(Sending this to wikipedia-l & OSM's legal-talk too)

On Mon, Jul 6, 2009 at 4:09 PM, Ævar Arnfjörð Bjarmasonavarab@gmail.com wrote:

On Thu, Jul 2, 2009 at 9:15 PM, Ævar Arnfjörð Bjarmasonavarab@gmail.com wrote:

...

So, what we should do is to author a document (on the wiki?) which clearly explains why such terms which restrict redistribution and fields of endeavor mean that free content projects like OSM can't use the data and will have to keep using SRTM.

Since nobody (especially someone with legal know-how) has offered to do this I've continued to my correspondence with NASA/USGS/METI using my own know-how and miscellaneous bits I've scraped from the recent ASTER threads on this list for support.

Below is an E-Mail I just sent to the NASA/USGS/METI people I'm corresponding with. I won't include the snippets I'm replying to since I haven't had permission to publish them, instead I'm going to replace them with little summaries of the original content. My summaries are one-liners while the originals are a few paragraphs so obviously information is lost in the process:

...

[What's this public OpenStreetMap forum you're referring to?]

It's being discussed on the main OpenStreetMap "talk" mailing list (and some other foreign language lists, e.g. the German one). Here's a list to the thread I started there:

http://lists.openstreetmap.org/pipermail/talk/2009-July/thread.html#38235

It's a public mailing list so you could sign up if you'd like, or continue corresponding with me and I could ferry information back-and-forth.

In any case I'll be submitting what I send to you to the aforementioned mailing list, but I won't quote any remarks from you (@nasa.gov/@usgs.gov people) unless I have explicit permission to do so. So I'll modify this E-Mail so that e.g. the paragraph I'm replying to now will be replaced by something like "[Where is this being discussed?]" before I post it. But that's bound to cause confusion so having permission to quote you when appropriate would be better.

I was hoping that someone with more legal knowledge would be willing to chime in but that hasn't happened already. I'm just a mapping hobbyist but I'll try to explain what would be about acceptable terms for open source/free software projects the best I can.

http://lists.openstreetmap.org/pipermail/talk/2009-July/038327.html

...

[Perhaps your intended use of the ASTER data is supported, e.g. if you derived tiles intended for some mapping software that would not be considered redistribution of the original product an could be pushed downstream] [However if you were intending to distribute the canonical ASTER data as-is that would be in violation of the terms]

I think I've correctly read between the lines of the download agreement in assuming that the purpose of that clause is to avoid Balkanization of the ASTER data, i.e. to make sure that NASA/METI will always be the canonical source for the source dataset.

If the terms were changed to something like:

You are not allowed to publicly distribute the original ASTER data files but any derived work can be redistributed freely with (only) the following restriction:

If you distribute, publicly display, publicly perform, or publicly digitally perform the Work or any Derivative Works or Collective Works, You must keep intact all copyright notices for the Work and give the original author (NASA/METI) credit reasonable to the medium or means You are utilizing by conveying the name (or pseudonym if applicable) of the Original Author.

Or something like that then the ASTER dataset could be used to its full potential by free data projects like OpenStreetMap, Wikipedia & others. But since there would be no restriction on the fields of endeavor that generated data could always be used to generate a DEM again, see a further explanation in this E-Mail:

http://lists.openstreetmap.org/pipermail/talk/2009-July/038327.html

For instance here's a map where the OpenStreetMap data which is under the Creative Commons Attribution Share-Alike (CC-BY-SA) license has been combined with SRTM contours:

http://osm.org/go/0CZyDpI--?layers=00B0FTF

The CC-BY-SA license specifies (as do most free software licenses) that when you distribute derived works you can impose no further restrictions on the data. That's a pretty much a universal feature of popular free content licenses to avoid data Balkanization and ensure compatibility so that e.g. someone doesn't specify the additional terms that you can't use the derived work for some specific use (e.g. military), or that you can't use it on a Sunday. Such accumulated restrictions would quickly make the data unusable for everybody.

Someone could take that map and generate a global DEM by analyzing the contour lines and distribute a global DEM derived from ASTER free of the original restrictions, thus circumventing the original limited use clause.

But in reality nobody is going to go to all this trouble and nobody is going to be confused about NASA/METI being the original and canonical source of ASTER data. The best support for this claim is that today nobody is confused about NASA being the canonical source for SRTM data. Even though it's in the Public Domain which means downstream distributors don't even have to attribute NASA for it (although they nearly universally do anyway).

In summary, not having restrictions on fields of endeavour will open the ASTER data to use by free content projects which otherwise wouldn't be able to use it, and nobody is likely to mistake NASA/ASTER as not being the canonical source for it, especially given attribution.

As for why I've changed the attribution in the latter paragraph of those example license terms is pretty much lifted from the CC-BY 2.0 license:

http://creativecommons.org/licenses/by/2.0/legalcode

The reason for changing the form of attribution from the current (When presenting or publishing ASTER GDEM data, I agree to include "ASTER GDEM is a product of METI and NASA.") is that asking distributors to include an exact string (in English) leads to what's called the Berkeley advertising clause problem (as pointed out on the list: http://lists.openstreetmap.org/pipermail/talk/2009-July/038237.html), see this page for an explanation:

http://en.wikipedia.org/wiki/BSD_licenses#UC_Berkeley_advertising_clause

Instead if distributors are merely asked to attribute the author (NASA/METI) that'll serve the same purpose in practice without the troubles associated with reproducing an exact string, e.g. the attribution can be translated or otherwise adjusted for the medium.

For instance if the ASTER data was used by someone to extrapolate the position of mountain peaks and this derived data added to the OpenStreetMap database we could add a node with a source=ASTER tag, e.g.:

http://api.openstreetmap.org/browse/node/308406749

Which is appropriate for a relational dataset, as opposed to reproducing "ASTER GDEM is a product of METI and NASA." for every object in it.

NASA/METI have updated their distribution terms with a FAQ in response to my questions:

https://lpdaac.usgs.gov/lpdaac/about/news_archive/friday_july_24_2009

Unfortunately the new terms aren't new at all, and they still look too restrictive to be incorporated into freely licensed datasets.