Re: why S doesn't require double properties [was: Datatyping Summary V4]

On 2002-02-05 19:05, "ext Daniel W. Connolly" <connolly@w3.org> wrote:

> On Mon, 4 Feb 2002, Brian McBride wrote:
> 
>> At 11:15 04/02/2002 -0600, Dan Connolly wrote:
>> [...]
>> 
>>> Yes, *if* one wants to use both idioms, one needs both sorts
>>> of properties.
>> 
>> That's my understanding and what the issue states, I think clearly.
> 
> There's some lack of clarity somewhere; folks seem to be arguing:
> 
> if communities need to use both datatyping idioms,
> they need double properties.
> 
> PRISM needs datatyping
> => therefore
> PRISM needs double properties
> 
> which isn't a sound inference; it needs another premise:
> 
> PRISM needs to use both datatyping idioms.
> 
> I'm quite confident PRISM (and dublin core and most other apps)
> will do just fine only using S-B.
> 
> So I don't think the statement of the issue is sufficiently clear.

Do we really have to re-re-re-rehash this again?

Both global and local idioms are needed in order to express
constraints on locally defined types. It's not whether a given
ontology needs both idioms, but whether a given application
does.

In a datatyped world, such as eCommerce, we expect that
local typing is used. Then, we define global types that
function as constraints/expectations of that local typing,
and we trap contradictions/conflicts.

This is the way things work in eCommerce. Strict typing
is essential, and that means both local and global.

Unless we just toss out the whole point of descriptive
metadata and hardwire it all in our software...

You, personally, may not need to express such constraints,
but then, RDF is not just for you, is it?

Can we please close this issue? It's been beaten to death.

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Tuesday, 5 February 2002 13:43:19 UTC