Re: Subjects as Literals

Hi Paul

Your argument is quite worth considering, but it begs the question. Why do
people produce such bad design showing they still confuse object properties
and datatype properties, and objects with data values? Someone a bit aware
of what RDF and linked data is all about should know that London (the thing,
not the string) is already identified by so many (too many actually) URIs.
See http://sameas.org/html?uri=http%3A%2F%2Fdbpedia.org%2Fresource%2FLONDONand
I guess it's far from exhaustive ...

Whatever the decision about literals as subjects (and BTW not "subjects as
literals", as the thread keeps being named, as if even here, the term
"literal" itself was taboo in subject position ...) we have still a lot of
simple and yes boring teaching and training and good practices showing to
bring data managers to think first about things their data are about rather
than in data structure.
If you have learnt to think in terms of things first, you don't do this kind
of error any more. When you make introductions and training to this stuff,
take any data familiar to the people in the room, say any spreadsheet and
take each field in it asking : is this data value, or does it stands for a
thing? And if yes, what is it? What is its type and properties? And browse
the cloud to figure if there is some URI available for it (and there is
often one) and if not how do you forge one etc. Again and again until they
have groked this central stuff.

And this is not so difficult as it seems. I've done it again and again with
people from very various background. Yes, as Dan put it, we've still to put
the barrow uphill. And thanks to folks such as Pat keeping pushing it,
despite his constant grumbling about people understanding nothing to basics
and pulling downhill, and his threats of throwing all this and go back to
more vital activities such as house-keeping and gardening. Don't believe
him, he'll still be here for a while I'm afraid :)

Cheers

Bernard



2010/7/2 Paul Gearon <gearon@ieee.org>

> On Thu, Jul 1, 2010 at 10:47 PM, Pat Hayes <phayes@ihmc.us> wrote:
> >
> > On Jul 1, 2010, at 10:06 PM, Jeremy Carroll wrote:
> >
> >> On 7/1/2010 7:51 PM, Pat Hayes wrote:
> >>>
> >>> The mistake here is to presume that simple character strings in RDF are
> >>> being used as though they were words. But this is such a basic error
> that I
> >>> doubt if anyone who holds it is going to be able to use RDF
> successfully in
> >>> any case.
> >>>
> >>>
> >>
> >> Pat - one of the key problems with subject as literals, is that this
> basic
> >> error will be compounded in "spades" - particularly by people who have
> been
> >> told by their managers to make their data public.
> >
> > Why will it be worse when they are subjects than it would be (is?) when
> they
> > are used in the third position?
>
> When the data hangs off the graph as a leaf, then it tends to get
> isolated, and the errors are restricted. For instance,
>
> ex:Fred ex:livesIn "London"
>
> This is wrong, but it's kind of isolated. If subjects become literals,
> then we're also likely to see:
>
> "London" ex:partOf "England"
>
> In other words, the poor modeling that was isolated to leaves now has
> the capacity to become a more integral part of the graph.
>
> Sure, none of this should be happening in the first place, but
> unfortunately it does. Consequently, this leaves a lot of people
> cleaning up the mess. Allowing this problem to become a central part
> of the graph makes the mess more likely to spread.
>
> Maybe this is a spurious complaint, since no one should be writing
> data that's this poor in the first place. But since they do, and
> people are constantly having to deal with it, then keeping the damage
> to a minimum doesn't seem like a bad approach.
>
> (Incidentally, I'm arguing this as a devil's advocate. I've wanted to
> see literals as subjects for years)  :-)
>
> Regards,
> Paul Gearon
>
>


-- 
Bernard Vatant
Senior Consultant
Vocabulary & Data Engineering
Tel:       +33 (0) 971 488 459
Mail:     bernard.vatant@mondeca.com
----------------------------------------------------
Mondeca
3, cité Nollez 75018 Paris France
Web:    http://www.mondeca.com
Blog:    http://mondeca.wordpress.com
----------------------------------------------------