Re: Post-last-call comments on N-Triples notation

On Mon, 2003-06-30 at 21:16, Martin Duerst wrote:

[...]

> [Compared with the other issues that are currently being discussed,
> this may be a detail]
> The I18N WG (Core TF) has looked at the escaping conventions
> for N-Triples at http://www.w3.org/TR/rdf-testcases/#ntrip_strings.

Wow.  That's effort above and beyond in that we'd previously agreed that
n-triples, as a format for representing test cases, need not concern
i18n.

DanC has explained why it is the way it is.  Unless I hear a howl of
protest, I consider this comment closed.

> 
> We do not think it is a good idea to have more escaping than
> necessary. This only introduces additional uncertainties.
> It would be better if N-Triples used UTF-8 directly, therewith
> not testing any escaping conventions, but only the actual RDF
> statements and entailments.
> 
> 
>  From the MIT team meeting
> -------------------------
> 
> The question came up from a team member whether there was a
> difference between N-Triples and N3. It was pointed out that
> a quite probable way to read "N3" is as 'n-triple'

Really?

I've just asked half a dozen people to pronounce 'N3'; people unfamiliar
with SW.  They called it "n-three".  I asked several of them if they
might have said "n-triples" (out of earshot of the others) and the
general response was "don't be silly".

The participants were all brits except one american.  could be a
cultural thing I suppose.

This wasn't exactly a well designed human factors experiment - but I
found no evidence of confusion.

>  (rather than
> as 'n-three'). It was suggested that choosing a different name
> than N-Triple would avoid (or would have avoided) such kinds
> of confusion.

Changing a name can also cause confusion.  I've seen no evidence to date
of the present name causing confusion, nor is there anything in your
comment beyond speculation that it might.  N-triples has been so named
for quite a while now with no complaint.  Weighing the non-zero cost of
changing it together with the possibility of the change creating the
very problem that it seeks to avoid against the evidence that a real
problem exists, I think its unlikely that the WG will want to take this
comment any further.

Say now if this is not acceptable.

> 
> 
> Personal comment: Both of the above comments seem to at least
>                    point to the need for a strong warning
>                    "Warning: N-Triples in not N3".

http://www.w3.org/TR/rdf-testcases/#ntriples begins:

[[
N-Triples is a line-based, plain text format for encoding an RDF graph.
It was designed to be a fixed subset of N3
]]

Is that sufficient?

Brian

Received on Tuesday, 1 July 2003 10:50:30 UTC