Re: A draft outline for the CSV2RDF document

On 21/05/14 11:35, Ivan Herman wrote:
>>> - Although I forgot to add this to [1] (and we were not sure
>>> whether
>>>>> that should go into the metadata spec in the first place) we
>>>>> did say that we can assign, say, an XSLT script for XML, or
>>>>> a SPARQL CONSTRUCT pattern for RDF that would be executed on
>>>>> the result of the RDF generation; such an extra step could
>>>>> take care of Jeremy's example, right?
>>>
>>> It is something that has been suggested but no one has worked
>>> through the details.
>>>
>>> Certainly possible in XSLT, but SPARQL CONSTRUCT isn't as
>>> powerful as XSLT.  Greeg has made suggestion for CSV-LD.  The
>>> XML publishing world commonly has XSLT.  Other communities don't
>>> necessary have the same degreee of conversion pipelines.
> But all communities have something; at the minimum, one can refer
> back to a javascript of python or whatever processing...
>

I don't think data consumers are necessarily programmers.

Template RDF fragments communicate important structure from publisher to 
data consumer.

If we agree (we don't) that the {} substitutions needs to do some text 
munging, then I think we get a long way but I'd like to be able to 
verify that from other people's experiences/requirements.

(also : http://stackoverflow.com/questions/1618038/xslt-equivalent-for-json)


>> See Jeni's
>> http://lists.w3.org/Archives/Public/public-csv-wg/2014May/0063.html
>>
>>want for conditionality and filed level processing.
>>
>> (where do you stand on that msg?)
>
> It makes me scared. "In all the real-life conversions I’ve ever done
> I’ve always ended up needing conditional statements of some sort". Do
> we really want to go there?

I'm waiting for concrete examples from Jeni on (1) because it may 
complicated or it may not be.  For example, applying different 
conversions to the same CSV file, extracting different parts in 
different ways, can get you a long way (probably 3NF).

I thought we'd established (2) a while ago.

>
> For the RDF world, I do not see why plugging in either an http URI
> for a specific SPARQL engine call using CONSTRUCT, or a textual
> literal with SPARQL CONSTRUCT would not work to massage the output.
> After all, the SPIN people have already done things like that...

You can set up complicated pipelines; it's not one just SPARQL 
CONSTRUCT.  Is that acceptable? (I think not - it's consumer side when 
the knowledge (more structure) is from the publisher).

>
> I am wary going down the line of defining the a complex pattern
> language. That is my problem. And Jeni's mail indicates that a simple
> replacement of {...} may not be enough. (Put it another way, even if
> we do use a template language, users will end up using SPARQL...)
>

	Andy

Received on Wednesday, 21 May 2014 11:05:20 UTC