Re: [Talk-GB] Use of OS OpenData in OSM

[Ed Avis]

Jon Stockill  writes:

>I've been adding 
>buildings from streetview, then going out to survey for addresses. If 
>the buildings get removed it'd better be done in a way that preserves 
>the address data, or we lose the results of a lot of surveying.

You could check with the legal mailing list, but I believe that the addresses
might also need to be removed, since they are still in some sense derived from
the OS data (or at least their position on the map is).

However, I'd expect that if the proposed relicensing goes ahead, some
compromise or rationalization will be found that lets the OS data continue
to be used.

-- 
Ed Avis 


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Ed Avis]

Graham Jones  writes:

>If I collect a GPS trace of a road that is tagged as say 'source=NPE', I will
>adjust the road to match the trace, and change it to source=survey.

I would put source=NPE;survey but nobody really expects the source tag to be
a bomb-proof, legally sound way to determine the source of an object.  It is a
hint at best.  Looking at changeset history is better (if people documented
what they were doing in the changeset comment, and the edits don't predate the
introduction of changesets to the API).  So I think it's okay in practice;
the contents of the source tag doesn't make any real difference to the map
or to its legal status.

-- 
Ed Avis 


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Ian Spencer]

I don't think that is fair enough if you are going to be legally 
pedantic about it. The basic track was derived from NPE and you have 
adjusted it, by implication you have taken into account the original 
work in two ways: as a validation that the GPS trace relates to this 
item, and also you have got an independent verification that the GPS 
trace is at least in the right ball park. To remove the "pollution" of 
the sourcing, you would need to delete the path and then start anew.


The mass of problems around licensing are exactly due to the pedantry 
necessary to be legally unencumbered. It is an unattainable nirvana - it 
is a fair bet that there are a fair chunk of footpaths in OSM that are 
effectively derived from current OS 25k maps where people may be 
surveying, but have used the OS map to check that their traces are not 
corrupt, or simply have the OS map by their side while they edit. The 
best OSM can hope for is that there is a GPS trail that vaguely matches 
map features to give plausible deniability, and that other "local 
knowledge" type sourcing can be shown to be properly surveyed if someone 
decided to challenge - which leaves surveyors with a burden of keeping 
notes and other evidence.


Given that it is an impossible task to be clean in a volunteer project 
without imposing some rule like "no edit without GPS or signature in 
blood" it would then in turn make sense to be far more pragmatic with 
regard to licensing of OS OpenData than seems to be the case with some 
hard liners.


In terms of goals, it does not make sense for OSM simply to be the sum 
of existing open source data, OSM will only be of worth if it produces a 
product (or products) superior to what is available, or unique in some 
way. The various national cycle maps that are evolving are an example of 
where OSM is producing value that is not available elsewhere. That is 
where the surveying input comes into its own.


Spenny

Graham Jones wrote on 23/07/2010 15:18:
If I collect a GPS trace of a road that is tagged as say 'source=NPE', 
I will adjust the road to match the trace, and change it to 
source=survey.
That means that unless you look through the history there will be no 
evidence that it was once derived from another source...but, once you 
have surveyed it, I don't think it is derived from the other source 
any more, so this is fair enough?


Graham.

On 23 July 2010 14:42, Chris Fleming > wrote:


On 22/07/10 16:25, Richard Fairhurst wrote:

Ed Avis wrote:

As an aside, I think the 'source' tag is a bit
misconceived; it would make
much more sense to tag source on the changeset, not on
each object it
touches.

Only if you solely use one source per changeset. I'll
typically use at least
a mix of NPE, OS OpenData, GPS survey and personal knowledge,
and sometimes
more.

I tend to do the same - although if I have a track for a road that
was previously  source = "not survey" I will generally modify it
to match the tracks and either delete the source tag or edit it to
be source=survey

Although I don't think I'm consistent.  What do people tend to do?


One I did recently is
http://www.openstreetmap.org/browse/way/4981553/history and shows
the evolution of what is initially a traced name = FIXME into a
fully surveyed way by 4 people over nearly 3 years :)

Although this is a good case of where an area appears done and so
I didn't visit it, until the the OS comparsion stuff came out. At
which point I've discovered lots of missing stuff.

