Re: Demo SPARQL notes

Well, this is a conundrum. As you know, if anything is inconsistent  
in DL, then everything logically follows.

What do you expect the behavior of the reasoner and the query engine  
to be in such a case? Some sort of
bounds around inconsistencies? If so, how to set the bounds.

It's been my experience that the cases of the data being wrong, or at  
best incomplete, in bio databases, vastly outnumber of the cases of  
the evidence actually
being contradictory. I'd be curious to have a look at papers  
supporting contradictory answers that you identified. Is is possible  
that upon rereading of the paper we will be able to find something in  
the biological context that will distinguish them.

A reasonable strategy might be to have 2 layers of representation.  
Assert the biology as best you can. Perhaps even overstate (we will  
do some of that for the demo). Have a reasoner identify the  
contradictions and then, based on a curator's best judgement, remove  
those findings that seem contradictory. Represent those separately,  
as statements, in a database of controversial  findings. As a  
scientist, understand that you will need to query both the consistent  
facts and the inconsistent facts if you want to know the whole story.

This seems a better option than what I usually see - silently leaving  
in a portion of actual encoding errors with a smidgen of controversy,  
and not marking any of it.

-Alan

On Apr 17, 2007, at 10:42 PM, samwald@gmx.at wrote:

>> In this case you need to make a choice about whether you want to
>> say something that we called in [1] the 'statement level' or the
>> 'domain level'. If at the domain level you need to put your neck on
>> the line and say which experiment is right.
>
> In many cases we cannot do that.
>
>> If at the statement level
>> you need to remodel so that you are clearly communicating that you
>> are representing author statements.
>
> In many cases we don't want to do that.
> Representing everything as author statements (i.e. statements about  
> statements) would require wholly different ontologies than those  
> most of us are developing. Consistency checking and inferencing  
> over biological relations would probably not be possible with such  
> an ontology. There would be no need to use a fancy OWL ontology and  
> reasoner in this scenario at all.
>
> -Matthias
>
>
>
>
> .
>
> -- 
> "Feel free" - 10 GB Mailbox, 100 FreeSMS/Monat ...
> Jetzt GMX TopMail testen: http://www.gmx.net/de/go/topmail

Received on Wednesday, 18 April 2007 02:58:46 UTC