Multiple definitions

Ed Shaya eshaya at umd.edu
Fri Feb 8 09:11:37 PST 2008


Rick,

Well, I suppose I should say something about why I think 1 term-1 
definition is best.  If you claim some term is narrower than  some other 
term,  then it better be clear which of the definitions of the first 
term is narrower than which of the definitions of the second term.  
Otherwise, I am searching for info on comets and suddently I am getting 
suggestions from the system that I should follow the path on optical 
distortions.  At which point I close that window.

Ed


Ed Shaya wrote:
> Rick,
>
> Wouldn't it work this way if we allow altLabels+prefLabels to be 
> non-unique?  So for this case,
> comet_coma would have id="comet_coma", prefLable="coma of comet" and 
> altLabel="coma".
> But the distortion coma would have id="coma", prefLabel="coma 
> distortion",  altLabel="coma".
> The point is to have unique terms for unique definitions.
> This is probably what you meant.  Yes?
>
> Ed
>
> Frederic V. Hessman wrote:
>> />/
>> />> I also would not underestimate the number of times a term has /
>> />> multiple meanings./
>> />/
>> /> Indeed!/
>> />/
>> />> That is unacceptable for us./
>>
>> Multiple definitions may be cumbersome, but they are reality.   I 
>> would rather have too many than too few to choose from.   Here's my 
>> best attempt at a "scenario" to show that computers should like 
>> multiple definitions:
>>
>> VO-app:  "What would you like to observe today?"
>> Astronomer: "The comet from yesterday's APOD."
>> (pause to look for an IVOA vocabulary which explains what "APOD" means)
>> VO-app: "OK"
>> (pause to find an HTN telescope)
>> VO-app: "The camera's FOV is only 10 arcminutes: what part of the 
>> object would you like to observe first?"
>> Astronomer: "Nucleus."
>> VO-app: "Completed, downloaded and displayed in your VO-viewer.   
>> Next target?"
>> Astronomer: "Tail."
>> VO-app: "Completed, downloaded and displayed in your VO-viewer.   
>> Next target?"
>> Astronomer: "Coma."
>> VO-app: "Slewing to Abell 1656..."
>>
>>
>>
>> Rick
>



More information about the semantics mailing list