Re: Comments on WD-rdf-testcases-20010912

>>>Bjoern Hoehrmann said:
> * Dave Beckett wrote:
> >These changes have been folded into
> >  http://www.w3.org/2001/08/rdf-test/
> 
> I'd still prefer to have a sole definition for N-Triples outside this
> specification, probably together with a formal specification of
> Notation3 (see my "Notation3 woes" [1]), but ok...

We aren't really worrying about the document structure at this time
(1st working draft) so N-Triples may be in a another document later.
Although N-Triples is a subset of Notation3, Notation3 is a different
semantic web research language, which is still changing, and the
authors are pretty friendly, so let them know.

> I still wonder whether N-Triple documents are considered to actually be
> written by humans or a plain RDF serialization format. If the former,
> I'd strongly recommend to allow usage of other encodings then US-ASCII.
> Note that if transferred via HTTP, those .nt files would need a
> Content-Type like
> 
>   Content-Type: text/plain;charset=us-ascii
> 
> since ommitted charset parameters on text/* types imply ISO-8859-1 as
> per RFC 2616. This would violate section 3 of the draft which requires
> US-ASCII. I don't know whether this is really a good idea at all.

We are considering the issues about making N-Triples using only UTF-8
encoding instead of ASCII - there are pros and cons, relationships
with XML Blueberry (since RDF is close to XML) and Notation3 to
consider.  This change will be in working draft #3 at the earliest.

> Is there any special reason why \U escapes need *8* digits even if only
> up to six digits are allowed?

I didn't change this since, as I note above, we might remove it.

> Anyway, the draft looks way better now.
> 
>   [1] http://lists.w3.org/Archives/Public/www-rdf-interest/2001Sep/0057.html

Thanks.

Dave

Received on Wednesday, 14 November 2001 07:02:20 UTC