Cheers
Chris





-- 
e: m...@chrisfleming.org 

w: www.chrisfleming.org 



___
Talk-GB mailing list
Talk-GB@openstreetmap.org 
http://lists.openstreetmap.org/listinfo/talk-gb




--
Dr. Graham Jones
Hartlepool, UK
email: grahamjones...@gmail.com 


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb
   


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[SomeoneElse]

On 23/07/2010 15:18, Graham Jones wrote:
If I collect a GPS trace of a road that is tagged as say 'source=NPE', 
I will adjust the road to match the trace, and change it to 
source=survey.
That means that unless you look through the history there will be no 
evidence that it was once derived from another source...but, once you 
have surveyed it, I don't think it is derived from the other source 
any more, so this is fair enough?


It may be, if (say) the name came from NPE and you forgot to write the 
name down while you were there collecting the GPS trace; in that case 
"source=gps;source:name=NPE" might be appropriate.


There are quite a few "countryside" roads near me that look NPE traced 
(i.e. the offset from other GPS traces is consistent with the NPE 
Potlatch overlay offset) but have no source assigned; I've added a 
source of something "looks like NPE" or similar when updating other 
attributes but not the position (i.e. I see that somethings a dirt track 
from one end but don't go down it to collect a trace).




___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Graham Jones]

If I collect a GPS trace of a road that is tagged as say 'source=NPE', I
will adjust the road to match the trace, and change it to source=survey.
That means that unless you look through the history there will be no
evidence that it was once derived from another source...but, once you have
surveyed it, I don't think it is derived from the other source any more, so
this is fair enough?

Graham.

On 23 July 2010 14:42, Chris Fleming  wrote:

> On 22/07/10 16:25, Richard Fairhurst wrote:
>
>> Ed Avis wrote:
>>
>>
>>> As an aside, I think the 'source' tag is a bit misconceived; it would
>>> make
>>> much more sense to tag source on the changeset, not on each object it
>>> touches.
>>>
>>>
>> Only if you solely use one source per changeset. I'll typically use at
>> least
>> a mix of NPE, OS OpenData, GPS survey and personal knowledge, and
>> sometimes
>> more.
>>
>>
> I tend to do the same - although if I have a track for a road that was
> previously  source = "not survey" I will generally modify it to match the
> tracks and either delete the source tag or edit it to be source=survey
>
> Although I don't think I'm consistent.  What do people tend to do?
>
>
> One I did recently is
> http://www.openstreetmap.org/browse/way/4981553/history and shows the
> evolution of what is initially a traced name = FIXME into a fully surveyed
> way by 4 people over nearly 3 years :)
>
> Although this is a good case of where an area appears done and so I didn't
> visit it, until the the OS comparsion stuff came out. At which point I've
> discovered lots of missing stuff.
>
> Cheers
> Chris
>
>
>
>
>
> --
> e: m...@chrisfleming.org
> w: www.chrisfleming.org
>
>
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-gb
>



-- 
Dr. Graham Jones
Hartlepool, UK
email: grahamjones...@gmail.com
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Chris Fleming]

On 22/07/10 16:25, Richard Fairhurst wrote:

Ed Avis wrote:
   

As an aside, I think the 'source' tag is a bit misconceived; it would make
much more sense to tag source on the changeset, not on each object it
touches.
 

Only if you solely use one source per changeset. I'll typically use at least
a mix of NPE, OS OpenData, GPS survey and personal knowledge, and sometimes
more.
   
I tend to do the same - although if I have a track for a road that was 
previously  source = "not survey" I will generally modify it to match 
the tracks and either delete the source tag or edit it to be source=survey


Although I don't think I'm consistent.  What do people tend to do?


One I did recently is 
http://www.openstreetmap.org/browse/way/4981553/history and shows the 
evolution of what is initially a traced name = FIXME into a fully 
surveyed way by 4 people over nearly 3 years :)


Although this is a good case of where an area appears done and so I 
didn't visit it, until the the OS comparsion stuff came out. At which 
point I've discovered lots of missing stuff.


Cheers
Chris





--
e: m...@chrisfleming.org
w: www.chrisfleming.org


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[SomeoneElse]

On 23/07/2010 08:52, Ed Avis wrote:
Fair point. In that case, to be scrupulous, you would need to add 
individual

source tags to each object as you change it.  But even then, the object data is
not sufficient to know where it has come from: you must check the change history
and see at which point the various sources were added, and what other changes
were made at the same time.

   
... or "source:blah" to indicate where the "blah" tag came from at the 
time that the "blah" tag was added or changed.




