Re: location vs. map scheme

Yes.  Of course, people refer to things indirectly all the time.  As long 
as we know when it is and isn't happening, that's OK.  So if, for example, 
I wanted to point you to Tim's book, I might very well send you:, 
but I would do so assuming that you knew, or would rapidly discover after 
following the link, that what I'd really identified was an Amazon page, 
which happens to be offering Tim's book for sale.  So, it's not completely 
off base to use the Google URI as an informal handle on a place.  Where we 
need to be careful is when we say that 
was published last month, and especially if we're doing it in an automated 
system where there's no human available to notice the ambiguity.  Maybe or 
maybe not Amazon published its page a month ago, but Tim's book has been 
out longer than that.


Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142

Erik Wilde <>
12/17/2007 05:21 PM
        Subject:        location vs. map scheme

hello noah.

> I'd put it a bit differently.  Google has registered, and 
> Linden Research has registered  That gives each of them the 
> right to associate resources with http-scheme URIs for those domains, 
> respectively.  So, if Google says that all URIs conforming to the 
><lat>,<long> refer to the corresponding 
> places on the physical earth, then they do.  If Google says that they 
> refer to a set of Google map documents that happen to depict those 
> on the earth, then that's what they identify.  I suspect that for 
> it's the latter (to the extent they've been careful in documenting one 
> the other.)  The URIs don't really directly identify the place:  they 
> identify Google maps of the places.

and i think this also is the point where it becomes clear that locations 
and mapping services are two different things. apart from the location 
coordinates, google has various parameters for the zoom factor and 
various other features of its mapping services. other mapping services 
have other features, and allow these to control via query parameter as 
well. that is good because it allows bookmarks which really identify a 
"view", not just the location.

so, ideally, location uris would be one thing, map uris another, and 
while for map uris i think a http-based scheme is a good solution 
(because the maps are mostly delivered over http anyway), the same is 
not automatically true for locations. of course, in many cases location 
uris will be mapped to map uris, but that is just one operation that can 
be performed on them.



Received on Monday, 17 December 2007 22:35:26 UTC