[Data-modeling] Events, again

[Ryan Shaw]

On Tue, Mar 18, 2008 at 11:14 AM, Kirrily Robert <kirrily at metaweb.com>
wrote:

>
> Based on various discussions and some attempts to mentally model how it'd
> all work, here's what I'd like to do for now:
>
> 1) remove the types "Historic period" and "Historic event".  They are
> really just events (though, in the case of historical periods, long-running
> ones.)  Anything that is currently typed as one of these things should be
> typed as "event" instead.


I agree that the "historic" modifiers should be removed, but I disagree that
periods are just long-running events. An event is something that happened:
even if we may not know exactly when or where it happened, in theory if we
could go back in time and witness it we could establish a specific time
and/or location. A period, on the other hand, is an abstraction. The classic
example is the Renaissance. There is no "fact of the matter" about when or
where the Renaissance occurred. Instead, it is an interpretive judgement,
about which there may be more or less consensus, but no "truth."

So I think that the distinction between event and period is worth keeping.
Perhaps period could just become a marker for a kind of topic. Ideally there
would be a way to associate a period with an approximate date range
reflecting current consensus about the time range it describes (as opposed
to the date(s) on an event which are a best guess about when something
actually happened).


> 2) modify the "event" type by removing "people involved".  This should be
> specified in more detail in related types.  Eg. "military conflict", which
> should become a kind of "event", already has properties for "commander",
> which is more useful in that context.


Agreed.


> 3) continue discussion of a generic "timeline" type, which would allow the
> aggregation of events into arbitrary interesting groups.  For instance, you
> could have a timeline of volcanic eruptions, or of major wars through
> history, or of dynasties in ancient Egypt.  There was some discussion about
> this at
> http://www.freebase.com/view/discuss/time#/guid/9202a8c04000641f8000000006f4819eif you're interested in reading the thread.  The problem with timelines is
> that it's hard to include eg. "birth of Albert Einstein" without creating an
> event for that, whereas ordinarily in freebase we'd just look at the "date
> of birth" property on the person to get their birthdate.  And I really don't
> think we want to promote all dates in the system to be events, do we?
>  Anyway, worth discussion, but can happen separately from the two points
> above.


Aren't things like this better handled through custom MQL queries?

Ryan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/data-modeling/attachments/20080318/b95a9487/attachment.htm