W3C home > Mailing lists > Public > semantic-web@w3.org > September 2009

Re: [ontolog-forum] Event Ontology

From: Dan Brickley <danbri@danbri.org>
Date: Thu, 3 Sep 2009 02:05:50 +0200
Message-ID: <eb19f3360909021705y583cb124i76d86b0a59a35019@mail.gmail.com>
To: AzamatAbdoullaev <abdoul@cytanet.com.cy>
Cc: "[ontolog-forum]" <ontolog-forum@ontolog.cim3.net>, semantic-web@w3.org
On Wed, Sep 2, 2009 at 8:12 PM, AzamatAbdoullaev<abdoul@cytanet.com.cy> wrote:
> John Bottoms wrote:
> "An interesting question might be: how do we migrate from detecting simple
> events to more complex ones. Or, we might discuss how those events are
> represented in an ontology so that they are anticipated and detected
> correctly."
>
> Indeed. Being the universal and ubiquitous real world phenomena, events
> deserve a full scale discussion. Events as happenings, occurrences and
> occasions are the most familiar things, as being everywhere and every time
> to everything. They are key elements in the very Nature, from subatomic to
> cosmic scale, as well as in life, mental life, social life, in technology
> and industry, mass media and computing, particularly. Natural events,
> physical events, chemical events, biological events, mental events, social
> events, political events, cultural events are just some types of generic
> Event.
>
> Imo, most confusion could be avoided with its adequate defining as "a change
> happening at a given place/time, followed and caused by some other events
> (changes, acts, or actions)."
>
> I am inclined to think that the current downgrading of event as a real world
> "nonevent" comes from its poor ontological study.

I don't get to say this very often, but ... well I think you might
enjoy reading "Events and their Names" by Jonathan Bennett, if you've
not run into it already. Google seem to have the whole thing online
for 'preview', see
http://books.google.com/books?id=hH1eCoFxlWUC&printsec=frontcover&source=gbs_v2_summary_r&cad=0#v=onepage&q=&f=false

Modelling things in terms of events is very seductive. Once you get
into that style of thinking, everything can seem best modelled that
way. Unlike the rest of the changing world, event descriptions don't
go out of date. This appeals to metadata people.

I don't find it productive to ask whether or not events are "real"
(any more than whether eg. beliefs or desires are "real"...). But we
can ask how useful the concept of an event is, how it works informally
and under various attempts at formalisation.

There are certainly costs involved when trying to make heavy use of
the notion of event, in data systems if not in general. Firstly, in an
open world where data may always be missing, partial, etc you don't
easily go from a pile of event descriptions, to knowing anything about
the current state of the world. If you tell me your car is blue, and
that's true, ... great, I know the colour of your car. If you tell me
you painted it blue yesterday, or last week, or last year, ... then I
know about an event in the past. I can't jump from that to a good idea
of your car's current colour without having a commonsense reasoning
system. And most people aren't holding their breath waiting for such.
So in practice, having a mountain of event descriptions might be less
useful than having some state-of-the-world descriptions that you can
treat as reliable.

A more abstract consideration is event identity; how can we count,
compare and cross-reference events? The Bennett book goes into this at
great length...

So I'm quite ok with the Allen & Fergusson quote in
http://motools.sourceforge.net/event/event.html#intro ... with the
aside that I wouldn't bother saying events don't exist, for reasons
pretty much in line with Stephen Stich's in
http://books.google.com/books?id=hbcsiGdZ0V0C&dq=stich+deconstructing&printsec=frontcover&source=bn&hl=en&ei=-wefSqW0Monp-QaA2d3WDw&sa=X&oi=book_result&ct=result&resnum=4#v=onepage&q=&f=false
...

cheers,

Dan
Received on Thursday, 3 September 2009 00:06:27 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:14 UTC