Re: Turtle implementation report for RDF::Trine

On Wed, 2013-04-17 at 23:12 +0900, Gregory Williams wrote:
> On Apr 17, 2013, at 11:02 PM, Eric Prud'hommeaux <eric@w3.org> wrote:
[...]
> I think having two different syntaxes for prefix declarations, with two different rules for the trailing dot, fails here.
[...]
> >  compatibility with SPARQL -- make it as easy as possible to copy stuff (triples and directives) between Turtle and SPARQL.
> 
> I'm sympathetic here, but for me the added grammar complexity required makes this a losing proposition.

I strongly agree with this.  Adding a new redundant prefix syntax to
Turtle that has absolutely no existing use in triple data anywhere is
ridiculous.

For better or worse, Turtle already has a prefix syntax, and giving it
two only makes it confusing, ugly, and inconsistent.  The very minor win
of copy pasting from a completely different language (SPARQL) is not
compelling enough.

> >  backward compatibility -- there's a *lot* of Turtle out there.

Exactly.

Publishing such data is just adding incompatibility to the world for no
benefit.  If specific implementations want to support *reading* this,
then they can.  They should NOT be encouraged to write it, and therefore
it should not be in the spec.

Lax reading, strict writing.  Strict writing of Turtle should always use
Turtle's existing and very well-established @prefix syntax.  It is
inappropriate to put completely new PREFIX and BASE syntaxes in the
Turtle grammar.

How this is still a debate is beyond me.  It's clear that many people
are not happy with this change.  Since the change is clearly
controversial, the appropriate thing to do is to just leave the grammar
as it was: directives in Turtle start with '@' and end with '.'.

-dr