___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Ed Avis]

Richard Fairhurst  writes:

>>As an aside, I think the 'source' tag is a bit misconceived; it would make 
>>much more sense to tag source on the changeset, not on each object it 
>>touches.
>
>Only if you solely use one source per changeset. I'll typically use at least
>a mix of NPE, OS OpenData, GPS survey and personal knowledge, and sometimes
>more.

Fair point.  In that case, to be scrupulous, you would need to add individual
source tags to each object as you change it.  But even then, the object data is
not sufficient to know where it has come from: you must check the change history
and see at which point the various sources were added, and what other changes
were made at the same time.

-- 
Ed Avis 


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Graham Jones]

I agree with what Kevin said earlier - that the OS OpenData is too good a
resource to ignore, so from my perspective the acceptability of any new
licence depends on making sure that the OS OpenData derived data stays in
the OSM database (either by persuading ourselves that the new licence is
compatible with the OS licence, or coming to some agreement with OS).

If that is not the case I can see us ending up with a few different
OpenStreetMaps - the new licence one (with extra data removed), a UK one
with OS OpenData, an Austrailian one with NearMap data etc.which would
be a mess!

Graham.

On 22 July 2010 16:50, Jon Stockill  wrote:

> Kevin Peat wrote:
>
> In 6 months time the OS data will be so entrenched in the UK map that we
>> could never strip it out in any useful way without vast amounts of fixup
>> being required and I can't imagine many people being interested in doing
>> that. So for me at least whatever license we change to must be OS
>> compatible.
>>
>
> The area around me is starting to get quite blue too - I've been adding
> buildings from streetview, then going out to survey for addresses. If the
> buildings get removed it'd better be done in a way that preserves the
> address data, or we lose the results of a lot of surveying.
>
> Jon
>
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-gb
>



-- 
Dr. Graham Jones
Hartlepool, UK
email: grahamjones...@gmail.com
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Jon Stockill]

Kevin Peat wrote:

In 6 months time the OS data will be so entrenched in the UK map that we 
could never strip it out in any useful way without vast amounts of fixup 
being required and I can't imagine many people being interested in doing 
that. So for me at least whatever license we change to must be OS 
compatible.


The area around me is starting to get quite blue too - I've been adding 
buildings from streetview, then going out to survey for addresses. If 
the buildings get removed it'd better be done in a way that preserves 
the address data, or we lose the results of a lot of surveying.


Jon

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Richard Fairhurst]

Ed Avis wrote:
> As an aside, I think the 'source' tag is a bit misconceived; it would make 
> much more sense to tag source on the changeset, not on each object it 
> touches.

Only if you solely use one source per changeset. I'll typically use at least
a mix of NPE, OS OpenData, GPS survey and personal knowledge, and sometimes
more.

cheers
Richard
-- 
View this message in context: 
http://gis.638310.n2.nabble.com/Use-of-OS-OpenData-in-OSM-tp5314389p5325850.html
Sent from the Great Britain mailing list archive at Nabble.com.

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Ed Avis]

As an aside, I think the 'source' tag is a bit misconceived; it would make much
more sense to tag source on the changeset, not on each object it touches.
The best way to see what data comes from OS or elsewhere is to look at the
history of an object and see which changeset added the 'name' tag or whatever.
It's a shame that we have no standardized way of tagging changesets so this
could be automated - though I hope that most people would note the sources used
in their comment.

-- 
Ed Avis 


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Kevin Peat]

Nice work, but as the OS data is a good dataset and compatible with our
current license why would anyone be surprised that people are using it. I've
uploaded woods and waterways for my area so it looks pretty blue but the
streets were surveyed on the ground and I would think that might be the same
for quite a few areas. This sort of data (woods, streams, etc.) is a bit
less core to the map than areas where people are filling in missing streets
from OS data.

In 6 months time the OS data will be so entrenched in the UK map that we
could never strip it out in any useful way without vast amounts of fixup
being required and I can't imagine many people being interested in doing
that. So for me at least whatever license we change to must be OS
compatible.

