- From: Graham Klyne <GK@ninebynine.org>
- Date: Mon, 27 Jun 2011 00:04:16 +0100
- To: Simon Miles <simon.miles@kcl.ac.uk>
- CC: Provenance Working Group WG <public-prov-wg@w3.org>
Simon Miles wrote: > Hello PAQ task force, all, > > As said in the telecon, we would like to identify clear points > regarding access which we can aim to decide on in the limited time > available at the F2F1. As there is not yet clear consensus regarding > proposed mechanisms, these points may be mechanism-independent, but > should help us move towards agreement on the technical details. If > there is time available, we can then have open discussion on the range > of mechanisms proposed. > > To start, I suggest the following points, all raised on the Wiki or > mailing list. It would be great if any of you could: > - Comment on the overall approach of starting with such points at F2F1 > - Suggest more such points > - Say whether the points suggested below are adequately precise > Note that it is not our aim to debate/decide on these points ahead of F2F1. > > P1. There may be data regarding the provenance of a thing accessible > from multiple sources. Yes. > P2. The information required to obtain access to some provenance of a > thing may be supplied in many different ways, and we do not aim to > enumerate them all. Yes. > P3. The WG effort will concern how the provider of a thing can supply > information required to obtain access to some provenance of that thing > (which may, as a side effect, include recommendations on how others > can do the same). I think the provider is a useful starting point. > P4. Regarding some provenance data obtained from dereferencing a > provenance URI, calling a provenance service, or some other means. > Which of the following is true? > (a) It is apparent from the data itself what single thing it > describes the provenance of. Isn't that a question for the model task force? (Though I think the answer should be yes.) > (b) Provenance data documents a set of things and how they are > related by past occurrences, so to extract the provenance of any one > thing requires knowing how it is identified in the data. I don't fully understand this; it seems rather non sequitur, though I think I agree with the conclusion. This also feels as it veers close to model territory and/or application design. > (c) Something else. > > P5. Regarding some provenance data obtained from dereferencing a > provenance URI, calling a provenance service, or some other means. > Which of the following is true? > (a) To meet the standard, it should be immutable. > (b) It can change over time without restriction. > (c) To meet the standard, there are particular ways in which it > should not change, e.g. any one account should remain as it was. I don't know. The standard should specify what is needed for interoperability. It may be that meaningful applications have to assume/assert further truths. I think the answer is (b) or (c). #g --
Received on Sunday, 26 June 2011 23:14:01 UTC