Re: Dataset URIs and metadata.

Welcome.
Hopefully not muddying the water…

We use (like many, I think) a voiD URI to identify our datasets, for example 
http://oai.rkbexplorer.com/id/void
This does the standard content negotiation, try:
curl -i -L -H "Accept: application/rdf+xml" http://oai.rkbexplorer.com/id/void
(The -i will show you what the content negotiation does.)

So I think that does the 3 things you asked about.
We basically treat the dataset as any other resource.
We don't do RDFa for it (although that would be fine), because we don't do RDFa in general.

We have gathered some void documents into a store at http://void.rkbexplorer.com/ if you want to explore some, and even submit your own.
You can look up the at the above dataset at http://void.rkbexplorer.com/browse/?uri=http://oai.rkbexplorer.com/id/void

Keith Alexander also has <a href="http://kwijibo.talis.com/voiD/">another voiD store</a>.

Best
Hugh


On 22 Jul 2011, at 11:16, Frans Knibbe wrote:

> Hello Michael,
> 
> On 2011-07-22 10:59, Michael Hausenblas wrote:
>> 
>> Frans,
>> 
>>> I had a quick look. But I could not find it. I had a closer look now and I see the URI probably is http://dbpedia.org/void/Dataset. I have tried it. Redirection to either HTML or RDF seems to be in place. HTML request lead to http://dbpedia.org/void/page/Dataset, which shows a table of VoID properties. The RDF redirection (to http://dbpedia.org/void/data/Dataset.rdf) does not seem to work, I get an error.
>> 
>> Sorry, you lost me here. Did you curl it or how did you get your findings?
> 
> Yes, I tried curl. I have just tried it again:
> 
> curl -H "Accept: application/rdf+xml" http://dbpedia.org/void/data/Dataset.rdf
> 
> The response seems to be different from yesterday. Yesterday I immediately got an error message. I am afraid I don't have the exact message any more. The response I get now is different, a transaction time out, which I get after waiting a bit.
> 
> Probably this is just a temporary situation and besides the point too.
>> 
>>> Apart from the error, somehow this is not what I expected. I assumed that the dataset URI is the URI of a dataset. It is the key to all other data. If you want something from a dataset, you only need to know this URI. So why is the dataset URI hard to find? Why isn't it used when references are made to DBpedia? Why isn't it the same as the base URI (http://dbpedia.org)?.
>> 
>> <http://lod-cloud.net/dbpedia> a void:Dataset;
>>                               foaf:homepage <http://dbpedia.org/>;
>> 
>> Says everything, or?
> Well, let me see...
> 
> First of all, please know that I am new to Linked Data and RDF, so there is a chance I don't fully understand everything.
> 
> I think what it says is that there is a thing identified by the URI http://lod-cloud.net/dbpedia, that that thing is a dataset and that its home page is http://dbpedia.org/.
> 
> So, does this mean that the URI of the dataset (DBPedia) is http://lod-cloud.net/dbpedia??
> 
> Sorry if I seem to be stupid, it is not my intention.
> 
>> 
>>> Probably VoID metadata/dataset URIs will be easier to discover once the /.well-known/void trick (described in paragraph 7.2 of the W3C VoID document) is widely adopted.
>> 
>> Agreed. But it's not a 'trick'. It's called a standard.
> Sorry about that!
> 
> Regards,
> Frans
>> 
>> Cheers,
>>    Michael
>> -- 
>> Dr. Michael Hausenblas, Research Fellow
>> LiDRC - Linked Data Research Centre
>> DERI - Digital Enterprise Research Institute
>> NUIG - National University of Ireland, Galway
>> Ireland, Europe
>> Tel. +353 91 495730
>> http://linkeddata.deri.ie/
>> http://sw-app.org/about.html
>> 
>> On 22 Jul 2011, at 09:42, Frans Knibbe wrote:
>> 
>>> On 2011-07-21 16:27, Michael Hausenblas wrote:
>>>> 
>>>> 
>>>>> But is this really common practice nowadays? Take DBpedia for example. What is the URI of the DBpedia dataset? Is it http://dbpedia.org? That does not seem to resolve to a set of metadata.
>>>> 
