Re: [Myexperiment-discuss] myExperiment RDF API and ID numbers

[David R Newman]

Hi All,

Sorry didn't quite finish my email there.  To continue:

However the REST API also allow you to recover any resource explicitly:

http://www.myexperiment.org/rest-access.xml?resource=http://www.myexperiment.org/workflows/20

myExperiment's RDF API allows you to recover the html page resource URI
as mebase:human-start-page (an anyURI-typed datatype property), that
could then be used to make this REST API call.  There has been a fair
amount of discussion of what is the correct way to reference the HTML
page from the RDF.  

Egon suggested that maybe the property foaf:homepage could be used to
reference the HTML page, which makes sense as this is a commonly known
property and sort of represents the relationship we want to describe.  

Linked Data people suggest a number of different ways of handling
different representations of the the same thing and further to that
people in the Open Repositories community think about this differently
again.

It has been something that has been troubling me for a time so I felt it
might be worth opening it up to the wider community to discuss the best
way of approaching the problem of being able to identify workflows
within myExperiment and how that translates across the HTML, REST and
RDF representations that myExperiment provides.

Regards

David Newman


On Fri, 2009-10-16 at 13:25 +0100, David R Newman wrote:
> Hi All,
> 
> Egon and I have been having a discussion of certain aspects of the RDF
> API after he made the following request:
> 
> "If I look up the RDF for a workflow (e.g.
> http://rdf.myexperiment.org/Workflow/889), I do not have access to the
> Workflow number as an integer field... while I can extract this from the
> above URI, I'd very much prefer an explicit field listing the number..."
> 
> This raised a few interesting question.  First the recognition of the
> emergent behaviour for users to refer to workflows by the ID number
> allocated to a workflow, ultimately through MySQL's auto increment ID
> field.  Egon cites Duncan's twitter post about car number plates:
> 
> "RV07NZT is in many contexts a much better id for my car than
> http://x.y.z/reg/uk/RV07NZT";
> 
> Clearly when you are in the context of talking about myExperiment
> workflows, often all people need to know to understand which one you are
> refering to is the the ID number:
> 
> "889, that's the Bioeclipse script that use DBPedia RDF data isn't it?"
> 
> Also the REST API uses ID as an explicit parameter:
> 
> http://www.myexperiment.org/workflow.xml?id=889
> 
> Therefore if you want to recover data using the SPARQL endpoint and then
> perform operations on it using the REST API having the ID a separate
> field would save any data munging.  However the REST API also allow you
> to recover any resource explicitly:
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> -----Original Message-----
> From: Egon Willighagen <address@hidden>
> Reply-To: Egon Willighagen <address@hidden>
> Date: Thu, 15 Oct 2009 18:12:31 +0200
> To: myExperiment <address@hidden>
> Cc: address@hidden, Duncan Hull
> <address@hidden>
> Subject: Re: (Case 1106) myExperiment feedback from Egon Willighagen
> (address@hidden)
> 
> >On Thu, Oct 15, 2009 at 6:01 PM, myExperiment <address@hidden>
> wrote:
> >> The following query will return to columns, the workflow URI and the
> URL for downloading the workflow.  From what you are doing I assume you
> are using the REST API to request the download of a workflow.  Using the
> "content-url" value returned by this query will save you having to go
> via the REST API.
> >
> >Yes, I know. But I still need to get the numbers. As I do not want to
> >download the workflows directly...
> >
> >> I can see how in general if you want to go an use the REST API it is
> useful to have the ID value to throw at it.  However, if this solves
> your problem in the meantime I will probably take some time to think how
> the REST and RDF APIs should best interact.
> >
> >An alternative solution for me would be to return the full workflow
> >URI... it's like that RT from Duncan Hull (cc-ed) today:
> >
> >"rt @michaelhkay  "RV07NZT is in many contexts a much better id for
> >my car than http://x.y.z/reg/uk/RV07NZT"; #linkeddata
> >http://bit.ly/JSJKR";
> >
> >I talk about workflow "889", and not workflow
> >"http://rdf.myexperiment.org/Workflow/889"; ...
> >
> >The MyExperiment website indeed does not seem to use it as such,
> >though the URLs of the www frontend tend to encourage it a bit, with
> >the REST like structure...
> >
> >I am patient... it's just a feature request :)
> >
> >BTW, while at it... what about a triple
> >
> >http://rdf.myexperiment.org/Workflow/889 foaf:homepage
> >http://www.myexperiment.org/workflows/889
> >
> >(as another thought) ?
> >
> >Oh, another thingy... are there plans to use the ontologies for
> >embedding the RDF as RDFa in the HTML directly? Just wondering... :)
> >
> >Egon
> >
> >-- 
> >Post-doc @ Uppsala University
> >Homepage: http://egonw.github.com/
> >Blog: http://chem-bla-ics.blogspot.com/
> >PubList: http://www.citeulike.org/user/egonw/tag/papers
> >  
> 
> 
> 
> _______________________________________________
> Myexperiment-discuss mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/myexperiment-discuss