@w3.org>
- Date: Thu, 29 Jan 2009 14:37:57 -0500
- To: public-rif-wg@w3.org
I just noticed that the CURIE syntax is much broader than I expected. All the IRI "sub-delims" are valid in CURIEs after the colon, including comma and both parentheses. This means most PS as written is not syntactically valid (using normal parsing techniques). For instance, in the BLD spec, in the first example, cpt:buy(?Buyer ?Item ?Seller) :- cpt:sell(?Seller ?Item ?Buyer) the first token a typical lexer would find is a CURIE, "cpt:buy(". Note the open-paren is part of the CURIE. Then of course we'd have a syntax error, because there's no open-parenthesis token. It might be possible to code around this with some clever back-tracking when we find the open-paren is needed, but I don't think that's a good idea. I suggest we switch (back) to qnames [1]. The only advantage to CURIEs I know of is that more IRIs can be abreviated, but I don't think that's worth a whole lot in our use cases. -- Sandro [1] As in http://www.w3.org/TR/REC-xml-names/ and I'm not sure whether we want QName or PrefixedName, really. If we allow unprefixedName, we need to think a little more about reserved terms. I could also see forbidding the "-", so infix math doesn't require spaces.
Received on Thursday, 29 January 2009 19:38:09 UTC