Fwd: The wasQuotedFrom relationship

Simon,

I'm directing comments to you in hopes that the wg can develop another response.

Begin forwarded message:

> Resent-From: public-prov-comments@w3.org
> From: "Morris, Chuck E (IS)" <chuck.morris@ngc.com>
> Subject: RE: The wasQuotedFrom relationship
> Date: February 7, 2013 2:18:48 PM EST
> To: "Miles, Simon" <simon.miles@kcl.ac.uk>, "public-prov-comments@w3.org" <public-prov-comments@w3.org>
> 
> Thanks for the response.  I can see an attempt to deal with the issue, but I’m not convinced that it has been resolved satisfactorily.  I still think there is a fundamental problem with the wasQuotedFrom relationship name.  I recognize that the pattern “X was quoted from Y” is used in colloquial language to attribute quotations,


This is exactly the intent of the relation. Some [quote] Entity is taken from some previous [larger] Entity.

I just copied Chuck's name from his own email:

[ a prov:Entity; 
   prov:value "Chuck"; 
   prov:wasQuotedFrom <http://www.w3.org/mid/B7183A6EEAB3B049B77D39BDE3F9D70F485E3259@XMBVAG71.northgrum.com>;
] .

<http://www.w3.org/mid/B7183A6EEAB3B049B77D39BDE3F9D70F485E3259@XMBVAG71.northgrum.com> a prov:Entity, :Email .




> but only when it is clear that X is a quotation. 


The use of the property implies that X is a quotation, by the definition of the property.
The domain of the property is not constrained because it wasn't necessary -- Entity is sufficient.



> No one would say “X was quoted from Y” if X is not a quotation,

+100

> but they may well say “X was quoted from in Y”, meaning Y contains a quote that came from X.

We do not offer this relation.
@Simon, is the tutorial leading people down this path? If so, I think it should simply be removed entirely.
The example seems to be misinterpreting the relation.
And I don't' think that a "scruffy" argument makes sense. 
I'm as scruffy as they come, and even THAT distinction (which involves a whole-part indirection) doesn't make sense to me.
When the part is quoted from, it does not imply that the whole is quoted from. It's simply too degenerate.


>  If you see “X wasQuotedFrom Y” and you do not know that X is a quotation,

You **DO** know that X is a quotation, simply by the fact that it wasQuotedFrom something. 
Chuck's assumption here is invalid, so the remaining argument does not work.

> I think it is more natural to assume that it means “X was quoted from [in] Y”


Yes, this would "swap" the direction of the "wasQuote" relation, where X is the big, earlier entity and Y is the "blog that includes a chunk of text from X".
BUT THIS IS NOT WHAT THE RELATION IS.

> than to assume that it means “X [is a quote that] was quoted from Y”.

NOR IS THIS.
I think the primer example misleads to this interpretation. 
Can we not directly describe the actual quoted value 

    :my_clipping dcterms:isPartOf :my_blog with prov:wasQuotedFrom :bible ?


>  That is especially true in scruffy usage when X is not a quote, but merely contains a quote.

This is NOT a valid "proper" or "scruffy" interpretation.


> I see from the discussion that quoted, wasQuoteOf, hadQuoteFrom, wasAQuoteFrom, isAQuoteFrom, and isQuoteFrom were all rejected.  How about just quoteFrom?  That isn’t past tense, but it seems to me that it would be consistent with some of the other relationship names like alternateOf and specializationOf.

It would be inconsistent with its own super property, wasDerivedFrom (where Quoted is a type of Derivation).


> And perhaps it would also be helpful to either change ex:blogEntry in the primer to ex:quoteInBlog to avoid scruffy usage,

+10000 (I think this was my biggest vote ever…)

> or add text to point out that the usage of the relationship is an example of scruffy usage.

IT IS NOT A GOOD USE OF SCRUFFY. And I don't think this argument can be used to violate the intent of the wasQuotedFrom.
I think it's best to pair an entity that was quoted from another entity with the prov:value of the quoted string, to make its use obvious.

-Tim


> Chuck
>  
> From: Miles, Simon [mailto:simon.miles@kcl.ac.uk] 
> Sent: Thursday, January 24, 2013 11:08 AM
> To: Morris, Chuck E (IS); public-prov-comments@w3.org
> Subject: EXT :RE: The wasQuotedFrom relationship
>  
> Hello Chuck,
>  
> Thanks for your comment. The Provenance Working Group has discussed this, and prepared a response
>  
> The main point is that we think this is just the primer text being misleading rather than the relation name being incorrect. The wasQuotedFrom relation should link a quote to the document it was quoted from. The primer currently can be read as linking something *containing* a quote to the place it was quoted from, which is allowable under "scruffy" use of PROV, but not ideal for illustrating the concept as it doesn't match the relation name, as you indicate. More generally, the working group previously extensively discussed the matter of the relation name, including considering hadQuotationFrom. While no relation name may be perfect, it was agreed wasQuotedFrom matches the intent of the relation and PROV-DM definition better than hadQuotation>From or other relations.
>  
> Full details of the response, and the clarifications we intend to make to the primer, are below
>  
> http://www.w3.org/2011/prov/wiki/ResponsesToPublicCommentsCR#ISSUE-616
>  
> Can you let us know whether that response addresses your comment?
>  
> thanks,
> Simon
>  
> Dr Simon Miles
> Senior Lecturer, Department of Informatics
> Kings College London, WC2R 2LS, UK
> +44 (0)20 7848 1166
>  
> Transparent Provenance Derivation for User Decisions:
> http://eprints.dcs.kcl.ac.uk/1400/
> From: Morris, Chuck E (IS) [chuck.morris@ngc.com]
> Sent: 10 January 2013 18:55
> To: public-prov-comments@w3.org
> Subject: The wasQuotedFrom relationship
> 
> I just looked over the provenance primer.  One thing I noticed is that the wasQuotedFrom relationship is very confusing semantically.  Take the example in the primer where Betty posts a blog entry with a quote from the newspaper article.  The provenance is expressed as (ex:blogEntry prov:wasQuotedFrom ex:article .) But that seems to imply that the blog entry was quoted by the newspaper article instead of the other way around.  I suggest that a better name for the relationship would be prov:hadQuotationFrom.
>  
> Chuck Morris
>

Received on Friday, 8 February 2013 15:11:53 UTC