We had quite a lively dicussion on the list and in face to face meetings
over the last few days about the "geo" URI, and i'd like to get the
discussion back on track so that we can progress the document - much of
what was brought forward on the list veers off what imho we had agreed
in previous meetings and discussions.
Regarding the specification of Coordinate Reference Systems:
------------------------------------------------------------
In the session in San Francisco, there was concensus that simplicity is
one of the primary desireable properties of the format. In the
discussion, it was pointed out that it seems to be a good idea that
pushing the complexity to the sender rather than the receiver is good.
This led to the conclusion that using WGS-84 as the most widely used
reference system was a good thing, and there was consensus that
concentrating on a single reference system has less potential to confuse
authors and receivers.
http://www.ietf.org/mail-archive/web/vcarddav/current/msg00992.html
http://www.ietf.org/mail-archive/web/vcarddav/current/msg00966.html
It was noted that almost any reference system can easily be tranformed
into WGS-84 with freely available open source software (as Ivan has
pointed out as well), however, it is desireable to push the load of
transforming to senders rather than clients - clients might not have the
resources available.
*However*, concerns were raised that if in the future, another reference
system would become the predominant CRS, the scheme would still be stuck
with WGS-84 without an option to specify a different CRS (GLONASS and
Galileo were mentioned, however Carl Reed pointed out that the
differences are negligable:
http://www.ietf.org/mail-archive/web/geopriv/current/msg07302.html
Ted Hardie then proposed that we should not *preclude* the potential
future use of other reference systems:
http://www.ietf.org/mail-archive/web/vcarddav/current/msg00989.html
Which i have finally added to the draft. However, this was never
intended as an "let all CRSes bloom" option for the URI - as outlined
above, this was more intended as an "emergency exit" once WGS-84 becomes
obsolete (which is not going to happen anytime soon, as far as i
understand - note that WGS-84 is being adopted over time - please don't
fall for thinking that "84" in "WGS-84" has the same meaning as "95" in
"Windows 95" - there's no need to "upgrade"!)
I think that letting thousands of CRSes bloom for the URI is dangerous -
it will clearly limit interopability, and impact the original goal of
having something that's simple, human readable, and easy to use (even
for people who don't even know what a CRS is, and just enter the
coordinates into a mapping service or their car navigation system).
So, please let's stay with primarily one CRS, and use the parameter as
an "emergency exit" mechanism only.
Obviously, that also means that i am opposing the use of an URI/URN as
the specification of the CRS.
Regarding uncertainty
---------------------
This dicussion recently came up (and, the inclusion of URI parameters
actually opened that "can of worm"). Generally, i have the impression
that the general consensus is that such information would be useful.
However, i would very much like to declare the "uncertainty" parameter
out of scope of the current document, and create a different document
for that parameter (Martin Thomson has indicated he would be willing to
work on this?).
There are various issues with this, including:
- dimensionality
- units
- relation to "crs" parameter?
- comparison of two URIs?
And i am sure there are more. Again, we should take care here to not
violate the original goal of simplicity and brevity - we have an
established mechanism for conveying more complex geometries called
PIDF-LO anyway...
Way Forward
-----------
I have only heard back from very few people within the group, so will do
the following to the draft document:
- Add a registry specification for the URI parameters, specifying that
registry
- Clarify text that significant digits should not misinterpreted as
uncertainty
- Extend / clarify the motivation behind a default CRS
- Add text in the style agreed with the chairs in hallway discussions
about privacy
I'd really like to Last Call that before Hiroshima if possible,
especially since i might not be able to attend there.
Comments?
Alex
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
