- See also Maps
Even on Earth within historical time (back to roughly 10,000 BCE = 12,000 years ago, see en:Ely in en:Cambridgeshire which has such a long documented history), some non-obvious considerations apply to specifications of blocks of spacetime. But certainly a key requirement is that coordinates of a particular place on the Earth ought to be as stable as its rocks and strata, so that a 'location' on the Earth's surface has its usual meaning.
Here are some of the non-obvious considerations of a terrestrial coordinate system going back 12,000 years and useful in en:geology and en:geography, en:paleontology and en:biology, en:ecology and en:climate, en:archaeology and en:history, while still being absolutely unambiguous about what time and place means in the present:
- Absolute blocks of time are often represented inexactly, e.g. "February 8, 2003" can represent one of thirty-nine different ranges of time, one in each en:time zone. In any project that crosses time zone borders, as this one does, you have said NOTHING if you have not implied or added the time zone. A default of UTC (Greenwich Mean Time) is the only sane standard when this is ambiguous, which makes many articles that assume PST or EST simply wrong. This must be part of the standard.
- Relative blocks of time, e.g. 'that week', 'a day earlier', etc., are unambiguous and actually better referents, if an absolute frame has been established. But those informal characterizations should be tagged to note that "that week means -7 days from reference", "a day earlier means -1 days from reference". This would be pretty easy to do automatically and sum up things that happened on a given day (which we do manually at this point with great effort. (This is the equivalent of tagging 'he', 'they', etc. as referring to a reference name - again relatively simple to do automatically and sum up all the things a given person did or said).
- Many calendars (e.g. en:Chinese calendar, en:Islamic calendar, en:Persian calendar) are still in use in the world. Others tend to alter the year but accept the en:Gregorian calendar for the "date". But in the Islamic calendar, for instance, "13 December 1979" is "24 Muharram 1400 A.H.". Chinese and Islamic calendars use lunar months, which is more reliable than Julian/Gregorian dating, so the phase of the moon is actually a better reference in most cases if poor records exist.
- Earth en:latitude and en:longitude are relative to the date, due to en:polar drift and en:continental drift - so a latitude/longitude pair far in the past is not necessarily the same place without some compensation. A complete solution to this problem would be able to accomodate this all the way back to the single continent en:Pangea which was once all land on Earth, and has split to become today's continents.
- Earth en:ecoregions are the most basic reference for living things, and a scientific and standard map of them now exists. Their borders determine most of what one would want to say about en:biology or en:climate, so it is better to organize this information around them, than around say national borders, which is a poor way to organize data about en:physical geography or en:ecology. This should be part of the system as early as possible, so that the ranges of animals can be specified in terms of ecoregions rather than countries or gross physical features.
If you know of an existing DTD that actually does this correctly and solves the problems of en:history and en:geography above, by all means dig it up and let's talk about it, and improve it here. But it appears at present that the Wikipedia already has problems that we should solve before we get into the issues of a Wikipediatlas.
- which problems ?
I agree. I looked up this "polar drift" (or "Chandler's Wobble"). It looks like it's not really a problem unless you want to know your location accurate to 1m. I really can't see any problems here that need fixing. -- Tim Starling
- As noted on the Maps page, HEML (Historical Event Markup Language) may be useful for this application.
- The Open GIS Consortium (OGC) has several standards that are
related to this topic, such as coordinate systems and WWW mapping services, with some reference implementations available.
- There is a a project on SourceForge that is working on space time analysis.