Kevin








On 21 July 2010 22:57, Graham Jones  wrote:

> Hi,
>
> I have tidied up my OS Opendata Map (
> http://www.maps.webhop.net/osm_opendata).
>
> The changes are:
>
>- Lines and dots are smaller so it looks less of a mess.
>- It excludes source tags containing '25k', 'os7' and 'photos', which
>were giving quite a lot of false positives, especially in Scotland.  Let me
>know if you see any others and I can exclude them.
>- I have left my original layer available as 'tiles1', but this is not
>displayed by default - you can add it with the '+' control to see the
>differences.
>- The about  page
>has been updated to describe how it works better (still crude, but more
>complicated SQL!).
>
> There are still some surprising things here - for example National Cycle
> Route 1 is highlighted, even though I know that the bits I added are not
> from OS Opendata (see the bit from Whitby to Sunderland 
> here).
>   It seems that someone has tagged the relation (Relation Number 9579) with
> 'OS_OpenData_StreetView' - I don't know why they would have done this?
>
> Regards
>
> Graham.
>
>
> On 20 July 2010 23:40, Graham Jones  wrote:
>
>> Thank you all for your comments.
>> I'll not get into the licence change debate here - plenty of that on
>> osm-talk
>>
>> -  I agree that there are a few surprises highlighted here.   There are a
>> couple of cycle tracks highlighted that I survryed myself, so I will have to
>> check the underlying data.  When I get home I will improve the filtering to
>> exclude os 1:25k references.
>> - I will see what I can do with the rendering as Gregory suggests.
>> - The supermarkets reference is copy-and-paste-itis on my behalf - sorry!
>> - Emilie is probably right that strictly I should be interested in
>> history, but I cant do that easily from a planet extract, and I don't think
>> it will matter too much with opendata being so recent.   A curious legal
>> point is that if a way was originally derived from os-opendata, but
>> subsequently re-surveyed, is it still derived from opendata?
>>
>> Graham
>>
>> 
>> Graham Jones
>> (from my phone)
>>
>> On Jul 20, 2010 4:41 PM, "80n" <80n...@gmail.com> wrote:
>>
>> On Tue, Jul 20, 2010 at 10:02 AM, Robert Whittaker (OSM) <
>> robert.whittaker+...@gmail.com > wrote:
>> > >...
>> What's more, because Produced Works can be published under a restrictive
>> license we couldn't get the additional data back by tracing either.  ODbL +
>> CT makes getting data back into OSM much harder than it is now by a massive
>> degree.
>>
>> BTW, how would a corporation agree to the Contributor Terms anyway?  The
>> sign-up page only caters for individuals.  Has, for example, CloudMade,
>> agreed to the contributor terms yet and how could we tell if they had?
>>
>> 80n
>>
>>
>>
>> ___
>> Talk-GB mailing list
>> Talk-GB@openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/talk-gb
>>
>>
>
>
> --
> Dr. Graham Jones
> Hartlepool, UK
> email: grahamjones...@gmail.com
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-gb
>
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Ed Loach]

I haven’t checked too closely, but assumed Norfolk was down to
parish boundaries or something similar, with the areas being shaded.

 

Ed

 

From: talk-gb-boun...@openstreetmap.org
[mailto:talk-gb-boun...@openstreetmap.org] On Behalf Of Gregory
Williams
Sent: 22 July 2010 09:19
To: Talk-GB@openstreetmap.org
Subject: Re: [Talk-GB] Use of OS OpenData in OSM

 

That looks much better now. It’s certainly interesting to see that
whole counties, like Norfolk, appear to be sourced from OS OpenData.
I know that that isn’t true, but I guess it’s a just a side-effect
of the last edit to existing ways adding a source / source:name tag
referencing OS, like adding the name to an existing way. It may be
interesting to differentiate between source and source:name using
different colours to get an idea of where it’s only the name that’s
been added.

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Gregory Williams]

That looks much better now. It's certainly interesting to see that whole
counties, like Norfolk, appear to be sourced from OS OpenData. I know that
that isn't true, but I guess it's a just a side-effect of the last edit to
existing ways adding a source / source:name tag referencing OS, like adding
the name to an existing way. It may be interesting to differentiate between
source and source:name using different colours to get an idea of where it's
only the name that's been added.

 

From: talk-gb-boun...@openstreetmap.org
[mailto:talk-gb-boun...@openstreetmap.org] On Behalf Of Graham Jones
Sent: 21 July 2010 22:58
To: 80n
Cc: Talk-GB@openstreetmap.org
Subject: Re: [Talk-GB] Use of OS OpenData in OSM

 

Hi,

 

I have tidied up my OS Opendata Map
(http://www.maps.webhop.net/osm_opendata).

