[OSM-talk] Things People Say

[Thomas Davie]

if (*ra4 != 0xffc78948) { return false; }

On 30 Dec 2011, at 15:57, Pierre-Alain Dorange wrote:

> Serge Wroclawski <emacsen at gmail.com> wrote:
> 
>> I agree wholeheartedly with the view that OSM should be providing
>> maps. I think as long as we continue to cling to this idea that we
>> want third parties to make the maps, then we limit the project's
>> viability, its success and its overall accuracy.
> 
> I've allready express my own opinion, but i do not really understand the
> point.
> 
> OSM hosted one rendering on its own server (using mapnik software)... Do
> you want OSM to provide other maps (with different rendering) ?

Sure – and we already do, e.g. OpenCycleMap the transport map, the osmarenderer layer, etc.

> I assume that what you want is OSM to provide a service without limit
> for other to use their tile server as base ?

Not at all – though I can imagine that we could make some cash by offering to serve tiles to people if they pay for our hardware/bandwidth, plus a bit.  Ofc, this requires quite a major investment in people to run it, so it's probably not immediately possible right now.

> I feel very unconfortable with this option. Managing a tile server has a
> cost (OSMF handle it now) and this cost goes higher has many user use
> it. We see actual limitation this summer (limited bandwidth).

I agree, providing unlimited tile data to unlimited numbers of people, for free is clearly not a reasonable option.  The status quo works pretty well though.

> So their is here 2 options :
> 
> 1. providing a free service open to everyone with no limits (google
> competitor to summurize) that will be adapt to demand (more power, more
> RAM...) so more cost every user use it.
> To handle cost there is 2 options : keep the service free (more
> donation, more money from ?)or made the service commercial (big users
> pay, this is what google is providing). This will require adding an API
> key (like google, bing or cloudmade).
> 
> 2. providing a basic service for small users (as it's now) and limit big
> usage without providing an laternative and let commercial compagny
> (cloudmade or openstreetmap has start this) providing services for big
> users.
> 
> Note that option 1 has a terrible issue : been a competitor to
> commercial compagny that would do business with OSM data...
Is that a terrible issue?  If a company is offering OSM with no added value on top of it, why do we care about competing with them?  If they're doing a better job than us, and make it financially unviable for us to do it, then all the better – we can stop worrying about this problem.

> My opinion is that OSM should provide a basic service (as now) without
> commercial issue (option 2). Big users should build their own tile
> servers or buy this service from commercial compagny : it's not OSM
> business (OSM is not in business).
There's no reason why it can't be run like one though – charities and philanthropic organisations usually are for one very good reason – it gives them sustainability.

Note – I have no problem with carrying on with 2 either for now, or indefinitely.  What I don't think is a good option though is option 3 that some people seem to be kicking about – that is, drop all rendered output to users or people considering what they can do with OSM, and instead concentrate on just having a huge database.

Tom Davie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20111230/b1ef000c/attachment-0001.html>