- From: Steve Harris <steve.harris@garlik.com>
- Date: Sun, 6 Mar 2011 18:57:33 +0000
- To: Richard Cyganiak <richard@cyganiak.de>
- Cc: nathan@webr3.org, RDF Working Group WG <public-rdf-wg@w3.org>
On 2011-03-06, at 15:25, Richard Cyganiak wrote: > On 6 Mar 2011, at 14:40, Nathan wrote: >>> My concern isn't all that much about graph literals, although I don't see the evidence for them being needed. >> >> Cool, okay - do you know of any problems that may be caused by including them? > > Like any feature, it adds cost for implementors. Also comprehension by users, there are a bunch of (mis)features that are in RDF which people are tempted to use until they find why they're broken. e.g. containers, collections, reification, XMLLiterals, and bNodes - though they're mostly only partly broken, none of them can be used in all the situations that users often expect. They might have been done right first time if they'd been based on experience, rather than guesswork. Containers and collections could have had a more appropriate graph structure. Reification could have been done completely differently (or not at all, named graphs are ore useful). XMLLiterals could have had more relaxed rules. bNodes could have not been existential variables, just onetime constants, disjoint with URIs. As it is there's an awful lot of cruft in the RDF spec left over from previous working groups. I can't remember the number of times I've said things like “yeah, I know there's a features called RDF Reification, but you really don't want to use that”, or “there's two ways of writing down vectors in RDF, but neither will work well in this situation”. - Steve -- Steve Harris, CTO, Garlik Limited 1-3 Halford Road, Richmond, TW10 6AW, UK +44 20 8439 8203 http://www.garlik.com/ Registered in England and Wales 535 7233 VAT # 849 0517 11 Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD
Received on Sunday, 6 March 2011 18:58:07 UTC