geo-extension-nonWGS84

From Microformats Wiki
Jump to navigation Jump to search

Geo Extension Straw-Man Proposal

Further to proposals for luna and mars equivalents to geo, the following is a "straw-man" proposal, to incorporate those ideas (and likewise for other bodies) into geo, and to make Geo available for other terrestrial schema than WGS84, in order that further debate may take place. Please feel free to critique it harshly but fairly!

Author

Andy Mabbett

Straw-Man

<span class="geo">

  <span class="body">
    Mars [1]
  </span>

  <span class="reference frame">
    [name of mapping schema] [2]
  </span>

  <span class="latitude">37.386013</span>, 
  <span class="longitude">-122.082932</span> [3]
</span>

Notes

  1. A list of acceptable, case-insensitive, values for 'body' would need to be drawn up (e.g. "Earth", "Mars", "Moon", "Venus", etc.) with "Earth" being assumed if none is specified.
  2. A list of acceptable values for 'reference frame' would need to be drawn up, for each body, with one being declared the default, to be used if no value is present (geo for Earth uses the datum of WGS84 by default. This extension would also allow for other terrestrial schema, of which there are many, such as OSGB36).
  3. As currently with geo, if the "latitude" and "longitude" classes are omitted, the two values MUST be separated by a semi-colon and latitude MUST be first:<span class=geo">37.386013;-122.082932</span>
Also:
  • If latitude is present, so MUST be longitude, and vice versa.
  • The same number of decimal places SHOULD be used in each value; zeros are significant
  • The Gazetteer of Planetary Nomenclature has coordinates for other planets, e.g. Venus and moons, e.g. Io.

Issues

  • Should other bodies be included in geo, or have stand-alone microformats?
  • What effect will this have on existing 'geo' parsers, and it is safe to ignore that?
  • What appropriate "reference frame" sets exist?
    • Earth: WGS84, ITRF2005, EtrS89
    • Moon: Mean Earth Polar Axis
    • Mars: IAU2000
      • 'Report of the IAU/IAG Working Group on Cartographic Coordinates and Rotational Elements of the Planets and Satellites: 2000', Celestial Mechanics and Dynamical Astronomy 82: 83-110, 2002 [1]
  • Moon is the preferred name, per [ http://planetarynames.wr.usgs.gov/ IAU nomenclature working group]
  • Is it appropriate to use the name "geo" (which means "Earth") for other bodies?
  • is body an acceptable class name, given that it's also an HTML element?

Planetocentric longitude

Geographical longitude on, for example, Mars is not 180 degrees west to 180 degrees east: 18N,226E is a legal Martian location.

This is from Timekeeping on Mars ([2])

Note that the modern standard for measuring longitude on Mars s "planetocentric longitude", which is measured from 0°–360° East and measures angles from the center of Mars. The older "planetographic longitude" was measured from 0°–360° West and used coordinates mapped onto the surface.

You can also check the description in JPL Horizons document, which uses planetographic longitude rather than planetocentric longitude, and gives much more longitude ranges used for different planets and moons, at http://ssd.jpl.nasa.gov/?horizons_doc.

Comments

thare

Comments from thare

  • geo does stand for Earth but it has previously been generalized for the planetary case. For example, the term geology is used for planetary bodies.
  • using body is fine by me and do not see a conflict as an HTML element.
  • by schema do you mean a natural surficial property like earthquake, volcano, or crater? Or do you also want to include man-made items building, house, landing site?
    • No, I mean, for example, WGS84, or the Martian or lunar equivalents. Is "schema" perhaps the wrong word?. Andy Mabbett
      • Perhaps "reference frame" would be a less confusing term than "schema"? --DavidCary 17:06, 5 Apr 2007 (PDT)
        • The HTML class attribute is a space seperated list, so 'class="reference frame"' is semantically equivalent to 'class="frame reference"' or 'class="reference" class="frame"' (though the latter is not syntactically valid). This could clash with, say, a theoretical microformat for footnotes, endnotes and references. Hyphenating ('class="reference-frame"') may be an improvement. TobyInk 12:29, 6 Jun 2007 (PDT)
  • I would suggest just using "Moon". For other spellings for planetary bodies I would look toward the IAU/IAG documentation (an international working group) IAU/IAG Working Group (WG) on Cartographic Coordinates and Rotational Elements
  • Since this is a simple case, I tried to stay away from this issue but just can't. What does Earth really mean? There a many definitions for the size of Earth that need to be resolved when using lat/lon coordinates. Is the size of the Earth defined by the WGS84 or NAD27 definition? For Mars, is it the definition used in 1991 or redefined in 2000? This has previously been solved in web GIS applications by using an "EPSG" code. For planetary bodies we are attempting to define a similar coded standard which currently is called the IAU2000 code or namespace. Thus for Mars the code might be IAU2000:49900.
  • still working - more later

Brent A. Archinal

(reproduced from e-mail to Andy Mabbett, by kind permission)

