Re: ISSUE-28 (was: Re: Draft agenda / final one in a few hours late)

Hi Sandro,

On 23 May 2012, at 13:49, Sandro Hawke wrote:
> it seems clear to me datasets with nesting can
> be reversibly rewritten in a standard way into datasets without nesting.
> The open questions to me are:
> (1) will we standardize on such a rewriting or leave it to
> applications;

I'd say it's not necessary, because

a) no one seems to be doing it at the moment yet, so there's no interoperability problem to solve
b) the use cases aren't terribly clear, so we can't be sure that we solve it the right way
c) it's likely that those who need it can realistically come up with a solution themselves

> There's some proposed text in my draft, but after writing it, I decided
> I was too far out on a limb and that I needed to implement it, which is
> mostly what I've been doing the past week.  (But I don't have anything
> to show for it yet.)

Well, even if we don't put it in the spec, this should be a good validation of the basic design, and could help us gaining confidence that the foundations are right.

> There is some danger that the rewriting gives us another abomination
> like RDF reification or lists.  I don't think so, but I'm not entirely
> convinced yet.

The bigger issue, I think, is that there are few use cases that people currently deeply care about that require this.


Received on Wednesday, 23 May 2012 23:16:27 UTC