Re: "Language-tagged strings Re: Toward easier RDF: a proposal"

Hi,

> On 25 Nov 2018, at 20:21, Andy Seaborne <andy@seaborne.org> wrote:
> 
> Hugh, Christian,
> 
> 
> You can do what you describe already : see for example SKOS-XL which also discusses some issues.
> 
> The web already has language tags RFC5646 and I guess that is how they ended up in RDF via xml:lang and HTML. We should work with and use the outputs of these communities, not redo their work.
Just to be clear: I agree - touching the language tags as defined by IANA or whoever would be a mistake.
Mind you, that RFC makes no mention (as far as I can see) of where the tag appears.
It's the XML inheritance that I have problems with, I think.
> 
> Often that's enough, and a concept general developers have come across (it's so common it is in standard programming language runtime libraries).
> 
> Sometimes, in some cases, more is needed and that is possible.
> 
> 
> The case for "Triples and only triples":
> 
> We've done that before - reification, RDF containers (alt/seq/bag), RDF lists - all using the "triples everywhere" approach.  Nice conceptualization but hasn't worked out in practice.
> 
> A: Acceptance: General developers don't relate to the idea. They expect lists etc, not more triples that have to put back together again.
> 
> B: Inconsistencies and partial representations (see SKOS-XL) make for hard development (SHACL/ShEX would help). Support in toolkits is a burden.
> 
> C: Implementation and deployment: More triples is a burden.
> 10 million -> 15 million - Meh.  10 billion -> 15 billion - that does make a difference.
I'm not sure that equivalence arguments such as those apply.
And one extra triple per xsd:string is not on anything like those scales.

But the world ain't a perfect place, as we know :-)

Thanks for the discussion.

Best
Hugh
> 
>    Andy
> 

-- 
Hugh
023 8061 5652

Received on Sunday, 25 November 2018 22:44:49 UTC