On Sep 14, 2010, at 09:15 , Toby Inkster wrote:
> On Tue, 14 Sep 2010 06:32:38 +0200
> Ivan Herman <ivan@w3.org> wrote:
>
>> I think that is the important point regarding the original issue.
>> There won't be _different_ triples (if the @profile is not loadable)
>> but _no_ triples. That was the issue of Jeni that led to this
>> resolution
>
> There will be different triples if the profile is loadable but no
> mappings can be extracted from it.
>
> e.g. if a profile defined a mapping for prefix 'dc' then the triple:
>
> <> <http://purl.org/dc/terms/title> "Example"@en .
>
> might become:
>
> <> <dc:title> "Example"@en .
>
> This situation could occur if the server hosting the profile serves,
> say, an error response (e.g. Not Found) with an incorrect HTTP status
> code.
Hm. I do not think I agree with your last statement, though I agree with the rest. An HTTP not found should mean the file is not loadable so the 'no triple generated' feature should kick in. But the situation can occur if there are errors in the profile file, ie, two URI-s are assigned the same prefix, which means that both mappings should be refused.
If we are really worried about that, I would not mind reinforcing Jeni's law, ie, if *any* error occurs in the profile file (as specified in the spec) than no triple should be generated. It sounds a bit harsh but may make sense...
Ivan
>
> --
> Toby A Inkster
> <mailto:mail@tobyinkster.co.uk>
> <http://tobyinkster.co.uk>
>
----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf