Re: Datatype support implementation matrix

Bijan Parsia wrote:
> 
> Mike Smith and I have started gathering information about existing 
> datatype (i.e., unary datapredicate) support in reasoners. We're 
> soliciting reasoner authors to fill in the following matrix:
>     http://spreadsheets.google.com/pub?key=pTmcCXR-dV6RpTEPxB0O-DQ
> 
> One thing we might consider is making some of the odder datatype 
> optional, e.g., Name and NCName. Similarly, we might be more explicit 
> about what facets are applicable, or, again, make some facets optional.

If a reasoner supports string and supports the facets with which Name 
and/or NCName are defined, then it should be strongly encouraged to 
support these. I tend to agree that they are not useful for typical 
semantic web apps, but in as much as the W3C should form a coherent 
whole, we should encourage SemWeb implementators to not wilfully not 
implement a part of some other spec.

> 
> This matrix could help drive test case generation as well.
> 
> We don't collect inline vs. external datatype support. AFAIK, only 
> Pellet supports pointers into XML Schema documents. I'd welcome 
> correction on this point!

Jena supports this too; I wouldn't be surprised if the Pellet support 
depends on the Jena support.


> 
> I don't think it's as interesting to gather data about editors because 
> it's much more trivial and it's pretty uninteresting to enumerate the 
> builtin types "supported" by an editor.
> 
> Cheers,
> Bijan.
> 
> P.S. We used a Google spreadsheet since 1) it's a heck of a lot easier 
> to manage than wikisyntax table and 2) we want reasoner authors who 
> aren't WG members to enter their own data :) We can always export and 
> script transform.

What about people without a google account?


Jeremy

Received on Wednesday, 14 November 2007 17:34:37 UTC