Re: testable properties of repositories that could be used to rate them

Hi.
> On 14 Sep 2014, at 22:06, Stuart Yeates <stuart.yeates@vuw.ac.nz> wrote:
> 
> The initial aim of this was to counter an apparently arbitrary repository ranking algorithm (which I won't deign link to) with a set of web standards that we (repository developers and maintainers) can collectively work towards, with an emphasis on breadth of different standards that could be applied.
Sure.
> 
> --
> 
> I've greyed out the 'everything' requirement, since I'm not sure that 'everything' is script-testable.
Yes, I was puzzling over that (how it could be made scriptable).
Certainly quite a lot of the other things in the list make assumptions about repository identifiers being available - otherwise how can you get started, or ask if dc:title is used, for example?
So how do you find the repository identifiers in a scriptable manner?
Let’s assume that there is no OAI-PMH support, for example.
So for this, maybe I could move it to after number 3 (where we know there is RDF) and then I could list the predicates that must have URIs (rather than strings)?
> 
> I've grey'ed out the content negotiation requirements since I'm not aware that any repositories or prototypes that try and do this (I'm happy to be corrected).
That actually seems rather a strange statement - if you had said that there was no interest in it, then that would be fine.
But surely your rating should list anything useful that a repository might offer?
Is there nothing else in your list that is not currently supported? Is RDFa supported anywhere?
But fear not, there are many examples in the wild!
The standard ePrints 3 software supports content negotiation - e.g. http://oro.open.ac.uk/id/eprint/40795
I see it does rdf+xml and text/n3 - I haven’t tried any others.
> 
> I've found a better URL for the RDFa requirement.
Nice.

> 
> cheers
Cheers
Hugh
> stuart
> 
> 
> 
> On 13/09/14 22:58, Hugh Glaser wrote:
>> The messages below should make sense.
>> Stuart is trying to make a doc for rating repositories.
>> 
>> I’ve added some stuff about Linked Data:
>> From http://www.w3.org/DesignIssues/LinkedData.html (Linked Data Principles)
>>  Everything has a URI - publications, documents, people, organisations, categories, ...
>>  These URIs are HTTP or HTTPS
>>  When RDF is requested, the URIs return RDF metadata
>>   RDF/XML supported
>>   N3 supported
>>   Turtle supported
>>   JSON-LD supported
>>  There are URIs that are not from this repository
>>  There are URIs from other repositories
>>  There is a SPARQL endpoint
>>  RDFa is embedded in the HTML
>> 
>> Is there somewhere I could have taken this from that would be suitable?
>> Anyone care to contribute?
>> It seems like it is a really useful thing to have (modulo a bit of specialisation for any particular domain).
>> (I didn’t want to go over the top on formats, by the way.)
>> Cheers
>> 
>>> Begin forwarded message:
>>> 
>>> From: Stuart Yeates <stuart.yeates@vuw.ac.nz>
>>> Subject: RE: testable properties of repositories that could be used to rate them
>>> Date: 13 September 2014 10:31:36 BST
>>> To: Hugh Glaser <hg@ecs.soton.ac.uk>
>>> Cc: "JISC-REPOSITORIES@JISCMAIL.AC.UK" <JISC-REPOSITORIES@JISCMAIL.AC.UK>
>>> 
>>>> I notice there is nothing about Linked Data and Semantic Web - would it be sensible to have something on this?
>>> 
>>> If there's something that's recommended by some standard / recommendation and is script-testable, you're welcome to add it.
>>> 
>>>> So for example does it provide RDF at all?
>>> 
>>> It has a question based on  http://validator.w3.org/feed/  which validates RSS, which in turn is either RDF (v1.0) or can trivially be converted to it (v2.0/atom). I've added a note that this is RSS.
>>> 
>>> cheers
>>> stuart
>> 
>>> Begin forwarded message:
>>> 
>>> From: Hugh Glaser <hg@ECS.SOTON.AC.UK>
>>> Subject: Re: testable properties of repositories that could be used to rate them
>>> Date: 12 September 2014 14:05:34 BST
>>> To: <JISC-REPOSITORIES@JISCMAIL.AC.UK>
>>> Reply-To: Hugh Glaser <hg@ECS.SOTON.AC.UK>
>>> 
>>> Very interesting (and impressive!)
>>> 
>>> I notice there is nothing about Linked Data and Semantic Web - would it be sensible to have something on this?
>>> Well, actually there is Semantic Web:- right up at the start there is a Cool URI reference, which is the the W3C "Cool URIs for the Semantic Web” note!
>>> 
>>> Perhaps there should be a section on this - maybe starting with with whether it is 5* Linked Data.
>>> http://en.wikipedia.org/wiki/Linked_data
>>> http://www.w3.org/DesignIssues/LinkedData.html
>>> 
>>> But it probably useful to unpick some of this in a less structured way.
>>> So for example does it provide RDF at all?
>>> Formats? RDF, N3, JSON-LD…
>>> 
>>> Best
>>> Hugh
>>>> On 12 Sep 2014, at 03:29, Stuart Yeates <stuart.yeates@VUW.AC.NZ> wrote:
>>>> 
>>>> A couple of us have drawn up a bit of a list of script-testable properties of repositories that could be used to rate them. We’re tried to both avoid arbitrary judgements and the implication that every repository should meet every item:
>>>> 
>>>> https://docs.google.com/document/d/1sEDqPS2bfAcbunpjNzHwB56f5CY1SxJunSBLFtom3IM/edit
>>>> 
>>>> cheers
>>>> stuart
>>> 

-- 
Hugh Glaser
   20 Portchester Rise
   Eastleigh
   SO50 4QS
Mobile: +44 75 9533 4155, Home: +44 23 8061 5652

Received on Sunday, 14 September 2014 21:25:42 UTC