W3C home > Mailing lists > Public > public-lod@w3.org > July 2009

Re: Merging Databases

From: Oktie Hassanzadeh <oktie@cs.toronto.edu>
Date: Mon, 20 Jul 2009 17:17:40 -0400
Message-ID: <cf9831bb0907201417i50994c0ag221389290622e190@mail.gmail.com>
To: Hugh Glaser <hg@ecs.soton.ac.uk>
Cc: Kingsley Idehen <kidehen@openlinksw.com>, Alan Ruttenberg <alanruttenberg@gmail.com>, "public-lod@w3.org" <public-lod@w3.org>, Anja Jentzsch <anja@anjeve.de>
On Mon, Jul 20, 2009 at 4:46 PM, Hugh Glaser<hg@ecs.soton.ac.uk> wrote:
> On 20/07/2009 21:29, "Kingsley Idehen" <kidehen@openlinksw.com> wrote:
>
>> Oktie Hassanzadeh wrote:
>>>
>>> Thanks for pointing this out. I agree that for this case, sameAs is an
>>> incorrect type for the links to other drug sources since intervention
>>> is not necessarily a drug, and has a description associated with it.
>>> We should treat drug as an entity for the sameAs links to make more
>>> sense. I'm working on this and will let you all know once I update the
>>> data source.
>>>
>> Oktie,
>>
>> All of us that are consuming the Linked Open Drug Data (LODD) really
>> need to be notified. I am a little shocked that this anomaly (as
>> identified by Alan) exists re. use of "owl:sameAs" in the current data
>> set release.
>>
>> Kingsley
> Hi Kingsley,
> I can't say I am.
> I don't see why you would have any expectation either way, as I don't know
> of any systematic study of the reliability of these links (other than
> specific studies of dblp).
> I actually am aware of quite a few problems in the LOD Cloud.
> One of the reasons that problems are not detected is that there are hardly
> any applications actually using the data, so it is unlikely problems will be
> found.
> I can actually tell you from personal experience that as soon as you start
> to build an application that uses more than a couple of LD sites, much of
> the published equivalence data has to be rejected.


I agree, and that's one of the main goals in LODD, to identify such
issues. Part of the group's focus is on developing usecases to make
sure we have the right data and the right linkages. We have already
identified a few problems and fixed them, but we have missed this one
(and probably many others) which also shows this part of the data
hasn't been in any of the usecases and applications so far.

And remember that this is still a preview release of the data, we hope
to gradually remove these anolamies, and we certainly need your
feedback.

Thanks,
Oktie


> Sorry to be the bearer of bad tidings - please don't shoot the messenger!
> Best
> Hugh
>
>
Received on Monday, 20 July 2009 21:18:22 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:29:44 UTC