 

The changes are:

*   Lines and dots are smaller so it looks less of a mess.
*   It excludes source tags containing '25k', 'os7' and 'photos', which
were giving quite a lot of false positives, especially in Scotland.  Let me
know if you see any others and I can exclude them.
*   I have left my original layer available as 'tiles1', but this is not
displayed by default - you can add it with the '+' control to see the
differences.
*   The about <http://www.maps.webhop.net/osm_opendata/about.html>  page
has been updated to describe how it works better (still crude, but more
complicated SQL!).

There are still some surprising things here - for example National Cycle
Route 1 is highlighted, even though I know that the bits I added are not
from OS Opendata (see the bit from Whitby to Sunderland here
<http://www.maps.webhop.net/osm_opendata/?zoom=10&lat=54.6778&lon=-1.37818&l
ayers=BFT> ).   It seems that someone has tagged the relation (Relation
Number 9579) with 'OS_OpenData_StreetView' - I don't know why they would
have done this? 

 

Regards

 

Graham.

 

On 20 July 2010 23:40, Graham Jones  wrote:

Thank you all for your comments.
I'll not get into the licence change debate here - plenty of that on
osm-talk

-  I agree that there are a few surprises highlighted here.   There are a
couple of cycle tracks highlighted that I survryed myself, so I will have to
check the underlying data.  When I get home I will improve the filtering to
exclude os 1:25k references.
- I will see what I can do with the rendering as Gregory suggests.
- The supermarkets reference is copy-and-paste-itis on my behalf - sorry!
- Emilie is probably right that strictly I should be interested in history,
but I cant do that easily from a planet extract, and I don't think it will
matter too much with opendata being so recent.   A curious legal point is
that if a way was originally derived from os-opendata, but subsequently
re-surveyed, is it still derived from opendata?

Graham


Graham Jones
(from my phone)

On Jul 20, 2010 4:41 PM, "80n" <80n...@gmail.com> wrote:

On Tue, Jul 20, 2010 at 10:02 AM, Robert Whittaker (OSM)
mailto:robert.whittaker%2b...@gmail.com> >
wrote: > >...

What's more, because Produced Works can be published under a restrictive
license we couldn't get the additional data back by tracing either.  ODbL +
CT makes getting data back into OSM much harder than it is now by a massive
degree.

BTW, how would a corporation agree to the Contributor Terms anyway?  The
sign-up page only caters for individuals.  Has, for example, CloudMade,
agreed to the contributor terms yet and how could we tell if they had?

80n

 

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb




-- 
Dr. Graham Jones
Hartlepool, UK
email: grahamjones...@gmail.com

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Graham Jones]

Hi,

I have tidied up my OS Opendata Map (http://www.maps.webhop.net/osm_opendata
).

The changes are:

   - Lines and dots are smaller so it looks less of a mess.
   - It excludes source tags containing '25k', 'os7' and 'photos', which
   were giving quite a lot of false positives, especially in Scotland.  Let me
   know if you see any others and I can exclude them.
   - I have left my original layer available as 'tiles1', but this is not
   displayed by default - you can add it with the '+' control to see the
   differences.
   - The about  page has
   been updated to describe how it works better (still crude, but more
   complicated SQL!).

There are still some surprising things here - for example National Cycle
Route 1 is highlighted, even though I know that the bits I added are not
from OS Opendata (see the bit from Whitby to Sunderland
here).
  It seems that someone has tagged the relation (Relation Number 9579) with
'OS_OpenData_StreetView' - I don't know why they would have done this?

Regards

Graham.


On 20 July 2010 23:40, Graham Jones  wrote:

