Re: Working list of BP requirements identified during the SDW WG f2f

> [snip]
> 1.1 URIs should be stable/persistent - I have seen suggestions that URIs can
> be easily minted using database keys. The risk with these is that keys are
> probably less stable than the things identified.
> Yes, that is a suggestion I made when the  discussion drifted towards
> already finding answers for the requirements that we were supposed to
> harvest. I think using database keys defers the problem of finding a scheme
> for minting URIs to the scheme that is used in the relational database. In
> some cases, database keys will be designed and assigned is such a way that
> persistence is guaranteed. In other cases not. But still, they are a good
> candidate to be looking at when thinking of a way to mint URIs.
> But more importantly, I think we will have to decide not to view a
> requirement like 'there should be a best practice for minting URIs of
> spatial phenomena' as something we should busy ourselves with. I can think
> of no reason to see the URI minting problem as something that is inherently
> spatial, temporal or spatiotemporal. Besides that, there already are some
> good guidelines on minting URIs out there.

Just to add the following:

During the UC review carried out by the BP group, one of the points
that were made was that requirements should be grouped based on
whether they were specific to spatial data or concerning any kind of
data on the Web.

The BP reqs page on the wiki reflects this "classification" [1].

PIDs can well be considered as a common requirement for any data on
the Web - and, actually, the DWBP WG covers this in BP-8:

BTW, the recent publication of the DWBP WG working draft raises the
question on how we plan to integrate their BPs (although still
unstable) in our work - especially on the scope of our BPs.




Received on Tuesday, 17 March 2015 13:42:18 UTC