ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] URIs [was: Truth]

To: "steve.ray@xxxxxxxxxx" <steve.ray@xxxxxxxxxx>
Cc: "'[ontolog-forum]'" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Edward Barkmeyer <edward.barkmeyer@xxxxxxxx>
Date: Thu, 12 Jul 2012 18:47:14 -0400
Message-id: <4FFF53F2.2050608@xxxxxxxx>
Steve Ray wrote:
> I'm willing to live with multiple authorities, as long as they provide
> stable URIs that I can point to. Then at least people can find out what
> definitions I'm using in my models.
>       (01)

But they can do that if you define your own terms.
The point of using a term from some established authoritative source is 
threefold:
 (a) your human audience may be familiar with the referenced term and 
thus better understand your ontology, and
 (b) integrating your work with that of others will be easier because 
you use common terms for common concepts
 (c) the authoritative definitions will use terms defined in 
authoritative sources    (02)

The problem with having 6 xxx#integer terms from authoritative sources, 
plus yyy#UnboundedNaturalNumber and 3 other such terms, is that it 
greatly reduces the probability of both (a) and (b) and creates a lousy 
basis for (c).  The assumption that xxx#time and yyy#time are synonyms 
is unwarranted, to say the least.  So can you be sure that xxx#integer 
and yyy#integer are the same concept?  No, you have to look at the 
definitions, and even then you may not be sure.  And software isn't 
going to do that.  Unless someone supplies the 'sameAs' axiom, for the 
software, xxx#integer and yyy#integer are exactly as closely related as 
xxx#cheese and yyy#chalk.    (03)

 From an ontological viewpoint, there is nothing wrong with diverse 
nomenclature if it is accompanied by formal term definitions.  But then, 
what vocabulary are those definition turtles resting on?  For software 
purposes, the successful solutions depend rather on accepted 
"equivalence tables".  But for n vocabularies, you need at least n-1 
tables.  The problem is identifying a handful of authoritative sources 
for the general terms of an industry domain.  Matthew's idea that an 
organization should be the authority for terms that it should own is, of 
course, undeniable.  But who takes ownership of the general terms of the 
industry?     (04)

Steve's question about 'data types' is a good example of 'general terms 
of an industry'.  Or maybe it isn't.  That is, one would think that 
owning the common terms of an industry would fall to a professional 
society or a respected industry association.  Perhaps the IT industry is 
simply not yet mature enough to have an authoritative version of either.    (05)

-Ed    (06)



> - Steve
>
> Steven R. Ray, Ph.D.
> Distinguished Research Fellow
> Carnegie Mellon University
> NASA Research Park
> Building 23 (MS 23-11)
> P.O. Box 1
> Moffett Field, CA 94305-0001
> Email:    steve.ray@xxxxxxxxxx
> Phone: (650) 587-3780
> Cell:      (202) 316-6481
> Skype: steverayconsulting
>
>
>
> -----Original Message-----
> From: Ed Barkmeyer [mailto:edbark@xxxxxxxx] 
> Sent: Thursday, July 12, 2012 11:57 AM
> To: steve.ray@xxxxxxxxxx; [ontolog-forum]
> Subject: Re: [ontolog-forum] URIs [was: Truth]
>
>
>
> Steve Ray wrote:
>   
>> Matthew,
>>      I completely agree with you, and would be interested to hear from 
>> this group what they believe are authoritative URIs for common data 
>> types such as:
>>
>> Float (or Real)
>> Integer
>> Boolean
>> String
>>   
>>     
>
> I have never met a datatype term that the information technology community
> couldn't argue about indefinitely. 
>
> The problem is:  Who is the registration authority for things like this?
> And who is willing to concede that authority?  The W3C attitude is that
> these have been defined in XML Schema part 2: Datatypes.  The JCP attitude
> is that, for matters related to Java, the Java standards are the authority
> for such datatypes.  Fortunately, tthe JCP lot were able to convince W3C to
> use their types in XML Schema, but then, so were the SQL folk, and most
> modelers prefer xsd:integer to xsd:long as the name for Integer.  And the
> ISO 11179 "metadata registry" folk think the registered names are ISO URIs
> for 'technical elements' of ISO 11404.  
> What we have is multiple communities who claim the authority, based on
> having published standards for certain implementation communities.
>
> So it seems to me that we will get multiple URIs for these datatype
> concepts, each of which has a definition in a standard written by whichever
> community owns that URI.  And some poor souls will have to determine whether
> the extensions of the defined concepts are really the same, or were intended
> to be, or not.  And then they will have to get everyone to agree that their
> repository of equivalences or subtypes or whatever is authoritative.  In
> short, nomenclature standards really don't work over large communities until
> the supporting technology stops changing.  And if we have to agree on both a
> term and its definition, it will be even harder.
>
> In a similar way, Matthew is certainly right about Ferraris, but who is the
> internationally accepted authority for motor vehicle terminology?  
> ODETTE?  Is it a "hood" or a "bonnet" or some other bizarre term no one
> actually uses?  Does everyone agree on the meaning of "power train"?
>
>   
>> And then, thoughts on the same for measured quantities such as:
>>
>> AbsoluteDateTime
>> Duration
>> (I'll stop here for now)
>>   
>>     
>
> Yeah.  You don't want to put the objective any further than Arhhem.  But you
> are already a bridge too far. 
>
> The unarguable authority for what Steve calls 'duration' is BIPM, reference
> source for all the basic measurement concepts.  Their publications, e.g.,
> the International System of (measurement) Units (SI), refer to that kind of
> quantity as 'time'.  The generally agreed upon authority for date and time
> expressions is ISO 8601, but I don't think it uses the term "Absolute Date
> Time" at all.  Unlike computational datatypes, which change with technology,
> however, these latter concepts have established international authorities. 
>
> There is still the issue of whether everyone agrees on the nomenclature for
> those concepts, as Steve's choice of terms indicates. This is not really a
> problem in RDF.  The term is not 'time' or 'duration', it is
> "http://www.bipm.org/publications/SI#time";, or something very similar.  
> The problem arises from humans trying to interpret only the last 4
> characters of that URI, without visiting the site and examining the
> definition.  As in most such things, we have met the enemy and they is us.
>
> -Ed
>
>       (07)


-- 
Edward J. Barkmeyer                        Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263                Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263                Cel: +1 240-672-5800    (08)

"The opinions expressed above do not reflect consensus of NIST, 
 and have not been reviewed by any Government authority."    (09)


_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (010)

<Prev in Thread] Current Thread [Next in Thread>