David, (01)
I get the part about ontologies being useful. Adam Cheyer and I
coauthored a paper titled "Just For Me: Topic Maps and Ontologies",
several years back when, on the CALO project, we observed that office
workers were having difficulties relating to terminology in the
ontology supplied to CALO. As has been stated in this forum from time
to time, different names for things, at least, facilitating handling
those different names is of value if ontologies are to be accepted in
the workplace. (02)
I also get that sometimes, a name string is, itself, a subject. In
some brain ontologies, brain parts are frequently named due to some
author, some date, that is, names are based on citations. There, you
are not scoping the string based on some language code but instead on
another subject. (03)
Reading this thread, end to end, I'm not certain that my original
comment has been covered. Then again, I can be dense at times. I'm
sitting in Seoul right now, installing a platform for studying climate
change; perhaps it's jet lag. (04)
Jack (05)
On Fri, May 21, 2010 at 5:00 AM, David Eddy <deddy@xxxxxxxxxxxxx> wrote:
> Jack -
>
> On May 21, 2010, at 4:58 AM, Jack Park wrote:
>
>> I say that as one
>> who gets confused by those who think I'm the Jack Park who writes
>> books about baseball.
>>
>> Jack
>> who writes books about wind energy, and XML topic map, not baseball.
>
> IF this ontological effort is going to be successful, at least two
> conditions MUST be met at minimum...
>
> (1) the ontological complexity MUST be hidden from 99.99999% of the
> end users. If it takes me more than maybe a half dozen keystrokes to
> find the label/name/thingy I need, I'm going to scream to turn the
> stupid ontology thing OFF!!!! because it's getting in the way.
>
> (precisely like a car... I have a 19 year-old new driver & who has
> confessed that he has no idea what a spark plug is... & he plans to
> be a mechanical engineer)
>
> If one must be an ontologist to use ontologies, the effort will never
> get to the stadium parking lot, much less the starting blocks.
>
>
> (2) "things" (test tubes, columns in data bases, etc.) MUST have
> "good names" (fit for purpose)
>
> (Real world examples: M0101 is not so good for a name these days,
> although it was serviceable 40 years ago when Fortran was a dominant
> software language, MSTR-POL-NO is a much better name for the
> insurance concept "policy number" in a COBOL program. "Policy
> number" is an excellent name on a report or in a letter to a
> customer, but totally inadequate in a Fortran or COBOL or Java
> software program. Try putting "Policy_Number" on a report to a
> business person or customer... good luck.)
>
> "Good names" does not mean a requirement for unique names... think
> the 3 Cs... (1) context, (2) container, & (3) contents. Context:
> baseball vs wind energy (someone looking for baseball will instantly
> know he's in the wrong place if wind energy is what he finds),
> container, your name (certainly not a unique string), & contents is
> the topical information of baseball vs wind energy.
>
> In the physical world we're accustomed to having the 3 Cs fly in
> formation... grocery store, milk jug, milk. Multiple parties (dairy
> farmer, milk processor, food inspector, store, etc.) are motivated to
> have these elements work together. In computers, of course, we're
> nowhere close to such congruence. I can easily put my birthdate
> into a field labeled Work Phone Number & the ontologically challenged
> database will be totally happy.
>
>
> If business people, business analysts, programmers, clerks, etc. are
> forced to learn anything about ontologies in order to use same, then
> ontologies will be an utter failure since no one will use them other
> than ontologists. (Again back to the example of the car... one needs
> to know absolutely nothing about the workings of a car to use it... a
> smidge of physics perhaps, but no more than a smidge.)
>
> "Good names" are the labels—good, bad & ugly—that people are
> accustomed to working with. Under the covers it's ok to have some
> unique (within the domain) identifier that reads like gibberish (e.g.
> 123-45-6789... but it's highly unlikely that end users will accept
> such junk as useful labels. Example: from the world of mainframes...
> there is an important IBM software product called CICS that we all
> depend on... that is pronounced KICKS by the British (who maintain
> the product now) and C-I-C-S by Americans. Call it synonyms or
> whatever, it is not possible to tell Americans it's KICKS or Brits
> that its CICS.
>
> ___________________
> David Eddy
> deddy@xxxxxxxxxxxxx
>
> 781-455-0949
>
>
> _________________________________________________________________
> 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
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
> (06)
_________________________________________________________________
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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (07)
|