[OSM-dev] Multiple temporary API instances

Brendan Morley morb.gis at beagle.com.au
Wed Aug 24 13:15:19 BST 2011


I've had some success running the rails port on Amazon Web Services 
instances.  Each one can run on its own DNS address.

However, I haven't tested them under recent versions of the rails port 
(e.g. in the last year) - in particular I've only used Potlatch 1.

Let me know if this approach may work for you, and I'll send further 


On 24/08/2011 9:53 PM, Josh Doe wrote:
> I'm interested in providing a facility whereby users can create and
> edit copies of OSM data for a small region, say 16 sq miles (~40 sq
> km). This is for a project [0] that I'm working on which would allow
> for users to experiment with adding/removing walking and bicycling
> infrastructure around a school to see how the changes impact
> accessibility (i.e. how many more kids can walk or bike to school
> given a specific change). While I'm quite a ways from implementing a
> full web app for all of this, I'd like to understand the challenges
> I'd be facing.
> I'd like to allow all of this to be done via a web interface, and I
> think Potlatch2 would be a good editor to provide. For that I'd need
> to set up a server side API, for which I believe the "rails port" is
> the only option. For a given school, it's possible that there could be
> dozens of users making changes to the same area, and each user may
> want to try out different combinations of changes. Therefore it makes
> sense to introduce the concept of "scenarios", where each scenario
> consists of a copy of the main OSM data. Users should be able to
> create, copy, and destroy these scenarios on demand.
> <begin-uninformed-conjecture>
> Each scenario should likely exist as a separate set of tables in the
> database sharing a unique prefix. A fork of the rails port would then
> need to be modified to allow the URL to specify which scenario
> (prefix) to use, and would then query those tables. Potlatch2 could
> then point to this scenario-dependent URL for the API, and everything
> would work magically. :) Seriously though I have no actual experience
> using the rails port or Potlatch2 (as a developer), this is all based
> off my reading of the wiki documentation, and I'm sure things are much
> more complicated and this project would require extensive modification
> of the rails port, if not necessitating writing something from
> scratch. I also have no idea how this sort of implementation would
> scale up to thousands of scenarios.
> Thanks for your input,
> -Josh
> [0]: http://wiki.openstreetmap.org/wiki/Safe_Routes_to_School_Mapping_Toolkit
> _______________________________________________
> dev mailing list
> dev at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev

More information about the dev mailing list