Re: URI for language identifiers

Bill de hÓra wrote,
> Here are two:
>
> * denotation by authrity is web architecture, according to a number
> of TAG members including the W3C Director. I don't know if the TAG
> will ever issue an actual edict on this; I only know of one member
> who has stated publicly he doesn't buy denotation by authority. It's
> a TAG permathread, see the archives.

No. Denotation by authority is a _candidate_ for being recommended web 
architecture, and, as the TAG permathreads illustrate, is a thoroughly 
contested candidate at that. There's no consensus and no edict and, 
IMO, there's never likely to be one either.

Sure we can read tealeaves and try and second-guess the Directors 
opinion du jour but, frankly, I think we've all got better things to do 
with our time.

> * case law, in cases involving deep linking set a precedent. It only
> takes a libel action to go from 'you can't link there' to 'you can't
> say that'.

Right. And it illustrates why we can't take the TAG's rumblings too 
seriously here. On the one hand we have a chunk of the TAG supporting 
denotation by authority, and on the other hand we have a TAG consensus 
that deep-linking is OK, hence that denotation by authority is bunk. 
This is emphatically _not_ a narrowly technical issue, and expecting a 
useful, consistent and authoritative statement from a technical body is 
foolish.

> In particular, I think the first makes the semantic web slightly
> crocked from an engineeering perspective - well and good to for the
> RDF MT to say that all URIs must have one denotation in the graph,
> well and good for some to say URIs are owned threfore their
> denotations are owned thanks to some axiom or other. But there is a
> whole layer of infrastructure to be put into place to assign
> interpretations to URIs before safely merging graphs, if Miles is
> correct.

And most of that infrastructure is social, political and economic rather 
than technical.

> The second is a potentially nasty aspect of the semantic web we'll
> be faced with before the decade's out. Until the lawyers come
> knocking no doubt we'll ignore this one as a non-technical issue.
>
> > I don't think so.  I think that Mile's comments are exactly
> > correct.
>
> Miles I belive is technically correct, but needs to take ICANN into
> account. Domains name are a form of property, sufficiently so you
> can go to court over them.

Oh, but I do take ICANN into account (I think you've always 
underestimated my political nous ;-). The contrast here is informative: 
ICANN's authority, and domain ownership case law cover domain names, 
*not* URIs. If there were an equivalent of ICANN and this case law for 
URIs, then Patrick might have a point.

But there isn't, so he doesn't.

Cheers,


Miles

Received on Wednesday, 2 April 2003 19:54:08 UTC