Re: PROV-ISSUE-233 (paq-dm-and-accounts?): If not in DM, should there be some form of account support in the paq? [Accessing and Querying Provenance]

Graham, Luc:

I think we should keep this issue open. I think the process should be as 
follows:

1) We agreed to revise or look at the service spec in the draft. This 
needs to be done first.
2) Once this revision is done, we should address this issue in 
particular with respect to a revised DM.

Otherwise, we are making changes against two things that are not yet 
fixed. Let's not premature in either discounting issues or making 
changes to address them.

Is that ok?
Paul


Luc Moreau wrote:
>
> On 02/08/2012 10:31 AM, Graham Klyne wrote:
>> Luc,
>>
>> On 08/02/2012 10:01, Luc Moreau wrote:
>>> Hi graham
>>>
>>> In short, to be able to retrieve prov from a service with the
>>> following parameters:
>>>
>>>    Http;//service/endpoint&entity=...&account=...
>> Hmmm, I think this is mostly covered by
>> http://dvcs.w3.org/hg/prov/raw-file/tip/paq/prov-aq.html#retrieve-provenance-information-for-a-resource
>>   (section 4.1.2).
>
> The service specification, as currently defined, only has one parameter
> (uri).
> My request is that we should have a second parameter (account).
>
> Of course, then, we need to look at ways of obtaining it.
>>> And have a mechanism to pass account ids.
>> This is the bit that I have trouble with.  What is this "account id"?
>> I think it's really inappropriate to introduce it as a concept in PAQ
>> because it's been dropped from DM.
>
> Not exactly.
> What is being dropped in prov-dm is AccountRecord  a construct that
> bundles up records *as part of the data model*.
> We have agreed that the bundling up is now done outside the data model.
>
> However, having named a bundle with a URI (with whatever mechanism we
> choose), we need to be able to express its
> provenance. It's then useful to have a subtype of entity (accountEntity
> or similar) for this very purpose.
> It will be in the working copy Paolo and I are currently working on.
>
>
> Luc
>
>> There is nothing in the specification that *prevents* you doing any of
>> this, but if the notion of account isn't part of the provenance model
>> it is bound to be a "private use" convention (which may become widely
>> used and a later candidate for standardization).
>
>
>
>> I think an appropriate way forward here would be to prepare a separate
>> document that covers the concept of accounts, account IDs and service
>> extensions to access provenance keyed by account.  This could be a
>> candidate for a WG NOTE (which we were told at F2F the WG may choose
>> to issue without reference to the charter).
>>
>> Which leaves me still inclined to close this "wontfix".
>>
>> #g
>> --
>>
>>> Professor Luc Moreau
>>> Electronics and Computer Science
>>> University of Southampton
>>> Southampton SO17 1BJ
>>> United Kingdom
>>>
>>>
>>> On 8 Feb 2012, at 08:35, "Graham Klyne"<Graham.Klyne@zoo.ox.ac.uk>
>>> wrote:
>>>
>>>> Luc,
>>>>
>>>> I find that don't really know what you are asking for here.
>>>>
>>>> I propose that either:
>>>> (a) you make a more concrete proposal that we can consider, or
>>>> (b) I close the issue "wontfix"
>>>>
>>>> In judging any new proposal, I would expect to apply the "razor" we
>>>> agreed in the F2F, in that if the proposal is not readily consensual
>>>> then we should err on the side of under-specification rather than
>>>> over-specification.  Generally, I think we should now be looking to
>>>> remove material from the documents rather than add it.
>>>>
>>>> #g
>>>> --
>>>>
>>>> On 07/02/2012 14:53, Stian Soiland-Reyes wrote:
>>>>> Then all bets are off beyond what you already know..
>>>>>
>>>>> Or do you mean paq could be a general URI resolver?
>>>>> On Feb 7, 2012 11:40 AM, "Luc Moreau"<L.Moreau@ecs.soton.ac.uk>
>>>>> wrote:
>>>>>
>>>>>> Hi Stian,
>>>>>> Sure, but what happens when accounts are identified by UUIDs ...
>>>>>> Luc
>>>>>>
>>>>>> On 02/07/2012 10:08 AM, Stian Soiland-Reyes wrote:
>>>>>>
>>>>>>> On Mon, Feb 6, 2012 at 10:35, Luc
>>>>>>> Moreau<L.Moreau@ecs.soton.ac.**uk<L.Moreau@ecs.soton.ac.uk>>
>>>>>>>    wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>> I already flagged the need of being able to retrieve the
>>>>>>>> provenance of an
>>>>>>>> entity in a given account.
>>>>>>>>
>>>>>>>>
>>>>>>> If you know the account/"provenance resource" URI, just fetch it.
>>>>>>>
>>>>>>> If not, then the PAQ will tell you about the provenance resources
>>>>>>> ("accounts") that it knows about. You can fetch them individually
>>>>>>> and
>>>>>>> choose yourself how you would like to separate or merge them.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>> --
>>>>>> Professor Luc Moreau
>>>>>> Electronics and Computer Science   tel:   +44 23 8059 4487
>>>>>> University of Southampton          fax:   +44 23 8059 2865
>>>>>> Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
>>>>>> United Kingdom
>>>>>> http://www.ecs.soton.ac.uk/~**lavm<http://www.ecs.soton.ac.uk/~lavm>
>>>>>>
>>>>>>
>

Received on Wednesday, 8 February 2012 11:00:56 UTC