[OSM-talk] WolframAlpha uses OpenStreetMap data

[SteveC]

So I've had phone and email chats with stephen wolfram about everything they're doing because incidentally I had an internship there a decade ago.

They love OSM. What's good is we're at a nice tipping point with many who want to use OSM and help it succeed but are having trouble figuring out how to. This is the same conversation I have with lots of companies right now, so abstracting away exactly what they and other companies keep telling me, and paraphrasing the message from a few million and billion dollar companies, here's what I hear:

They would like to link to us directly but don't think a) we can handle the load and b) don't think it would be a good user experience to dump people on to osm.org, what with the site design. Of those that would like to help us scale the servers, they don't want to be seen to try to 'take over' by hosting OSM in their data centers but can't really justify throwing money at us when they have perfectly good resources we could use. Most would like to have some feedback or edit button from their site that goes to OSM but know we'd fall over from the load, so they'd basically be down to forking the OSM dataset and hosting it themselves which nobody wants to do because then you're a bad guy. I've had a couple of offers of design and coding help but most are scared by the responses they've seen to other people who've tried to help with obvious things here, and don't want community wrath to hurt their brand image.

Basically there's a big decision tree that I have half worked out that follows from 'big company wants to *significantly* help OSM but how?'. I've been through this decision tree about 6 or 7 times now I think. It would be interesting to graph it, and each end point node in that tree is 'we can't do that because of X'. By '*significant*', I mean throwing millions of dollars at the problem (OSM) because they're already throwing tens to hundreds of millions at NT/TA and so OSM a viable side bet. Of course you could say 'start small' but the problem there is that it's usually much easier to release large resources than small in large organisations.

There is a slight contradiction here though because the other thing I hear a lot is they'd like to try something with us but keep it quiet - i.e. try something small, but that's extremely hard with an open community. One of the many reasons to try something quietly is that you might not want to piss off your multi-billion dollar data supplier, NT or TA and OSM isn't at the point yet where it's a drop in replacement.

So you have a sort of prisoners dilemma where we're the company is acting rationally, OSM is acting (sort of) rationally... and yet it leads to the worst possible outcome: no big help given. I find it frustrating on both sides of the table.

So what will happen is that they do something themselves without OSM involvement and it just pops up in the world one day like wolfram alpha or flickr did and then build their own site reminiscent of maps.cloudmade.com with either just plain browsing the map functionality or some simple tools. You will see this happen multiple times with other companies over the next year or two.

That's of course totally fine and allowable by the license blah blah blah, but what we, OSM, lose as a community is all of those eyeballs to help fix the map. I think that's a terrific loss. There are things we can do to fix it, and there are things they can do.

I'm hoping that we'll reach a point where the dam will bust and it will be cool and fashionable to support OSM openly and loudly, but we're not quite there yet.

Steve

stevecoast.com


On Jun 17, 2010, at 5:30 AM, Tobias Knerr wrote:
> 16.06.2010 18:38, Martin Koppenhoefer wrote:
>> it might be OK or not, but generally we explain in our wiki how we
>> like to be attributed in the web-context, that is supplying cc-by-sa
>> linked to creative commons and Openstreetmap, not too difficult, is
>> it?
> 
> Apparently it is counter-intuitive at least for users not reading our
> wiki, otherwise we wouldn't see those incomplete attributions this often.
> 
> I even doubt that insisting on that license URI is a good idea at all.
> The URI isn't really useful in practice as long as people know where the
> data comes from and are can easily go to osm.org (where we have the
> possibility to advertise our license conditions in any way we like).
> It is cumbersome in many environments and enforcing it would require a
> lot of unpleasant communication: Informing others that they don't
> conform to a license isn't a great start for relations. Initiatives such
> as the "Lacking proper attribution" wiki tables have caused conflict
> even within the OSM community.
> 
> Tobias Knerr
> 
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk
>