- From: Sandro Hawke <sandro@w3.org>
- Date: Fri, 08 Apr 2011 13:14:03 -0400
- To: Dan Brickley <danbri@danbri.org>
- Cc: Richard Cyganiak <richard@cyganiak.de>, Steve Harris <steve.harris@garlik.com>, RDF Working Group WG <public-rdf-wg@w3.org>, RDF Working Group Issue Tracker <sysbot+tracker@w3.org>
On Fri, 2011-04-08 at 13:28 +0200, Dan Brickley wrote: > On 8 April 2011 11:23, Richard Cyganiak <richard@cyganiak.de> wrote: > > On 8 Apr 2011, at 10:05, Steve Harris wrote: > >>> rdf:Seq does have some merit, and some serious 'in the wild' usage. > >> > >> Agreed. While far from perfect, in some situations it's preferable to RDF Lists. > > > > Can you give examples where rdf:Seq is preferable to rdf:List? > > :_1 when the Seq numbers shadow some real world assignment eg room > numbers, and each description only mentions a few In that case, I think it would be much, much better to put those room numbers in the data as real data and not use a Seq or a List. > :_2 when each triple costs, eg. on pay as you go bandwidth (eg mobile > web in india people are), or huge datasets I've never heard of paying by the triple. I think everyone charges by the byte or packet or something like that. So if you use Turtle with the "(...)" construct, RDF Collections will be much cheaper than RDF Containers. If you use RDF/XML then I expect parsetype=Collection would solve your problem as well. It's mostly in N-Triples that you'd have a problem, and I don't think you'd use that if you were paying for bandwidth. > :_3 in Rdfa where the list structure has no sugar syntax ( does 1.1?) *shrug* This seems like something to address in RDFa, then. > :_4 (maybe? unenthusiastically?) when dealing with legacy code that > has special support for Seq, eg. old Mozilla XUL stuff That's a strong argument against removing it. But not one is suggesting removing it. The issue is whether to deprecate it, and current usage doesn't have a bearing on that. -- Sandro
Received on Friday, 8 April 2011 17:14:18 UTC