At 17:09 12.04.99 -0700, Andrew Daviel wrote:
>On Mon, 12 Apr 1999, Harald Tveit Alvestrand wrote:
>
>> hi,
>> stumbled across your I-D "draft-daviel-html-geo-tag-00.txt" on the Web
>> (your site, not the IETF site).
>>
>> Some comments:
>>
>> 1) This may be best done as a W3C activity, since they're doing most
>> HTML related stuff these days?
>
>Maybe, but I can't afford to join. I think it costs $5000/year, or
>something.
The only way to access the W3C is to get in touch with people on the
Consortium. They have ways around their own rules.
Try Henrik Frystyk Nielsen, for instance - you will find him on their Web.
>
>>
>> 2) you might want to look at the various proposals for geodata
representation
>> in the DNS, especially RFC 1712 and RFC 1876.
>> 1876 in particular adds the encoding of extent - something I find
>> extremely interesting when talking about named entities - I don't
>> want the city of Trondheim reduced to a point in the town square, and
>> with extent, "overlaps", "within" and "near" are much more codifiable.
>
>I'd seen RFC 1876, and referred to it by saying that this was different -
>location of what is beeing described, not an IP node.
>Hadn't seen RFC 1712. Interesting, it has a serious error - confusing
>Latitude with Longitude...
>
>I take your point about extent, but I want this to be accessible to
>a wide audience, and I don't want to confuse people. Maybe in version 2..
>
>Anyway, how do you suggest specifying extent? Many GIS metadata schemes
>define a rectangular area, because most maps are rectangular aligned
>North-South. Aerial photographs and satellite images are a bit more
>tricky, and may be described in terms of a polygon or linear track.
>The metadata schemes for these are horrendous for mere mortals and I'm
>looking for something to describe a restaurant, ski resort etc. A single
>point that, if someone goes there, they'll realize they've arrived, is
>probably good enough for most people.
I like "size" as a number that gives distance from centre to edge,
with the implication that the object you are seeking is contained within
that sphere. Note that 1876 gives a VERY rough measure for this; we're not
aiming
for accurate extents here!
This allows, for instance, sizing of circles on a map, so that
"Trondheim city square" is a circle of diameter 100m, while "Trondheim"
is a circle of diameter 15km, centered on the same spot.
It tells a user that to differentiate between the city square and the
city hall, the geographical designator "Trondheim" is useless.
(They're about 500 metres apart)
Harald
-- Harald Tveit Alvestrand, Maxware, Norway Harald.Alvestrand@maxware.no