RE: prov-issue-11: what is version?

A couple of comments on the last few emails:

 

There are two concepts running around - version, which is parent-child, and revisionOf, which is sibling/peer. Both are useful and one is basically derivable from the other, though going from version to revision requires information about the order/branch structure of versions to be captured somewhere (and going the other way requires knowing the parent...). Another useful notion I've seen in versioning models, also derivable, is the idea of the 'current version'.

 

While we'd all like version/revision to be something we can derive from more basic provenance information, I don't think we can make that work. Versioning is really an intellectual decision, not a direct result of processing. Or said differently in a way that connects to the document/file IVP relationships - versioning is a part of the lifecycle of documents whereas edits/byte corruption, copy, and other operations occur on files. Without someone specifying (asserting) which operations maintain the link between document and file, and which processes (probably even which process executions) cause versions to occur, the processing history of the file is potential evidence for versioning happening, but not something we can directly infer from. I can edit a file and create a new version of my document, or I can edit a file and create a new document - I don't think you can come up with a rule that would allow you to distinguish these without asking me.

 

This is basically why I see versioning as a separate ontology that we could connect via a profile along the lines of the way we discussed time. Version 2 shouldn't be appearing before version 1 was created according to our provenance information, version 3 should have some provenance connection to version 2 (perhaps you just told me about your text on the phone and I started with a new empty file to create it, so we can't infer much about the connection). The document lifetime should span that of the versions (document must be generated before/at the same time as version 1, etc.) 

 

Jim

 

From: public-prov-wg-request@w3.org [mailto:public-prov-wg-request@w3.org] On Behalf Of Thomas FRANCART
Sent: Wednesday, June 29, 2011 8:31 AM
To: Provenance Working Group WG
Subject: Re: prov-issue-11: what is version?

 

Hi Simon, Luc

I feel users/implementors of the PIL will need to express versioning information; at least I know I need it :-) now whether that notion of version is part of the PIL, a profile of the PIL, or a mapping to another vocabulary is another question.
I like Luc's definition.
Here are some comments on the notion of version (that are "engineering-driven" I would say) :

*	provenance/history information can be automatically recorded by an application, while versioning (in the way I see it) requires someone to say "I have reviewed all the modifications and I consider that this artifact is now of version x+1" (there is an explicit action involved to create a new version);
*	a version is (I think) associated with the notion of publication of the artifact (even if some modification happened to version x, you don't consider it being version x+1 until you decide to publish it);
*	there is somehow a link between all the atomic modifications made between version x and x+1, and version x+1 (so that you can compute/display the difference between 2 versions);

Cheers
Thomas

On Wed, Jun 29, 2011 at 11:42, Simon Miles <simon.miles@kcl.ac.uk> wrote:

Hi Luc,

OK. I'm fine with your definition if the group considers we need one.

As a more general point, I'm unclear about the restrictions of the
standardisation process. Is there no way to say "this proposal in the
charter was considered, but was decided to be ambiguous and/or
superfluous"? Otherwise, WGs must risk producing standards which are
compliant to the letter of their charters, rather than good and easy
to adopt. Where we can say less, there is less for people to read,
misinterpet and disagree with.

Thanks,
Simon

On 28 June 2011 17:30, Luc Moreau <L.Moreau@ecs.soton.ac.uk> wrote:
>
> Hi Simon,

>
> If we agree with the definition I suggested (possible Jim's too, I am
> not sure), it shows that version (or is revision of) is not a primitive
> notion in PIL, but can be derived from more primitive concepts.
>
> I think we still need to take a view on this concept, since it is part
> of the charter,
> and we can't simply ignore it.
>

> Luc
>
>
>
> On 06/28/2011 03:36 PM, Simon Miles wrote:

