RE: PROV-DC Note ready for internal (final?) review

Hi Dani,

Regarding referencing and derivation, I think that if the document said "The construction of this document (Doc1) was not based on REF 1" it would simply be an untrue statement. In order to reference REF 1, you need to give identifying information about it within Doc1. Therefore, Doc1's construction (its bibliography if nothing else) is based on pre-existing entity REF 1.

Your argument seems to be assuming a particular form of derivation, where sentences within Doc1 were quoted, paraphrased or inspired by sentences in REF 1. But I'd argue this is not the only way that one entity may be derived from another, and I think the "based on" part of the wasDerivedFrom definition tries to account for this.

I'm also still unclear what the PROV relation between Doc1 and REF 1 would be if it is not derivation. As Doc1's existence required REF 1's existence (else Doc1 wouldn't be able to refer to it), there must be some provenance relationship, surely?

thanks,
Simon

Dr Simon Miles
Senior Lecturer, Department of Informatics
Kings College London, WC2R 2LS, UK
+44 (0)20 7848 1166

Mapping Dublin Core (Attribution Metadata) to the Open Provenance Model:
http://eprints.dcs.kcl.ac.uk/1386/
________________________________
From: dgarijov@gmail.com [dgarijov@gmail.com] on behalf of Daniel Garijo [dgarijo@delicias.dia.fi.upm.es]
Sent: 22 February 2013 16:17
To: Miles, Simon; Kai Eckert; Kai Eckert
Cc: <public-prov-wg@w3.org>
Subject: Re: PROV-DC Note ready for internal (final?) review

Hi Simon,
thanks for your quick response!

2013/2/22 Miles, Simon <simon.miles@kcl.ac.uk<mailto:simon.miles@kcl.ac.uk>>
Hello Daniel, Kai,

Thanks for the detailed responses and changes. I've skimmed the document, and it looks good. I'm also happy with the updates made in response to my comments. I have two remaining comments below. Neither are blocking issues and I'm happy for the note to go to FPWD.
Great!

1. In the abstract, one sentence doesn't quite make sense: "The direct mappings are broad and usually a more complex mapping can be specified," It is not clear what you mean by "broad", or why this is connected to specifying complex mappings. "Broad" is generally a positive term, while "complex" is a negative one. Intuitively, if the direct mappings are "broad", then this means they apply to many cases, but if so, why would you want to specify complex mappings? Complex mappings can be specified but, given that we want to avoid complexity, why do you need to? It needs to be rephrased.
I will look into it.

2. I still can't understand how dct:references can be not a subproperty of prov:wasDerivedFrom. Expanding the example in your response, the complete document DOC1 might say:
  "This work has nothing to do with this random reference [REF1].
   [REF1] Kai and Daniel, Discussion on derivation, 2010."
First, DOC1 can only have the content it does if REF1 first exists, so REF1 is part of the provenance of DOC1. If the connection between the two, as expressed in PROV, is not derivation, what is it? Furthermore, DOC1 has to identify REF1 in order to reference it, e.g. this is done by title, authors and year above, and these identifiers are part of what REF1 is. Therefore, how can DOC1 not be derived from REF1?

Regarding the definition of derivation, "a transformation of an entity into another, an update of an entity resulting in a new one, or the construction of a new entity based on a pre-existing entity", it is clear that this case is not a transformation or an update of an entity into another. So we are discussing the third part (highlighted). My concern here (and Michael's) is that stating a derivation is kind of contradictory. What if the text was "The construction of this document (Doc1) was not based on REF 1". Stating the derivation would imply exactly the opposite. Derivation is too strong in this case, don't you think?.

Best,
Dani

thanks,
Simon

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

Mapping Dublin Core (Attribution Metadata) to the Open Provenance Model:
http://eprints.dcs.kcl.ac.uk/1386/
________________________________
From: dgarijov@gmail.com<mailto:dgarijov@gmail.com> [dgarijov@gmail.com<mailto:dgarijov@gmail.com>] on behalf of Daniel Garijo [dgarijo@delicias.dia.fi.upm.es<mailto:dgarijo@delicias.dia.fi.upm.es>]
Sent: 20 February 2013 17:19
To: <public-prov-wg@w3.org<mailto:public-prov-wg@w3.org>>; Ivan Herman; Miles, Simon; Luc Moreau
Subject: PROV-DC Note ready for internal (final?) review

Hi all,
Kai and I have gone through the issues and we have given a pass through the whole document.
I have reestructured it and now I think it reads better.
The latest version can be accessed at:
https://dvcs.w3.org/hg/prov/raw-file/default/dc-note/dc-note.html

A detailed answer to the reviews made by Simon and Luc can be accessed here:
http://www.w3.org/2011/prov/wiki/AnswersToProvDCReviewers (on the bottom of the page)

The wiki page also summarizes the decissions over the main changes proposed to the mapping.
All issues are now pending review. Once I get the confirmation from Simon and Luc, I'll proceed to stage
the note.

@Ivan: I need to update 2 documents that are linked from the note (in particular
http://www.w3.org/ns/prov-dc-directmappings.ttl and http://www.w3.org/ns/prov-dc-refinements.ttl.
Both have changed a little bit). Who should I contact to do so?
The right versions can be accessed at https://dvcs.w3.org/hg/prov/file/tip/dc-note/files/prov-dc-directmappings.ttl
and
https://dvcs.w3.org/hg/prov/file/tip/dc-note/files/prov-dc-refinements.ttl respectively.

Best,
Daniel