Elevation info in the osm database

The primary objective of the OSM project is to store geo data, which are basically tagged nodes, or collections of nodes described in various “meta” constructs like ways, relations, etc… Every and all of the basic building blocks - a.k.a. nodes - are defined by their lon/lat values. However, the earth is not flat and in many scenarios it would be very useful to know also node(s) elevation. Some of such cases are:

  • export osm route relations into gpx files. From gpx files that contain elevation info, consumers could infer interesting information, for example energy needed to traverse a particular route.
  • list elevation of places (villages, towns, airports, …) on maps
  • use elevation data in engineering and scientific estimates
  • etc.

Currently, if a data consumer needs elevations, osm data must be augmented by information from additional sources, which requires engineering expertise and thus poses a barrier to use of the data. Many times these efforts are made for the same data sets by many independent consumers, which leads to waste of human effort and energy in general.

I would like to get feedback from the community about the following proposal, whose ultimate objective is to make elevation information an integral part of OSM data.

  1. resolve and clarify issues with the ele tag: http://wiki.openstreetmap.org/wiki/Key:ele; That discussion should include also addition of a new tag giving accuracy of the elevation.
  2. encourage builders of OSM data editors to preserve elevation info, if it is present in the imported sources (gpx traces, geotagged images…)
  3. create automated processes that would add missing elevation data for nodes imported into OSM DB. There are sources of elev. data out there (SRTM) and even though their accuracy is in places questionable, even less accurate data would be useful in many situations.
  4. use the software solution developed for functionality in the above point (3) to add missing information to all of the existing nodes in the database.

I do not have enough insight into technical aspects of this proposal, but I believe there are people who do and will provide a commentary about feasibility from that point of view.
Of course I might also be unaware of similar proposals that could have been made in the past. I checked the osm forum and did some web search prior to writing this post, but it is possible I have missed something.