>>>> Did you have a look at the URI I gave you? I mean http://lod-cloud.net/void.ttl
>>> I had a quick look. But I could not find it. I had a closer look now and I see the URI probably is http://dbpedia.org/void/Dataset. I have tried it. Redirection to either HTML or RDF seems to be in place. HTML request lead to http://dbpedia.org/void/page/Dataset, which shows a table of VoID properties. The RDF redirection (to http://dbpedia.org/void/data/Dataset.rdf) does not seem to work, I get an error.
>>> 
>>> Apart from the error, somehow this is not what I expected. I assumed that the dataset URI is the URI of a dataset. It is the key to all other data. If you want something from a dataset, you only need to know this URI. So why is the dataset URI hard to find? Why isn't it used when references are made to DBpedia? Why isn't it the same as the base URI (http://dbpedia.org)?.
>>> 
>>> Probably VoID metadata/dataset URIs will be easier to discover once the /.well-known/void trick (described in paragraph 7.2 of the W3C VoID document) is widely adopted.
>>> 
>>>> 
>>>> BTW, some 30% [1] of the LOD cloud datasets are using VoID ...
>>>> 
>>>>> Is there a general way of obtaining datasets URIs?
>>>> 
>>>> Not to my knowledge. We're working on it in LATC [2] - Keith?
>>>> 
>>>> Cheers,
>>>>    Michael
>>>> 
>>>> [1] http://www4.wiwiss.fu-berlin.de/lodcloud/state/#data-set-level-metadata
>>>> [2] http://latc-project.eu/
>>>> 
>>>> -- 
>>>> Dr. Michael Hausenblas, Research Fellow
>>>> LiDRC - Linked Data Research Centre
>>>> DERI - Digital Enterprise Research Institute
>>>> NUIG - National University of Ireland, Galway
>>>> Ireland, Europe
>>>> Tel. +353 91 495730
>>>> http://linkeddata.deri.ie/
>>>> http://sw-app.org/about.html
>>>> 
>>>> On 21 Jul 2011, at 15:19, Frans Knibbe wrote:
>>>> 
>>>>> 
>>>>> Thanks for the replies. It seems that there is agreement that a dataset should have a URI and that dereferencing that URI should return metadata about the dataset. That is good to know.
>>>>> 
>>>>> But is this really common practice nowadays? Take DBpedia for example. What is the URI of the DBpedia dataset? Is it http://dbpedia.org? That does not seem to resolve to a set of metadata.
>>>>> 
>>>>> Is there a general way of obtaining datasets URIs?
>>>>> 
>>>>> I can imagine an RDF dataset comprising all known dataset URIs. And of course that dataset will have a URI itself. Does such a dataset exist at the moment?
>>>>> 
>>>>> Regards,
>>>>> Frans
>>>>> 
>>>>> 
>>>>> On 2011-07-21 12:35, Frans Knibbe wrote:
>>>>>> 
>>>>>> Hello,
>>>>>> 
>>>>>> I have just placed a Linked Data dataset online and now I am struggling with finding the best way to publish the metadata of the dataset. I wonder if there are best practices for referencing a dataset and its metadata, and for linking the two.
>>>>>> 
>>>>>> I did find out that using the Vocabulary of Interlinked Data (VoID) is a good way to publish the metadata of a dataset. But I still need some guidance. I have come up with three questions:
>>>>>> 
>>>>>> 1) Is it common practice/recommendable to regard a dataset a resource? If it is, then all datasets should have a URI, right?
>>>>>> 2) If having a dataset URI is a good thing, what should be behind the URI? Should dereferencing the URI lead to the dataset metadata (a VoID file for example)?
>>>>>> 3) If dereferencing a dataset  URI leads to the dataset metadata, should there be separate HTML and RDF versions of the metadata? Or is it better to have a HTML page with embedded (RDFa) RDF data?
>>>>>> 
>>>>>> Thanks in advance for your help,
>>>>>> Frans
>>>>> 
>>>> 
>>>> 
>>> 
>>> 
>> 
>> 
> 
> 
> 

-- 
Hugh Glaser,  
              Intelligence, Agents, Multimedia
              School of Electronics and Computer Science,
              University of Southampton,
              Southampton SO17 1BJ
Work: +44 23 8059 3670, Fax: +44 23 8059 3045
Mobile: +44 75 9533 4155 , Home: +44 23 8061 5652
http://www.ecs.soton.ac.uk/~hg/

Received on Friday, 22 July 2011 12:10:19 UTC