- From: Sandro Hawke <sandro@w3.org>
- Date: Sun, 16 Oct 2011 23:27:46 -0400
- To: Pat Hayes <phayes@ihmc.us>
- Cc: Andy Seaborne <andy.seaborne@epimorphics.com>, public-rdf-wg@w3.org
On Sun, 2011-10-16 at 20:13 -0500, Pat Hayes wrote: > > How about adapting some kind of bracketing convention, maybe [[ a b c > ]] for a list with three items, and having SPARQL treat [[?x]] as a > variable list, ie it is required to match a (wellformed, simple, > proper) list? That seems very intuitive to me. > Yeah.... How about Prolog lists: [] is the empty list [a,b,c] is the list of a, b, and c [a|x] is the list with a first of a and rest of x [a,b,c|x] is the list with the first being a, second being b, third being c, and the rest being x Now, in Prolog you never need it, but it seems simple enough to imagine [|x] would be the list x, which differs from x in that it's forced to be a list, which is what we wanted here. Turtle/SPARQL use (a b c) instead of [a,b,c]; I'm not sure if "|" is still available to use this way or not. A simpler approach would just be to rev the results format, with some element in the protocol indicating you can handle the 1.2 results format. -- Sandro
Received on Monday, 17 October 2011 03:27:57 UTC