>Ted,
>
>I agree with what you're saying, but the actions implied aren't clear.
>Do you think we should go ahead and define a datum parameter, or just
>not preclude one? If we do, do we define values, and are clients
>required to process it?
>
>--Richard
>
I think we should have a datum parameter; I think we can do that
and give it a default value to be assumed when it is not present.
We should define the behavior for clients when they see a parameter
they do not understand.
The current doc has this syntax:
geo-URI = geo-scheme ":" geo-path
geo-scheme = "geo"
geo-path = geo-location
geo-location = latitude "," longitude [ "," altitude ]
latitude = [ "-" ] 1*2DIGIT [ "." *DIGIT ]
longitude = [ "-" ] 1*3DIGIT [ "." *DIGIT ]
altitude = [ "-" ] *DIGIT [ "." *DIGIT ]
I think we could generalize this to be
x, y, z, with the datum defining the units
for x, y, z. Add an optional parameter that
is datum=1*CHAR and specify that
it has a default value of WGS84. When it is
WGS84, x is degrees with the value limitations
currently in the draft, y is degrees with the
value limitations currently in the draft,
and z is altitude in meters, with the caveats
about *not assuming it means zero when
it is not present*.
Definitions of other datums are left for other
documents. We could specify a registry (either
an existing registry or an IANA one), but I
personally don't see a great need for that.
Again, just my two cents,
Ted
>
>Ted Hardie wrote:
>> At 8:24 AM -0700 6/3/09, Carl Reed wrote:
>>> I can check.
>>>
>>> Carl
>>
>> Both Galileo and GLONASS use datums which are distinct from
>> WGS84. PZ-90, which is the datum used by GLONASS, does differ
>> from WGS84, though the difference is minor (Wikipedia puts
>> it as "less than 40cm in any given direction").
>>
>> In some applications, 40cm is significant, and if you can eliminate
>> an error by noting that the reference datum was PZ-90 rather
>> than WGS84, it seems useful to me personally to do so.
>>
>> To back up a step, I have to ask what we think we're getting
>> by using a URI here. This is clearly not a URI intended to be
>> used to trigger protocol processing; it is instead a way of
>> capturing the context of the data. For those of us who use
>> GPS and systems designed to work with GPS, lat, long, altitude
>> and WGS84 are the syntax and context for location. This
>> group could agree to mint a URI scheme that works only
>> in that context and it would have a very simple syntax.
>>
>> It could instead have a slightly more general syntax and
>> mechanism for indicating context (the parameter indicating
>> datum). I think the latter is more in keeping with the purpose
>> of URIs. Shorn of even the possibility of using new or different
>> datums, I think we will either see other URI schemes,
>> the same URI schemes used with different datums (thus introducing
>> invisible differences in reference coordinates). Though I
>> usually fall into the "strings are cheap" view of URI schemes,
>> having a URI proliferation like geo:, geoGLONASS:, geoWGS84-2009:
>> etc strikes me as unlikely to be a long term win. And having
>> folks use the same URI scheme with different but unmarked
>> datums makes me very concerned indeed.
>>
>> The parameter method suggested before is easy to specify,
>> relatively cheap to parse, and it meets my personal test
>> for simplicity. I understand it doesn't meet others, but
>> speaking personally, I think the resistance to it doesn't
>> match the real likelihood of problems. Unless someone
>> builds a parser without reading the spec or referring to
>> a standard URI library, there's a fair chance they'll get it
>> right.
>>
>> My two cents,
> > Ted
>>
>>
>>
>>> ----- Original Message -----
>>> From: "Alexander Mayrhofer" <alexander.mayrhofer@nic.at>
>>> To: "GEOPRIV" <geopriv@ietf.org>; <vcarddav@ietf.org>
>>> Sent: Wednesday, June 03, 2009 7:18 AM
>>> Subject: Re: [Geopriv] [VCARDDAV] The "geo" URI draft
>>>
>>>
>>>>
>>>>> Also, the forthcoming Galileo system:
>>>>>
>>>>> <http://en.wikipedia.org/wiki/Galileo_positioning_system>
>>>>>
>>>>> will not, I understand, use WGS84.
>>>> I don't claim to be an expert in reference systems / geopotential models
>>>> etc. so please take the following information with caution (Carl? ;)
>>>>
>>>> I couldn't find any information about the reference system that Galileo
>>>> uses on the wikipedia page. However, a bit of research shows that
>>>> Galileo seems to use ITRF ("International Terrestrial Reference Frame"),
>>>> which, however, seems to be practically identical to WGS84, according to
>>>> http://www.dqts.net/wgs84.htm (web site by EUROCONTROL). Also, this
>>>> http://www.cambridgeconference.com/2007_conference_information/Conferenc
>>>> e%20proceedings/w2_5_cross.pdf presentation notes on slide 15 that
>>>> "WGS84 and IRTF are aligned".
>>>>
>>>> Which means that Galileo will use a reference system that is practically
>>>> identical to WGS84.
>>>>
>>>> Again, Disclaimer: i'm not an expert in this field.
>>>>
>>>> Alex
>>>> _______________________________________________
>>>> Geopriv mailing list
>>>> Geopriv@ietf.org
>>>> https://www.ietf.org/mailman/listinfo/geopriv
>>> _______________________________________________
>>> Geopriv mailing list
>>> Geopriv@ietf.org
>>> https://www.ietf.org/mailman/listinfo/geopriv
>>
>> _______________________________________________
>> Geopriv mailing list
>> Geopriv@ietf.org
>> https://www.ietf.org/mailman/listinfo/geopriv
>>
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