>> OK, but I think that defining it generally even in a profile may go
>> too far. Given that "version" means quite different things in
>> different application contexts, as I think you capture by the notion
>> of typed process executions in your definition, is there a value in
>> defining it generally at all? I could imagine it may be defined in
>> various ways in a few different domain-specific profiles, and there
>> could be a mappings from the PIL model to version in DC and elsewhere
>> etc., but defining it as part of the model seems to help no-one while
>> adding to the complexity. This differs from time, where though it has
>> different conceptions in different domains, I could imagine a default
>> conception defined in a profile would be useful for applying the model
>> to common kinds of web resource.
>>
>> Thanks,
>> Simon
>>
>> On 28 June 2011 15:25, Myers, Jim<MYERSJ4@rpi.edu>  wrote:
>>
>>> I was just trying to use version as an example of IVP in the last email, hence it shouldn't be different. Looking at whether we need version explicitly as a concept - perhaps it is a 'profile' like time...
>>>
>>>   Jim
>>>
>>>
>>>> -----Original Message-----
>>>> From: public-prov-wg-request@w3.org [mailto:public-prov-wg-
>>>> request@w3.org] On Behalf Of Simon Miles
>>>> Sent: Tuesday, June 28, 2011 10:18 AM
>>>> To: Provenance Working Group WG
>>>> Subject: Re: prov-issue-11: what is version?
>>>>
>>>> Luc, Jim,
>>>>
>>>> I notice that you two take different views on what the concept "version" is
>>>> intended to describe. With the example things:
>>>>    T1. The government data
>>>>    T2. The government data with incorrect values
>>>>    T3. The government data with corrected values Under Luc's definition T3 is
>>>> a version of T2, but under Jim's definition T3 is a version of T1.
>>>>
>>>> I'm not clear that "version", under either definition, is beneficial to keep in
>>>> the model. Jim's definition seems to be only subtley if at all different from
>>>> IPVT, while Luc's is distinct but just a simple composition of other concepts
>>>> which could be recognised by any query.
>>>>
>>>> My counter-proposal would be to remove "version" from the model.
>>>> Simplicity of the standard is surely a good thing where possible.
>>>>
>>>> If that is unacceptable, I think that Luc's definition makes sense but would
>>>> be more clearly called "is revision of" or similar.
>>>>
>>>> Thanks,
>>>> Simon
>>>>
>>>> On 27 June 2011 16:11, Luc Moreau<L.Moreau@ecs.soton.ac.uk>  wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Exploiting the  most recent definitions of Derivation and IVP of, I
>>>>> tried to propose a definition of version.
>>>>> http://www.w3.org/2011/prov/wiki/ConceptVersion#Definition_by_Luc
>>>>>
>>>>> What do you think?
>>>>> Regards,
>>>>> Luc
>>>>>
>>>>> --
>>>>> Professor Luc Moreau
>>>>> Electronics and Computer Science   tel:   +44 23 8059 4487 <tel:%2B44%2023%208059%204487>  University
>>>>> of Southampton          fax:   +44 23 8059 2865 <tel:%2B44%2023%208059%202865>  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/%7Elavm> 
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>> ______________________________________________________________
>>>> ________
>>>>
>>>>> This email has been scanned by the MessageLabs Email Security System.
>>>>> For more information please visit http://www.messagelabs.com/email
>>>>>
>>>>>
>>>> ______________________________________________________________
>>>> ________
>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Dr Simon Miles
>>>> Lecturer, Department of Informatics
>>>> Kings College London, WC2R 2LS, UK
>>>> +44 (0)20 7848 1166 <tel:%2B44%20%280%2920%207848%201166> 
>>>>
>>>
>>> ______________________________________________________________________
>>> This email has been scanned by the MessageLabs Email Security System.
>>> For more information please visit http://www.messagelabs.com/email
>>> ______________________________________________________________________
>>>
>>>
>>
>>
>>
>
> --
> Professor Luc Moreau
> Electronics and Computer Science   tel:   +44 23 8059 4487 <tel:%2B44%2023%208059%204487> 
> University of Southampton          fax:   +44 23 8059 2865 <tel:%2B44%2023%208059%202865> 
> 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/%7Elavm> 
>
>
>
> ______________________________________________________________________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email
> ______________________________________________________________________
>



--
Dr Simon Miles
Lecturer, Department of Informatics
Kings College London, WC2R 2LS, UK
+44 (0)20 7848 1166 <tel:%2B44%20%280%2920%207848%201166> 




-- 
Thomas Francart
CTO - Mondeca
3, cité Nollez 75018 Paris France
Tel. +33 (0)1 44 92 35 04 - fax +33 (0)1 44 92 02 59
Web: www.mondeca.com
Blog : Leçons de choses <http://mondeca.wordpress.com>