> Thank you all for your comments.
> I'll not get into the licence change debate here - plenty of that on
> osm-talk
>
> -  I agree that there are a few surprises highlighted here.   There are a
> couple of cycle tracks highlighted that I survryed myself, so I will have to
> check the underlying data.  When I get home I will improve the filtering to
> exclude os 1:25k references.
> - I will see what I can do with the rendering as Gregory suggests.
> - The supermarkets reference is copy-and-paste-itis on my behalf - sorry!
> - Emilie is probably right that strictly I should be interested in history,
> but I cant do that easily from a planet extract, and I don't think it will
> matter too much with opendata being so recent.   A curious legal point is
> that if a way was originally derived from os-opendata, but subsequently
> re-surveyed, is it still derived from opendata?
>
> Graham
>
> 
> Graham Jones
> (from my phone)
>
> On Jul 20, 2010 4:41 PM, "80n" <80n...@gmail.com> wrote:
>
> On Tue, Jul 20, 2010 at 10:02 AM, Robert Whittaker (OSM) <
> robert.whittaker+...@gmail.com > wrote:
> > >...
> What's more, because Produced Works can be published under a restrictive
> license we couldn't get the additional data back by tracing either.  ODbL +
> CT makes getting data back into OSM much harder than it is now by a massive
> degree.
>
> BTW, how would a corporation agree to the Contributor Terms anyway?  The
> sign-up page only caters for individuals.  Has, for example, CloudMade,
> agreed to the contributor terms yet and how could we tell if they had?
>
> 80n
>
>
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-gb
>
>


-- 
Dr. Graham Jones
Hartlepool, UK
email: grahamjones...@gmail.com
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Graham Jones]

Thank you all for your comments.
I'll not get into the licence change debate here - plenty of that on
osm-talk

-  I agree that there are a few surprises highlighted here.   There are a
couple of cycle tracks highlighted that I survryed myself, so I will have to
check the underlying data.  When I get home I will improve the filtering to
exclude os 1:25k references.
- I will see what I can do with the rendering as Gregory suggests.
- The supermarkets reference is copy-and-paste-itis on my behalf - sorry!
- Emilie is probably right that strictly I should be interested in history,
but I cant do that easily from a planet extract, and I don't think it will
matter too much with opendata being so recent.   A curious legal point is
that if a way was originally derived from os-opendata, but subsequently
re-surveyed, is it still derived from opendata?

Graham


Graham Jones
(from my phone)

On Jul 20, 2010 4:41 PM, "80n" <80n...@gmail.com> wrote:

On Tue, Jul 20, 2010 at 10:02 AM, Robert Whittaker (OSM) <
robert.whittaker+...@gmail.com > wrote: >
>...
What's more, because Produced Works can be published under a restrictive
license we couldn't get the additional data back by tracing either.  ODbL +
CT makes getting data back into OSM much harder than it is now by a massive
degree.

BTW, how would a corporation agree to the Contributor Terms anyway?  The
sign-up page only caters for individuals.  Has, for example, CloudMade,
agreed to the contributor terms yet and how could we tell if they had?

80n



___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[80n]

On Tue, Jul 20, 2010 at 10:02 AM, Robert Whittaker (OSM) <
robert.whittaker+...@gmail.com > wrote:

> Emilie Laffray  wrote:
> > The second point is that I don't see the relation between knowing how
> much
> > OS OpenData and the switch to the new licence. Talks of losing data is
> > partially a self fulfilling prophecy. It is impossible right now to gauge
> > how much data IF ANY we would lose since we don't have any means to know
> who
> > is in support of what until the voluntary licence is put in place.
>
> The last I heard (albeit informally) from OS is that they're worried
> about the lack of formal attribution requirements on Produced Works
> under ODbL. It's also unclear whether they would agree to DbCL for
> individual data items. From the first point alone, I don't think it's
> safe to conclude that their current license would allow their data to
> be used in an ODbL database, although I'd be hopeful that they could
> be persuaded to agree to a license that allows this.
>
> More importantly though, the current contributor terms [1] (in
> particular clauses 2 and 3) require mappers to grant certain rights to
> their data to OSMF, which in particular would allow OSMF to re-license
> the data without the Share-Alike or Attribution requirements if that's
> what the community voted for. I can't see OS ever agreeing to this, as
> it would mean they could loose their attribution requirements.
>
> So, unless those contributor terms are amended / removed, or there's
> an exception for certain data providers, we will have to loose any OS
> OpenData derived information from the database, by either deleting
> objects or reverting them to a prior state. Until this issue is
> resolved, I'd suggest not investing any time in adding anything
> further derived from OS OpenData.
>
> (Personally, I think those clauses in the contributor terms need to be
> removed entirely, as (a) it's unreasonable to expect people to agree
> to the use of their data under an as-yet-unspecified license, and (b)
> the terms provide a loop-hole that would prevent us from benefiting
> from the Share-Alike provisions of ODbL -- A third-party can take our
> data, do some cool stuff with it, add some of their own data, and
> release the result under ODbL to fulfil their SA responsibilities.
> There's nothing to force them to agree to our contributor terms, and
> by not doing so, they'd prevent OSM from using their combined dataset.
> This renders the SA provisions all but useless for us.)
>
> What's more, because Produced Works can be published under a restrictive
license we couldn't get the additional data back by tracing either.  ODbL +
CT makes getting data back into OSM much harder than it is now by a massive
degree.

