some semantic puzzles from VOEvent

[Bernard Vatant]

Hello Rob

> Bernard Vatant comments on Roy's VOEvent issues:
>
> > (1) [...]  We would like a formal vocabulary so that, for example,
> > the computer can understand that "Supernova" and "SN" are the same
> > thing..

> >> everybody knows (or thinks so) or at least implicitly agrees upon
> >> what Jupiter *is* and looks like, what a planet *is*, and so on.
> >> There is a unique (implicit) interpretation in the common
> >> (implicit) ontology.

> "Jupiter" is a proper name for a specific entity.  That
> identification is simply a definition under the control of the
> appropriate IAU committee or some other international body.

OK. When I wrote "Jupiter", I did not mean the name, but was implicitly referring to this
"entity generally agreed upon to have some sort of permanent existence as a System Solar
major component, and which is called Jupiter ... etc." My point is that, generally
speaking, given the proper context, there is not much ambiguity in the astronomical
community about the definition of this particular object.

> There is  intrinsic ambiguity, however, with the definition of classes of
> objects - with nouns that don't correspond to proper names.

Certainly, this is a general issue. Actually all efficient disambiguation processes assume
there is inter-agreement on the class(es) of objects, but this agreement is often implicit
and impossible to formalize. You have a lot of ways to identify a person among six billion
ones (e-mail address, postal address, name + birthdate + birthplace, welfare number,
passport number, ... or any combination of those to make sure). But nobody has found so
far a consensual definition of what an human *is*. Ditto for the instance "Jupiter", more
easy to define than its class "planet". Think of poor Pluto :)

> What is a planet?  What is a gas giant?  For many purposes the answers to
> these questions are obvious, something like: a gravitationally bound
> non-stellar entity orbiting a star.  (Although it quickly becomes
> obvious that all definitions depend on other definitions.)  For other
> purposes, such as the characterization of planetary atmospheres or
> magnetic fields, the precise transition from Jupiter to
> interplanetary space may be hazy indeed.

Sure : even more simple questions are tricky : what is the minimal size for a planet? for
an asteroid? for a comet? Every value would be arbitrary, although nobody I guess would
call a body of mass a few grams orbiting the Sun a planet ... All those issues boil down
to the good old undecidable question "What is the minimal number of grains to make a
sandheap?"

> > (2) [...] One report says the Event is a Supernova, another says it
> > is a SupernovaType1a, another says it is a SN and another says
> > "bright source associated with a galaxy". Are these four in agreement?
>
> >> Observer X has taken an image later interpreted as being a SN 1c
> >> in NGC 4038. Other, looking to the same sky
> >> coordinates, have obtained other data that they interpret as being
> >> images of the "same" object in the "same" galaxy (whatever this
> >> "sameness" actually means ...).
>
> I presume synonyms such as "Supernova" and "SN" will be trivial to
> handle.

Sure enough. It's plain synonymy here.

> Less trivial might the the example at the bottom of section
> 3.7 of the VOEvent specification.  The object or process that has a
> proper name of "Tycho's Stella Nova" is classified as "SN 1a".  An
> assertion is made that this is associated with the object/process
> called "3C 10" which is classified as a "supernova remnant".
> Astronomy abounds with famous objects for which large numbers of
> reliable identifications and classifications of various sorts are
> available.  We are more interested in how to classify the question
> marks - in precisely the terra incognito of the emerging ontological
> maps.  A description of the science that was current last year is not
> going to prove exceptionally helpful.  A successful ontology will be
> one that includes an easy process for growth and hypothesis.

I agree completely. That's why I insisted on the different ontological layers that seem to
be blurred in the definition of an event, or for that matter an object. And, first of all,
the process of isolating such or such pack of data from the rest of the data continuum (so
to speak) is itself on the level of interpretation. This act of separating a particular
set of data (e.g., this particular pack of photons captured by this instrument under such
and such experimental conditions) is somehow arbitrary. Why this part of the sky? Why this
part of the camera field? One could try fuzzy but generic "objective" definitions, based
on some kind of discontinuity in the data background, either space-like for an object (a
more or less steady source of light is defined as an object because the sky background is
more or less black), or space+time-like for an event or a process (the data flow changes
at some point). But such definitions would be as arbitrary as the definition of a planet
minimal size. Moreover, this separation is most of the time based on pre-existing
classification schemes. "Our eyes see light, our brains see things".

IOW, there is no object or event pre-existing to interpretation of data.

> > (3) [...] If I have a lot of VOEvent packets that refer to the same
> > astrophysical event, can I extract a SED or light-curve (or time-
> > dependent SED) by federating all the separate observations?
>
> >> As long as you put together data from different sources, the
> >> agreement about those data being about the same thing supposes an
> >> (there again, most of the time implicit) agreement on a world
> >> model where this object is supposed to "exist".
>
> I think the issue here is broader than VOEvent ...

Certainly! That's why maybe the ontology approach should not stick too much to the
specific data model of VOEvent to begin with, but incorporate this into a high-level
description of the scientific *process*.

More to come ...

Bernard

**********************************************************************************

Bernard Vatant
Senior Consultant
Knowledge Engineering
bernard.vatant at mondeca.com

"Making Sense of Content" :  http://www.mondeca.com
"Everything is a Subject" :  http://universimmedia.blogspot.com

**********************************************************************************