W3C home > Mailing lists > Public > public-rdf-wg@w3.org > October 2011

Re: ISSUE-77: Should we mark rdf:Seq as archaic (cf ISSUE-24)

From: Andy Seaborne <andy.seaborne@epimorphics.com>
Date: Sat, 15 Oct 2011 20:24:23 +0100
Message-ID: <4E99DDE7.2030100@epimorphics.com>
To: Ian Davis <id@talis.com>
CC: "public-rdf-wg@w3.org" <public-rdf-wg@w3.org>


On 15/10/11 20:00, Ian Davis wrote:
> I consider that a feature not a bug. There are times when you want to
> rank things equally in a list. Its not that hard to handle in code
> with nested loops.
>

I agree there are can be such times -- and bags are useful sometimes as 
well (the opposite - no order, with duplicates).

rdf:Seq is not spec'ed as a rank though, only the order,

[
"the numerical ordering of the container membership properties of the 
container is intended to be significant."
]

and it says nothing about merging two graphs and dealing with Seq 
overlap. Best to use a blank node for the subject of the Seq.

You could subclass and impose addition definitions.  Are there any 
example of subclassing rdf:Seq?

It does go against the idea of making the encoded structures nicer to 
use by APIs if in some places the structures have a specific meaning of 
the offsets.  An API that presents a "Seq" may mangle them.

A bag of (rank, item) pairs would presumably be better but more complicated.

	Andy

(at least these could have been :1 :2 :3 in Turtle!)
Received on Saturday, 15 October 2011 19:24:55 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:45 GMT