Re: Alternatives to containers/collections (was Re: Requirements for a possible "RDF 2.0")

On Wed, Jan 20, 2010 at 3:04 AM, Jeremy Carroll <jeremy@topquadrant.com> wrote:
> Michael Schneider wrote:
>>
>> In /my/ experience, you will find for virtually every RDF feature some
>> gang
>> of people claiming that the feature is not helpful or even broken.
>
> In desperation, ... rdf:Alt ? :)
>

If one wishes to be principled, one  could either/both

1) Just deprecate all things (containers, reification, the use of
owl:import etc.) that have no working formal semantics, unless they
can be given some formal semantics that makes sense (i.e. building
some sort of ordered lists into RDF semantics if you want lists,
building the relatively straightforward named graph semantics into
RDF, etc.)

2) Make sure everything that is in the spec that people claim should
not be deprecated has some evidence of empirical - and one could add
consistent - use in the wild, rather than just people arguing for or
against it on listserv. If there are multiple inconsistent uses of
something in the wild (sameAs, seeAlso), then split the feature so
that it clearly matches the usecases people are actually using it for.

And then make sure things that people find difficult to do using
obscure features of RDF can be done simply (i.e. the question about
metrics discussed earlier, and even things like time, changesets, and
provenance if one wished to go really far)

I do agree some sort of principles are needed here rather than ad-hoc
updates and changes. Empricial driven changes with a clean semantics
sound like a place to start.


>
> Jeremy
>
>

Received on Wednesday, 20 January 2010 02:28:24 UTC