BTW, how would a corporation agree to the Contributor Terms anyway?  The
sign-up page only caters for individuals.  Has, for example, CloudMade,
agreed to the contributor terms yet and how could we tell if they had?

80n
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Robert Whittaker (OSM)]

Emilie Laffray  wrote:
> The second point is that I don't see the relation between knowing how much
> OS OpenData and the switch to the new licence. Talks of losing data is
> partially a self fulfilling prophecy. It is impossible right now to gauge
> how much data IF ANY we would lose since we don't have any means to know who
> is in support of what until the voluntary licence is put in place.

The last I heard (albeit informally) from OS is that they're worried
about the lack of formal attribution requirements on Produced Works
under ODbL. It's also unclear whether they would agree to DbCL for
individual data items. From the first point alone, I don't think it's
safe to conclude that their current license would allow their data to
be used in an ODbL database, although I'd be hopeful that they could
be persuaded to agree to a license that allows this.

More importantly though, the current contributor terms [1] (in
particular clauses 2 and 3) require mappers to grant certain rights to
their data to OSMF, which in particular would allow OSMF to re-license
the data without the Share-Alike or Attribution requirements if that's
what the community voted for. I can't see OS ever agreeing to this, as
it would mean they could loose their attribution requirements.

So, unless those contributor terms are amended / removed, or there's
an exception for certain data providers, we will have to loose any OS
OpenData derived information from the database, by either deleting
objects or reverting them to a prior state. Until this issue is
resolved, I'd suggest not investing any time in adding anything
further derived from OS OpenData.

(Personally, I think those clauses in the contributor terms need to be
removed entirely, as (a) it's unreasonable to expect people to agree
to the use of their data under an as-yet-unspecified license, and (b)
the terms provide a loop-hole that would prevent us from benefiting
from the Share-Alike provisions of ODbL -- A third-party can take our
data, do some cool stuff with it, add some of their own data, and
release the result under ODbL to fulfil their SA responsibilities.
There's nothing to force them to agree to our contributor terms, and
by not doing so, they'd prevent OSM from using their combined dataset.
This renders the SA provisions all but useless for us.)

Robert.

[1] http://www.osmfoundation.org/wiki/License/Contributor_Terms

-- 
Robert Whittaker

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Use of OS OpenData in OSM

[Emilie Laffray]

On 19 July 2010 23:37, Graham Jones  wrote:

> Hi Folks,
> Given all of the talk on the osm-talk mailing list about the possibility of
> losing data if we move to the new licence, I started to wonder just how
> widespread OS OpenData use is in OSM.  I couldn't find a visualisation, so I
> made one this evening which is visible at
> http://www.maps.webhop.net/osm_opendata/.   Not all zoom levels are
> rendered yet - it could easily take most of tomorrow to finish the higher
> zoom levels.
>
> The way I have detected the OS derived data is very crude - basically if
> 'source' or 'source:name' matches %os%, the element gets highlighted in
> blue.  This means that tags relating to out of copyright OS maps might also
> be highlighted (I see some 'OS7' tags).
>
> Anyway, I thought you might be interested in seeing how widespread the use
> is - I was very surprised that at low zoom levels, the UK (or rather GB) is
> covered in blue, but if you zoom in enough you can start to see which towns
> are 'old mapping' and which contain a lot of tracing from StreetView.
>
> I'm not sure what to do with this now - I will work out some statistics for
> how much of the OSM data includes OS opendata.   If anyone can think of a
> use for it now that I have the database imported, please let me know!
>
>
Well the first thing to say is that the approach is a bit flawed. You would
need to look also at the history to find out what is new (completely OS
OpenData) or what has been completed with data from OS. It is very likely
that some might come from "old mapping".
The second point is that I don't see the relation between knowing how much
OS OpenData and the switch to the new licence. Talks of losing data is
partially a self fulfilling prophecy. It is impossible right now to gauge
how much data IF ANY we would lose since we don't have any means to know who
is in support of what until the voluntary licence is put in place.

Emilie Laffray
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb