Re: Blank Nodes Re: Toward easier RDF: a proposal

On 11/26/18 8:55 PM, Thomas Passin wrote:
> On 11/26/2018 9:49 PM, David Booth wrote:
>> It is hard for very smart people to see why concepts that are simple 
>> to them are *not* so simple for others who have significantly less 
>> intellectual horsepower. Typically it is not any one concept that 
>> makes   a subject too hard, but the totality of the interaction 
>> between multiple concepts, each with its own exceptions and caveats, 
>> that pushes the user over his/her mental threshold of "too hard".
>
> It's not necessarily the case of not enough horsepower.  Being able to 
> read and digest the content of formal specifications, and being able 
> to put that to use, isn't easy.  It's very abstract, even more 
> abstract than a lot of everyday programming.  A lot of folks just 
> don't do that well.  It's been remarked before in this thread (I think 
> by you?) that there isn't a lot of good tutorial material out there on 
> how to understand and use RDF-related tools and data design.
>
> If you mean to make it easier for the middle ranks of programmers to 
> work with linked data, then those people are going to have to 
> understand quite a lot about semantics - since it's probably rare for 
> two data sets to have exactly the same semantics for things that are 
> apparently the same - and about data cleaning, truth maintenance, and 
> imputation of missing data.  Absent those things, for someone to 
> blithely forge ahead linking data right and left is only going to 
> cause trouble.
>
> So arguably it will be important to make *those* things easier for 
> middle rank workers to understand and deal with.  RDF details, not so 
> much.  We could probably put together a profile that disallows certain 
> RDF usages and encourages certain idioms, and be done with that part 
> of it.  Then the hard work would start.
>
> TomP
>
>

I have to say that the entire argument about the limited /intellectual 
horsepower/ strikes me as odd to a point where it may hinder us as a 
community to admit that there is room for improvement (which the 
discussion here is all about).

To recall just one example, a group of ontology engineers argued last 
year over the relation between owl:class and rfds:class, particularly 
also with respect to (legacy) tool support. As we could not fully agree, 
we asked two people we trusted would certainly be able to come up with a 
clear recommendation. Unfortunately, their answers seemed not to align 
very well with one another but after a few more iterations  we got it 
sorted out: simply using owl:class and rdfs:class together to type each 
class in our ontology would be 'most certainly harmless' [quote]. This 
is what we ended up doing.

I also vividly recall how we decided to describe our ontology in a 
minimal fashion (really just stating that it is a vocabulary, has a 
creator, follows a certain license, and has a preferredNamespacePrefix) 
and ended up with a combination of vann, voaf, foaf, and dcterms. As 
your fellow non-SW 'developer' friends would confirm, this is not normal.

Thanks to everybody for having this lively discussion,

Krzysztof


-- 
Krzysztof Janowicz

Geography Department, University of California, Santa Barbara
4830 Ellison Hall, Santa Barbara, CA 93106-4060

Email: jano@geog.ucsb.edu
Webpage: http://geog.ucsb.edu/~jano/
Semantic Web Journal: http://www.semantic-web-journal.net

Received on Tuesday, 27 November 2018 05:52:00 UTC