Definitions

[Ed Shaya]

 Perhaps we should begin prefacing our subject lines with "Ont: ..." or 
" Vocab: ...".  If the Vocab folks vote to leave out definitions (which 
would be fine, because it speeds up the completion of that task), then 
when we discuss definitions and precise class inheritance we put "Ont: 
..." in the subject line.  Does that make sense?

Ed


Norman Gray wrote:
>
> Whoa, everybody -- I spend a day out of wireless range and come back 
> to... this?
>
> We are emphatically _not_ in the business of creating a dictionary of 
> astronomy here.  Our job here is to take vocabularies that already 
> exist, such as the A&A, AOIM or IAU-93 vocabulary, and SKOSify them.  
> Nothing more.
>
> If this process results in some ambiguities (and Ed and Brian have 
> illustrated where those might be), then that's just hard luck.  Those 
> ambiguities will remain until someone who is materially affected by 
> the ambiguities can make a case to expend resources on producing 
> updated vocabularies, and ideally a full-blown ontology, with the 
> ambiguities removed.
>
> That means that the IAU-93 term 'ABSOLUTE MAGNITUDE' will turn into
>
> <#AbsoluteMagnitude> a skos:Concept;
>     skos:prefLabel "Absolute magnitude"@en, "Magnitudine assoluta"@it;
>     skos:definition "Absolute magnitude"@en.
>
> Not very exciting, I know, but it's all we've got, and if we try to 
> add fuller definitions we will talk and talk and talk and produce 
> nothing.  Also, as Rob has mentioned more than once now, we don't have 
> the authority to start adding fuller definitions.  (By the way, I 
> think I misunderstood Brian's recent question about definitions and so 
> didn't answer usefully: Yes, we need skos:definition properties to be 
> present; no, we can't afford to expand these definitions as much as 
> we'd possibly like; sorry for any confusion).
>
> Brian asserts that a messy vocabulary will be fairly useless for 
> practical applications.  But the use-cases in the document are 
> practical applications, and they can be satisfied with occasionally 
> imprecise vocabulary terms.  If these use-cases are inappropriate, 
> Brian and everyone, then let us know, because the document is aiming 
> at satisfying those, and not others.
>
> Remember that the terms in the A&A vocabulary have no real definitions 
> at all, but they've been used for some years now without the sky 
> falling down, because the terms are _clear enough_ for their intended 
> purpose.
>
>
>
>
>
> I've added text to the beginning of the document which I hope 
> clarifies the scope.  Comments on that are welcome.
>
> We've rehearsed a variety of opinions about definitions here; I doubt 
> it would be useful to rehearse them again.  If anyone disagrees with 
> the text in the document as it stands can you phrase the disagreement 
> as an issue, summarising the alternative points of view, and I can add 
> it to the outstanding list.  Remember that 
> <http://www.astro.gla.ac.uk/users/norman/ivoa/vocabularies> will be 
> the most up-to-date built version, in between 0.x releases, and 
> <http://www.astro.gla.ac.uk/users/norman/ivoa/vocabularies/issues> is 
> the issues list.
>
> All the best,
>
> Norman
>
>