[Data-modeling] observatories

[Jonathan W. Lowe]

If one of the modeling problems being discussed in this thread is how to
model the location of a telescope that is comprised of a geographically
distributed collection of structures, then the
(nascent) /location/geometry type might be useful.  This is more a
placeholder than immediate solution, as the documentation and a query
API are still being developed, but here's the gist of the modeling
approach:

To model a location that is just a single point, use
the /location/geocode type (as per normal), but use /location/geometry
to model locations delineated by more complex geometries such as lines,
polygons and "multi-geometries".  One of the multi-geometries is
multi-point, which just a set of anonymous points stored in a BLOB and
referenced by a /location/geometry topic.

The focus to date of this work has been on lines and polygons rather
than multi-points, so others (Jamie?) may prefer a different approach
for modeling groupings of points.  But if deemed
appropriate, /location/geometry can store a multi-point geometry as a
single entity.

More detail will follow in the next 2-3 weeks as we lurch toward a
proper release of complex geometry support in Freebase.

(This approach obviously doesn't include assignment of a location to a
moving (er...orbiting?) structure such as the Hubble telescope.)

Jonathan

On Fri, 2008-04-18 at 16:33 -0700, brendan wrote:
> Yes, the issue of conflating observatories and structures/buildings is  
> a problem.
> 
> In many cases they are referred to as "telescopes" and the  
> *collection* of telescopes and perhaps some other buildings is what is  
> called the "observatory".  In some cases the telescopes are considered  
> to *be* the observatories.
> 
> /architecture/structure is meant to be pretty generic, though, and I  
> still think it's useful here.  It's meant for any man made, fixed  
> thing (and it includes properties describing when it was build/opened,  
> where [on earth] it is located, who owns it, etc)
> 
> So perhaps there should be a type /astronomy/observatory_telescope  
> which describes these individual structures we're talking about, that  
> includes the type /architecture/structure.
> 
> and a type /astronomy/observatory that describes the collection of  
> telescopes that make up the observatory
> 
> the property:
> /astronomy/observatory/telescopes -> expected type: /astronomy/ 
> observatory_telescope
> 
> As for space-located telescopes, perhaps space_telescope could just be  
> a co-type.
> 
> The original list I sent out has over 400 reconciled "observatories".   
> I'd guess most of them are actually the simple case: a telescope  
> housed in a building that people have named "an observatory."  But  
> I'll now need to sort through them to figure that out.
> 
> Ok, looks like this won't be happening this week ;)  Let's continue  
> the conversation...
> 
> Brendan
> 
> On Apr 18, 2008, at 3:50 PM, Alexey Goldin wrote:
> 
> > Many observatories have a lot of buildings, many are not in any city.
> >
> > How would you classify this  :
> >
> > http://www.ifa.hawaii.edu/mko/
> >
> > ?
> >
> > On Fri, Apr 18, 2008 at 5:40 PM, Robert Cook <robert at metaweb.com>  
> > wrote:
> >> It might make sense to co-type as /architecture/building and /
> >> architecture/structure so that you can put the city in the address  
> >> CVT.
> >>
> >> R
> >>
> >> On Apr 18, 2008, at 10:44 AM, brendan wrote:
> >>
> >>>
> >>> http://en.wikipedia.org/wiki/List_of_observatories
> >>>
> >>> data monster hungry, yumm!
> >>>
> >>> I'm going to scrape these out and pump them into freebase. Seems  
> >>> like
> >>> a simple type: /astronomy/observatory (included type /architecture/
> >>> structure) is called for.  Astronomy admins? What think you?
> >>>
> >>> Emailing the list just in case anyone happens to be doing the same.
> >>> fyi.
> >>>
> >>> thanks,
> >>>
> >>> Brendan
> >>> _______________________________________________
> >>> Data-modeling mailing list
> >>> Data-modeling at freebase.com
> >>> http://lists.freebase.com/mailman/listinfo/data-modeling
> >>
> >> _______________________________________________
> >> Data-modeling mailing list
> >> Data-modeling at freebase.com
> >> http://lists.freebase.com/mailman/listinfo/data-modeling
> >>
> > _______________________________________________
> > Data-modeling mailing list
> > Data-modeling at freebase.com
> > http://lists.freebase.com/mailman/listinfo/data-modeling
> 
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling
>