[Talk-GB] OpenKent, OSM coverage estimation

[TimSC]

On 07/06/11 13:34, Richard Fairhurst wrote:
>
> If you go out and find the exact position of a service, with a piece of
> paper and a pen (or a GPS or whatever), that's your data, not theirs. So of
> course it's compatible.
>    
I think you miss my point. The datasets contain more than just their 
postal address. If the licenses are compatible, we can mash up the data.

> And you should do that anyway.
This implies I don't already, which is a false. (Otherwise, why are you 
telling me I should?)

> OSM is meant to be a crowd-sourced,
>    
This is a meaningless statement in my way of thinking. Even if it was 
meant to be something, by some one, at some stage proves nothing. Just 
because it has been crowd sourced to some extent doesn't preclude other 
approaches. Some types of data are in OSM that are almost impossible to 
survey with our crowd sourcing resources. UK streams, for example, 
mostly were not crowd sourced (in terms of surveying).
> constantly updated representation of what's on the ground, not some cheapass
> mirror of any dataset that Government happens to have lying around.
>    
Any what if the government dataset is open and stomps on OSM's attempt? 
(Don't bother saying "improve OSM" because that IS the approach we use 
and still the government set is better, in some cases.) Duplicating 
other open data sets seems a waste of time - as you seem to imply by 
resurveying stuff already available elsewhere. I am not advocating we 
only import data either. A hybrid approach - import AND crowd source - 
is better. If you want crowd sourced surveying only, I suggest you start 
another project.

Regards,

TimSC