[doap-interest] Bug: doap:os

[Damian Steer]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

[bah, forgot to Cc list]

On 28 Jul 2004, at 19:47, Edd Dumbill wrote:

> On Wed, 2004-07-28 at 14:34 -0400, Norman Walsh wrote:
>> / Damian Steer <pldms at mac.com> was heard to say:
>> | "Operating system that a project is limited to. Omit this property 
>> if
>> | the project is not OS-specific"
>> |
>> | Omitting a property to mean negation (i.e. not OS specific) is a 
>> mistake
>> | given the open world of RDF [*]. I'd suggest changing the 
>> rdfs:comment
>>
>> Good catch.

Credit to Morten's schema viewer, and Edd for linking to it in his 
article. I'm lazy :-)
...
>> And I wonder if we shouldn't have a list of URIs for languages and
>> platforms, just like we now have a list for licenses.
>
> I'm not sure about that, primarily as I don't want to maintain it.

I understand that. One option would be using a URI, but not maintaining 
a list. Is that any worse than uncontrolled literals? I'm not sure. 
Perhaps there's more chance of convergence with URIs.

If some brave soul wants to categorise those URIs later then they can, 
which would be harder with literals.

Another option is indirection via an intermediate node to a literal. 
The XML isn't so bad using this (unusual) rdf construct:

<doap:os doap:osName="...whatever..."/>

If people want add more info on that intermediate node then there's 
space. It doesn't even need to be a blank node, which gives us the best 
(or is it worst?) of both worlds.

Anyway, there are a couple of ideas.

Damian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iD8DBQFBCB3qAyLCB+mTtykRAq71AJ9QsXc18zbHcsFYlmyZRmkFo2CZPgCgtBWW
0Lc/SPquFweuMc0d5X5c4W4=
=zDPQ
-----END PGP SIGNATURE-----