[Talk-GB] OpenKent, OSM coverage estimation

[TimSC]

On 08/06/11 08:15, Peter Miller wrote:
>
> My experience is that the LWG never makes definitive statements!
>    
I find that annoying sometimes but, if we are to follow to Spinoza's 
example that we should "made a ceaseless effort not to [...] scorn human 
actions, but to understand them", LWG have to deal with legal advice 
that is also not definitive. Hopefully they can offer a definitive 
position on matters such as good mapping practice - like if we should 
import data of uncertain compatibility.

> I suggest that you turn the tables on them and send them an email
> saying that you will import the OGL-licensed data in xx days unless
> you get a statement from them in the mean time saying that it would be
> violating the OSM licensing terms and compromising your status as a
> contributor.
>    
I have set one or two deadlines on LWG in the past but it doesn't fit 
with their working pattern. Until now, nothing gets decided, or is put 
to discussion leading up to a decision, in any forum other than the 
teleconference. But to their credit, they are quite open and 
understanding when you do phone in and discuss matters. This is 
something I want to work on: to have a medium-long term discussion with 
LWG outside the weekly teleconference. I think the suggestion was met 
with a mixed response - discussions will continue. In the modern world 
with email, wikis, face to face, etc, there is more to life than 
teleconferences!

> Regarding data formats. Can I suggest that that we gratefully accept
> data in whatever format it is provided. We can ask politely for it to
> be in an better format but please don't complain either about the
> quality of the data or the suitability of the format which may support
> councils who will argue that they should delay releasing anything
> until they have got it right and in the perfect format. The phrase is
> 'raw data now' (warts and all).
>    
Agreed. If the data is even slightly usable, someone in the community 
can convert it.

I am currently working on a legally gray dataset (which I am not 
importing, obviously) which is currently a mixture of closed data and 
data that a government agency aspires to make open data. They seem to 
lack the urgency or resources to separate the two, so I am doing it for 
them (without them asking) and I will ask nicely if they will release 
"my" data subset (for which they have the copyright).

> On a separate note. Would you be able to do a comparison between place
> names in NatGaz and in OSM. I think we will be surprised how many
> places we are still missing from OSM. My guess is that OSM only
> contains about 65% of the 50K places in that database. Here it is:
> http://data.gov.uk/dataset/nptg
>    
That is an interesting data set. I might use a different approach 
because it seems unlikely the original data contains significant 
errors(?). Currently, I use XAPI to query OSM for objects near to a 
record in the government database. I am not sure if the admins would 
appreciate me hammering the XAPI server with 50K requests! or that might 
be fine... I could use the UK dump, "slice" it to get place=*, import it 
into a separate microcosm server on my laptop, and then do XAPI requests 
to my laptop server. I will have a think.

Regards,

Tim