- From: Peter Frederick Patel-Schneider <pfps@research.bell-labs.com>
- Date: Tue, 1 Mar 2011 13:46:16 -0500
- To: <andy.seaborne@epimorphics.com>
- CC: <sandro@w3.org>, <nathan@webr3.org>, <public-rdf-wg@w3.org>
From: Andy Seaborne <andy.seaborne@epimorphics.com> Subject: Re: [TURTLE] Extending Turtle before Rec Date: Tue, 1 Mar 2011 12:41:11 -0600 > On 01/03/11 17:15, Sandro Hawke wrote: > >> I think there's a parent-issue here, which we should perhaps raise and >> even decide first, of whether we're going to be extending turtle now, >> making a superset which will not work in existing turtle parsers. >> Alternatively, we could nail down the Turtle that is already widely >> deployed. I guess the issue might be called "Extend Turtle Before >> Rec?". >> >> I'm torn on this issue. On the one hand, there's a big deployed base, >> to which we have a responsibility. On the other hand, there a lot more >> that can be done to make Turtle useful. For myself, I lean towards >> saying "no", considering Turtle more or less done, and letting >> extensions happen in other languages. >> >> If we say "no", then our work is much simpler; we're basically operating >> inside the space of existing good faith implementations. > > Good points - I think we definitely need to do "existing turtle" and > maybe "superTurtle" so at least keep it as two strands. > > It's only really possible to see if superTurtle is a good idea when all > the features are defined so we can see the interactions. That takes > time so it's not a good idea to depend on that as the outcome. > > Andy It's turtles all the way down.
Received on Tuesday, 1 March 2011 18:47:11 UTC