[OSGeo-Discuss] Integration test bed

[Mark Leslie]

Rajs comment about structure struck me as dead on.  I would expect the 
test bed to be much more engaging if we developed some actual use 
cases.  This would require an early commitment from participants, but 
would help define what we expect of them and give us a list of people we 
can harass to make things happen. This does, of course, make it a whole 
lot more work for everyone involved.
Mark

Cameron Shorter wrote:
> I was very excited to hear about the Integration test bed concept and 
> hope that any setbacks can be taken as a learning experience and help 
> us put some achievable steps in place to make it more successful at 
> future events - be they FOSS4G or conferences with an OSGeo presence.
>
> Key take home points for me:
> 1. There should be a standard dataset that all applications use as 
> part of their default demos. I've seen this to a small degree between 
> Geoserver, Mapbuilder and Openlayers. There are sure to be others and 
> I'm sure we can do a better job.
> 2. An interoperability testbed is difficult to set up just before a 
> conference. So lets do it gradually over a year or so as we build our 
> software, open source workshops and presentations.
> 3. A version dependency tree will greatly ease installation which in 
> turn will help open source uptake.
>
> Jody Garnett wrote:
>> On problem was expecting vendors to supply their own hardware for 
>> something that was a conference theme. If you made several machines 
>> available prior to the conference for the vendors to configure with 
>> their software you may get some traction. Vendors are very busy 
>> getting their own act together; fitting into a conference theme (like 
>> showcasing interoperability) is an extra that is hard to plan for.
>>
>> For some projects (like MapServer) we did not have a specific vendor 
>> or organization to pester into setting up an example service.
>>
>> An integration showcase would be a great message for the OSGeo booth 
>> however. Making machine available for community members to set up and 
>> configure would perhaps be a better approach.
>>
>> Jody
>>
>> Tim Bowden wrote:
>>> I believe the integration test bed didn't go quite as well as planned.
>>> Any thoughts as to why, and what could be done next time to improve the
>>> experience?
>>>
>>> Regards,
>>> Tim Bowden
>>>
>>> _______________________________________________
>>> Discuss mailing list
>>> Discuss at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/discuss
>>>   
>>
>> _______________________________________________
>> Discuss mailing list
>> Discuss at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/discuss
>>
>
>