Re: [wbs] response to 'W3C PROV Vocabulary Usage Survey'

On 03/05/2013 03:00 PM, Stian Soiland-Reyes wrote:
> Hi, I am one of the members of the W3C Provenance WG.

Hi Stian,

> Thanks for your registrations. Our survey system is a bit 'lossy' as
> it overwrites based on email address at
> https://www.w3.org/2002/09/wbs/46974/prov-vocabulary-usage-survey/results
> - I'll see if we can find a way to list them all.
>
>
> The rest of this email is just a personal comment and does not
> (necessarily) reflect the view of the WG.
>
>
> I think your datasets at  http://fao.270a.info/ and
> http://oecd.270a.info/ show beautiful use of PROV and VoID, has a very
> simple and easy way to present the metadata as HTML, and are overall
> very understandable and well structured.
>
>
> Is the group and/or myself OK to promote these in say a blog post as
> an example of best practice, or is the site not yet public? I find no
> links to them with Google (although they have been indexed).
>

The sites are public. There is also http://bfs.270a.info/ (I'm not sure 
if that made it through your survey). I will announce another major 
dataset this week.

>
> A few questions or bugs, which I hope you don't mind me raising:
>

On the contrary, I need more of these! Thanks.

> If I retrieve http://fao.270a.info/provenance/activity/20130214190859
> with Accept: text/turtle, I get redirected to
> http://fao.270a.info/provenance/activity/20130214190859.rdf rather
> than http://fao.270a.info/provenance/activity/20130214190859.turtle
>
>
>
> Looking at http://fao.270a.info/provenance/activity/20130214190859.html
> it includes alternate-links to json, rdf and turtle (yay!) - but the
> turtle link has:
>
>   <link rel="alternate" type="text/plain"
> href="http://fao.270a.info/provenance/activity/20130214190859.turtle"
> title="Turtle version of this document"/>
>
> Here I assumed type should be text/turtle. The dc:format of the ...rdf
> also shows text/plain instead of text/turtle.
>

Indeed. The "feature" probably originates somewhere in Paget. The 
platform is on the edge of being archaic and I haven't looked at it 
closer in awhile. I'm aware of some of these annoyances, so, will attend 
soon enough. Might have to drop Linked Data Pages/Paget/Moriarty/ARC2 
altogether.

Sorry for the inconvenience. You could use a tool like rapper (if you 
are not already) and look the other way for the time being? :)

>
> In the prov:Activity
> http://fao.270a.info/provenance/activity/20130214190859, you list:
>
>      prov:used <http://www.fao.org/figis/...9>,
>           <https://launchpad.net/ubuntu/+source/wget> .
>
> I like how you add prov:wasDerivedFrom between the generated resources
> and the 'source' entity - this helps distinguish the source entity
> from the wget tool - like in
> http://oecd.270a.info/code/1.0/CL_TIGS_IO_LOCATION - without needing
> to qualify it with custom prov:Role's.

I would take my actions not as someone that has a clever way of working 
with PROV-O but rather someone that understood enough to put it to use. 
Or at least I thought I understood enough. A part of the property 
selection was due to things being within reach without hurting my head. 
I'll take a closer look at prov:Role next time. Thanks.

>
> I initially wondered about the fact that the wget software is also
> 'used', as there are multiple approaches (ex:WGet specialization of
> prov:Activity; or a prov:Agent  (the server?) executing prov:hadPlan
> <http://wget>; or perhaps a specialization with prov:wasAssociatedWith
> [ prov:specializationOf <http://wget>, prov:actedOnBehalfOf :you ] . )
>
>
> Is the wget 'used' here because the activity is considered to be
> executing a script (or manually executed command) that combines wget
> and the FAO query..?   If an actual script was used, perhaps it could
> be referred to with prov:hadPlan in a qualified association:
>
> <http://example.com/activity> prov:qualifiedAssociation [
>      prov:agent <http://csarven.ca/#i> ;
>      prov:hadPlan <http://github.com/some/script.pl> ;
> ]
>
>
> Or just embedded commands:
>
> :a prov:qualifiedAssociation [
>      prov:agent <http://csarven.ca/#i> ;
>      prov:hadPlan [ a ex:Command ;
>          prov:value "wget http://asdjfhjsadhfjsadfh"
>      ] ;
> ]

This is a lot closer to what I intended. That is, wget is just being 
employed by a Bash script to retrieve the datasets. For the 
transformation activity, I will also include the saxonb-xslt command.


I'm not sure if it shows right now or not (I had some bugs in the past 
week or so), but, one of the cool things that I like about PROV-O is 
prov:wasInformedBy. It lets me say this:

Transformation activity:
http://oecd.270a.info/provenance/activity/20130214234147

prov:wasInformedBy

Retrieval activity:
http://oecd.270a.info/provenance/activity/20130206112045

>
> Although I can understand if you think that would be too much detail! :-)
>

Not at all. Thank you for the feedback. I'd love to give PROV-O another. 
Right now I'm trying not to break things along the way to publishing 
these new datasets :)

-Sarven

>
> On Sun, Feb 17, 2013 at 5:57 PM, WBS Mailer on behalf of
> info@csarven.ca <webmaster@w3.org> wrote:
>> The following answers have been successfully submitted to 'W3C PROV
>> Vocabulary Usage Survey' (Provenance Working Group) for Sarven Capadisli.
>>
>>
>> ---------------------------------
>> Vocabulary Usage Information
>> ----
>> Please provide the name and url of the dataset, website or other set of
>> content that uses PROV  to describe provenance. We also encourage you to
>> fill out this form if your site uses an extension to PROV.
>> Name: FAO Linked DataURL: http://fao.270a.info/Description: FAO (Food and
>> Agriculture Organization of the United Nations) data published using the
>> Linked Data design principles
>>
>>
>> ---------------------------------
>> Contact Information
>> ----
>>
>> Name: Sarven CapadisliEmail: info@csarven.ca
>>
>> ---------------------------------
>> PROV Encodings Supported
>> ----
>> Choose all that apply
>>
>>   * [x] PROV-O
>>   * [ ] PROV-N
>>   * [ ] PROV-XML
>> Please list any additional supported encodings (e.g. PROV-JSON, PROV-CSV,
>> etc.) in the free-text area below:
>>
>>
>> ---------------------------------
>> Feature Coverage
>> ----
>> Indicate covered features by selecting one of the following below:
>>   * 1) I Don't Know
>>   * 2) Used
>>   * 3) Will be used in the future
>>
>>   * Entity: [ 2 ++ ]
>>   * Activity: [ 2 ++ ]
>>   * Agent: [ 2 ++ ]
>>   * Generation: [ 2 ++ ]
>>   * Usage: [ 2 ++ ]
>>   * Communication: [ 2 ++ ]
>>   * Derivation: [ 2 ++ ]
>>   * Attribution: [ 2 ++ ]
>>   * Association: [ No opinion ]
>>   * Delegation: [ 2 ++ ]
>>   * Start: [ 2 ++ ]
>>   * End: [ 2 ++ ]
>>   * Invalidation: [ No opinion ]
>>   * Revision: [ No opinion ]
>>   * Quotation: [ No opinion ]
>>   * PrimarySource: [ No opinion ]
>>   * Person: [ No opinion ]
>>   * Organization: [ No opinion ]
>>   * SoftwareAgent: [ No opinion ]
>>   * Plan: [ No opinion ]
>>   * Influence: [ No opinion ]
>>   * Bundle: [ No opinion ]
>>   * Specialization: [ No opinion ]
>>   * Alternate: [ No opinion ]
>>   * Collection: [ No opinion ]
>>   * EmptyCollection: [ No opinion ]
>>   * Membership: [ No opinion ]
>>   * Identifier: [ No opinion ]
>>   * Attributes: [ No opinion ]
>>   * Label: [ No opinion ]
>>   * Location
>>
>> : [ No opinion ]
>>   * Role: [ No opinion ]
>>   * Type: [ No opinion ]
>>   * Value: [ No opinion ]
>> Rationale:
>>
>>
>> ---------------------------------
>> Provenance Exchange
>> ----
>> Is this vocabulary extension generated or consumed by an implementation,
>> which one(s)?
>>
>>
>> These answers were last modified on 17 February 2013 at 17:56:20 U.T.C.
>> by Sarven Capadisli
>>
>> Answers to this questionnaire can be set and changed at
>> https://www.w3.org/2002/09/wbs/46974/prov-vocabulary-usage-survey/ until
>> 2013-03-30.
>>
>>   Regards,
>>
>>   The Automatic WBS Mailer
>>
>>
>
>
>