User requirements (was: Threads)

[Rob Seaman]

Norman Gray wrote:

> Andrea Preite Martinez wrote:
>
>> My modest opinion is that the discussion of vocabularies AND  
>> ontologies, of SKOS or OWL could be very important if related to  
>> the right requirements.
>
> [...]  It's probably no coincidence that the solution I favour --  
> of preening (or normalising or standardising) existing deployed  
> vocabularies, and avoiding creating new terms except where there  
> are obvious lacunae -- seems to match well the requrements I  
> perceive.  We can retain the value of a large vocabulary by adding  
> the specifications which allow us to link vocabularies where  
> appropriate.
>
> That might include SKOS-ifying the IAU Thesaurus, or even producing  
> a comprehensive update.

This strategy also seems close to the optimally damped path to  
introducing robust manly ontologies to the astronomical community.   
The jump from 2500 years of raw Aristotelian logic to reduced and  
rendered computer palatable semantics may simply be too far for  
practitioners in any domain discipline to make at one step.  Rather,  
the IVOA could demonstrate the value of such technologies by reviving  
the moribund Thesaurus into something lithe and lively and useful not  
just to programmers, but to astronomers writing their next papers,  
for instance.

Meanwhile, VO technologies like VOEvent should be coupled to just the  
right level of semantic scaffolding needed to  meet their own  
requirements.  Each VOEvent packet then would become a little  
semantic ambassador.  It does nobody any good to lard excess semantic  
payload into a packet that will never benefit either author or  
subscriber.

Ontological weapons of mass instruction like OWL (or whatever comes  
after OWL) will only build interest among any community when the time  
is right - when both the scale of the problem demands it and the  
particular community is ready for it.

Rob