Re: Turtle implementation report for RDF::Trine

On 18 April 2013 00:12, Gregory Williams <greg@evilfunhouse.com> wrote:

> On Apr 17, 2013, at 11:02 PM, Eric Prud'hommeaux <eric@w3.org> wrote:
>
> > I think everyone's gazing into their crystal balls and trying to figure
> out how to balance these competing constraints:
> >
> >  simplicity -- primarily, don't confuse authors. secondarily, don't be
> cruel to developers.
>
> I think having two different syntaxes for prefix declarations, with two
> different rules for the trailing dot, fails here. I've implemented this
> stuff many times over and use these syntaxes all the time, and I still fail
> to get the trailing dot syntax correct.
>
> Personally, the only issues that I have had have been since the experiment
with PREFIX/BASE in Turtle started. Previously both were separate and they
had different use cases, now Turtle seems to be inheriting part of the
SPARQL use case (although without making @base/@prefix case-insensitive, it
seems). Having different rules for the trailing dot, while also having
different rules for case-sensitivity doesn't look to be helping authors.


> >  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. And this wouldn't be a simple copy-paste
> job, either. It would require at least two copy-paste operations to avoid
> the SPARQL query type, projection, braces, etc. that come between the
> prefixes and any ground triples that appear in the query graph pattern.
>

As Greg says, you cannot paste a Turtle document into a SPARQL Update query
verbatim, so what is the benefit? It wouldn't only be two copy-paste jobs
if you have a Turtle serializer that creates a large number of prefixes
inside of the body as it needs to, based on new URIs that it finds. In
addition, Turtle parsers cannot be used to parse SPARQL Update queries,
that is what SPARQL parsers are for, and there are no cases where people
will be copy and pasting from SPARQL Update directly to Turtle.

It would be easier just to recommend that people use N-Triples if they want
to paste directly into SPARQL Update queries.

Making it easier to write SPARQL Update queries (by allowing Turtle @base
and @prefix inside of the query graph pattern for instance, which would be
enormously valuable!) is something that should have been done by the SPARQL
working group, not by the RDF/Turtle working group.

Why did the SPARQL working group syntax become static so early if this was
an immediately known issue for the RDF/Turtle working group back in 2011,
given that SPARQL-1.1 has only *just* gone through recommendation in 2013?

Cheers,

Peter