[Geodata] wiki-nature geodata aggregators

[Andrew Turner]

David William Bitner wrote:
> On Tue, Jul 28, 2009 at 9:25 AM, Arnulf Christl 
> <arnulf.christl at wheregroup.com <mailto:arnulf.christl at wheregroup.com>> 
> wrote:
>
>     > I'll also add that just because a geodata set has a stamp or
>     > "authority", and some apparent quality assurance process, does not
>     > mean that the resulting data is any better, or the process any more
>     > full proof, than the many eyes approach of OSM.
>
>     True. At the same time for some it does make a difference whether an
>     authority of some kind stamps information or not. OSGeo may well (be)
>     develop(ed) into a kind of authority because it is not ...
>
>
> Often times when liability is concerned sometimes -- often more 
> wrongly than right -- it is necessary to use an "official" version of 
> something merely for the reason that you can then pass-the-buck if 
> there are problems.  Along the same lines, when you are doing work 
> that could end up in a courtroom or more simply end up getting 
> questioned or audited or .... it is necessary to build off something 
> that is at least a snapshot such that those who are 
> questioning/auditing/whatevering you can go back and recreate your 
> work from the same version of the data that you had used.
>
> As with source code, versioning, having snapshots, sometimes having a 
> branch away from trunk that gets released (think a snapshot of OSM at 
> a certain time that has been deemed by a local authority as acceptable 
> for planning work and tagged as such) can be very valuable things that 
> can help to bridge the gap between the "wilds" of crowd sourcing and 
> the need for control of "official" datasets.  Along with this,  as an 
> entity considers tagging a blessed version for whatever purpose, as 
> the trunk continues to change, having good "diff" tools becomes 
> important for the authorities such that they can monitor what is 
> changing and can respond accordingly (accepting the edits, going back 
> in and saying "no, we surveyed this with super new fangled gadget and 
> know that it is right" etc).
>    
There has been some good brainstorming on this idea at a few recent 
conferences and sessions. An entity could "sign" a snapshot/region/way. 
Calculate the MD5 hash, apply your GPG public key, and you now have an 
'official', or accepted, version of data. This is similar to how Git 
(and other DVCS tools) work.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/geodata/attachments/20090728/d725787a/attachment.html