Re: deterministic naming of blank nodes

Le 21/05/2015 18:37, a écrit :

>> FWIW I agree.  I'm pretty sure I advocated for .well-known at the
>> time (*ugh*), but in hindsight a URN prefix would have been a
>> better hack.
> Ah well. Never mind who advocated for it. I can see that the pressure
> of finishing the sepc may have had something to do with that.

In fact, no. The design for Skolem IRI has been discussed at the
beginning of the WG and was present in the First Public Working Draft of
RDF 1.1 Concepts, remaining pretty much unchanged in the 7 iterations of
public drafts and candidate rec and proposed rec until Rec.

> As long as it is something that we can acknowledge is not a good
> idea, this could be fixed  by putting a note in the current document
> that this is controversial, and perhaps provide a better solution in
> the next version.

A better solution could be devised in a future version, but nothing can 
be noted as controversial in a document that validated all the steps 
towards recommendation.

Controversial parts of a W3C document are marked explicitely with red 
boxes during the development of the spec. This one was never deemed 
controversial during the life-time of the WG, as far as I can remember. 
There has been debate on this, though, before the current scheme was chosen.

Besides, no one forces anyone to use the /.well-known/genid/ scheme. 
There is not even a MUST in the section. And there is no obligation to 
parse the IRI to find the .well-known part. Skolem IRIs are IRIs and 
should be treated as such.

> Clearly there is nothing wrong with skolemisation and stable blank
> nodes would be very useful in many cases, such as for PATCHes to a
> document.

Right, but I don't see where this last comment clashes with the spec.


> All the best,
> Henry
>> David Booth
> Social Web Architect

Antoine Zimmermann
ISCOD - Institut Henri Fayol
École des Mines de Saint-Étienne
158 cours Fauriel
CS 62362
42023 Saint-Étienne Cedex 2
Tél:+33(0)4 77 42 66 03
Fax:+33(0)4 77 42 66 66

Received on Friday, 22 May 2015 10:29:41 UTC