  • I appreciate the information you've provided and certainly the general idea of using "microformats" to indicate geographical location of points on solar system bodies is interesting.
  • As to any comments on the general idea or any specific implementation, I think my colleague, Trent Hare, has already addressed any primary concerns or questions (above).
  • The main issue seems to be that both the body, and the specific definition (technically a "reference frame") in question need to be specified, e.g. 1) Earth, ITRF2005, 2) Moon, ULCN 2005, 3) Mars, "IAU2000" or similar. Note that these frames are actually realizations of various "reference systems" 1) ITRS, 2) mean Earth/polar axis system, 3) IAU2000 Mars body fixed, respectively), but the system is apparent once the frame has been identified.
  • A secondary issue is that obviously a third coordinate may need to be specified, i.e. some sort of "radius" (from the body center of mass) or "elevation" about a reference surface. If the latter, then that reference surface needs to be specified as well (e.g. Mars, "IAU2000", "IAU2000 sphere").
  • If the microformat(s) were expandable to handle these cases when necessary (and others we haven't yet thought of) then obviously they could have a lot of utility.
  • There are plenty of GIS and standards organization formats out there for handling this type of data, even for the solar system case. I'm not at all an expert on what's available, but obviously for something like the microformats to be generally used, their advantages over these existing formats would have to be made clear.
  • As to usage here, we'll let others know of the existence of these formats, but I can't say how much immediate interest there might be in using them. We already (and Trent could say more) use a wide range of GIS and even flat file formats to carry e.g. lists of point-like data.
  • Regarding one other point on the web pages, I believe the "Moon" (capitalized) is the primary recognized name for the Moon. There is an [IAU nomenclature working group http://planetarynames.wr.usgs.gov/] that handles not only the names of features on planetary bodies, but of the bodies themselves, and I believe this is their position on this (since confirmed - Andy Mabbett).
Brent A. Archinal, Geodesist, Astrogeology Team, U.S. Geological Survey

multiple representations of the same location

I assume that someday, some other reference frame will replace the WGS84 reference frame. During the changeover, I expect many people to post 2 descriptions for a particular location (giving slightly different lat/long coordinates for each reference frame).

Even today, I can imagine someone wanting to tag a single location with WSG84, UTM (universal transverse Mercator), MGRS (military grid reference system), and the Maidenhead locator system. So readers can use whichever system they find most convenient, and ignore the others.

  • Would it be better to put all these ways of describing a location inside a single

<span class="geo"> ... </span> span? If so, what is the best way to make sure that each set of lat/long numbers gets associated with the correct reference frame?

  • Or would it be better to put all those descriptions in their own independent

<span class="geo"> ... </span> span? In that case, how do we indicate that they all (are indended to) indicate the same location, rather than a series of locations ?

--DavidCary 17:06, 5 Apr 2007 (PDT)

  • Why not use the same approach as used in GML? Use the srsName attribute when referring to a Coordinate Reference System (CRS) (or Datum). This way a coordinate can be recalculated to whatever CRS.

--Richard Flapper 23:46, 21 May 2008 (MET)

Other

I'm no expert on these matters, and am not sure if this is the right place for such comments, but these are are a few ideas/issues that came to mind after reading this page:

Multiple (Image) geo-tagging

A single XHTML document might make reference to (or include images of) multiple locations. It is not clear from the documents on this site how thee are to be dealt with. Should an <IMG ... > be child of the < div class=geo > or should both tags be on the same level in a parent < div >?

Geo-tagging areas and images of landscapes Geo-tags specify a single point which occupies no space. It might be desirable to be able to define areas, or single locations in a more fuzzy manner. Sometimes, as in the case of pictures of landscapes, you might want to indicate the Point of View (i.e. the location of the viewer)

To accomplish all of this would require three additions to the geo microformat.

  1. a precision attribute: This can be generally useful as even GPS devices have a margin of error, which you might want to indicate. Otherwise it can be used to give an idea of the bounding sphere of a large object like a building.
  2. an area attribute: This could use a system similar to that used by SVG for defining paths, but using lat and long for defining points. It is rather odd that, for example on wikipedia, entire countries or cities are identified by a single point.
  3. some sort of vector attribute: While the only use I can think of for this if for photos of landscapes, it would be good to be able to indicate that the Geographic information relates to the position of the viewer rather than the subject. The vector would indicate where the viewer (camera) is looking

I'm not sure whether the area and vector features are well suited to be extensions of geo or should be independent microformats. Particularly the vector might be better suited for inclusion in some 'EXIF-data equivalent micro format'. Unless of course it you can think of a use beyond photos of landscapes.

--Inkwina 01:54, 30 Nov 2007 (PST)

Astronomical Coordinates

Comments from Stuart

Rather than limit the format to the surfaces of planets and moons, would it be possible to include astronomical coordinates for objects in celestial coordinate frames? This would be very useful for tagging astronomical objects (e.g. M41 or NGC 7027).

If you do not want to break the existing geo tag might I suggest the more generic term coord?

As suggested already it is important to define a reference frame. The term coordinate-frame (or coord-frame) might also avoid ambiguities that exist for the word reference. In astronomical applications it is necessary to provide both a frame (e.g. FK5) and an equinox/epoch (e.g. 2000.0). Astronomical coordinate frames that are widely supported are: ICRS, FK5, FK4, ECL, GAL, SGAL.

Harmonization with Geo URI and h-geo

As geo has always been supposed to match vCard's geo thing (currently RFC 5870), I believe we should change the "reference frame" class name to simply "p-crs". The semantics are pretty well-defined there.

As for the class-less format, I propose just referring to RFC 5870 parsing. Stick a geo: prefix to the front and parse it. --Artoria2e5 (talk) 04:18, 17 February 2021 (UTC)

Related pages

See also

NASA shares space with Google (2006-12-20)