On 7 Sep 2009, at 13:41, Jason Cunningham wrote:
In the recent discussion about Anerley bus stops, we spoke about
several problems. I've been out and looked at most of the stops in
that area and have made some changes.
For the moment, until agreement is reached I've decided not to edit
tags with are labelled or start with "naptan" (except the verified
tag)
I did change tags that I consider to be general OSM tags used to
'map what's on the ground'. The following link shows an example of
how one bus stop can differ from the Naptan Data
http://tinypic.com/m/5lb29w/2
Name: Some were wrong or needed altering. I changed the name to what
was on the bus stop, but left the Naptan:CommonName tag untouched.
One issue is agreeing on how to separate two lines of names. (see
the picture) I used the following eg [Anerley Road / Croydon Road].
Local_ref: I changed this to the reference letter on the bus stop.
The tag Naptan:Indicator holds the Naptan version.
There are errors in tags starting with Naptan and it seems we
shouldnt change these, but instead create a new tag/note with the
different data. This looks like a good idea if it makes it easier
for Naptan to compare data, without dealing with the history of the
node.
All this makes sense and yes, I will stop editing the NaPTAN field
itself. If the bearing is wrong I will create a bearing field with the
correct value. If it is a marked location and not a customary stop as
indicated in NaPTAN then I will tag is as shelter=yes, or pole=yes as
appropriate. If it is marked as being there in NaPTAN but is not there
on the ground then I will check with a local or a bus driver and mark
it as customary=yes or as 'physically_present=no' as appropriate.
Should we add a tag about out investigations, for example:
'note:customary=asked a local and they said is way 07Sep09'
Possibly we should agree the whether the physical name on the pole
goes in name field or in alt-name field.
Regards,
Peter
Jason Cunningham
user:jamicu
2009/9/7 Chris Hill <[email protected]>
I thought we had agreed not to change the NaPTAN fields until an
update
process had been agreed. My only exceptions to this are to remove the
naptan:verified=no field (not a real NaPTAN field) and to move the
stop
to a location that is more accurate, since the location rendered on
maps
benefits from being correct. I record differences in a note= tag, and
record any discrepancies in a separate list.
Any stops that are missing I add physically_present=no and remove the
highway=bus_stop tag (as per the wiki), add a note and update my list.
If the name, bearing etc are wrong I add this to the note and record
it. If we want a name field ('The Range' rather than 'B&Q') we have a
name field to use, or invent stop_name rather than adjusting the
NaPTAN
fields. When we reload this data in the future we don't yet know what
basis the matching might be on, since there are discrepancies with the
AtcoCode we may need other fields too.
I have sent my first set of comments (~25% of the stops) to Hull
council's transport team for their comments and analysis.
Cheers, Chris
Peter Miller wrote:
> On 7 Sep 2009, at 09:46, Ed Loach wrote:
>
>
>> Is it worth questioning the stop names? The ones named "B&Q" in
>> Clacton for example are outside their old site (now "The Range"),
as
>> B&Q moved to new premises about 5 years ago. I've not yet been to
>> the B&Q stops to see if they have the name on (some stops around
>> here do, some don't - and these are two I've not yet checked).
>>
>
> One of the motivations for adding NaPTAN data to OSM was to get a
new
> feedback path for differences. There are of course licensing
> restrictions about using data from OSM to improve a (c) Crown
dataset,
> but the list of places where there is a difference of opinion is of
> considerable use.
>
> Also, OSM aims to be better that official data so we need to make
such
> that is the case and that we can lead with data accuracy, and not
just
> wait for the official data to be correct which might take some time
> (in my county I have issues I reported 3 months ago that are still
not
> resolved).
>
> As such I think we need to agree on how we deal with discrepences.
>
> How about the following:
>
> Situation: One finds at stop in reality that does not exist in
NaPTAN.
> Response: Add it to OSM with highway=bus_stop, shelter=yes/no etc,
> note='not in NaPTAN 7 Sept 2009'
>
> Situation: NaPTAN says it is a marked bus stops but there is no sign
> of a pole
> Response: removed highway=bus_stop tag and add a note, for example
> note=no sigh of this one on the ground'
>
> Situation: NaPTAN says it is customary, but actually it is marked
> Response: Should we have a tag stop_type which can reflect the
NaPTAN
> types, ie customary, marked, hail and ride etc? For now I have been
> changing the naptan tag
>
> Situation: The bearing doesn't match the road and the description
> Response: For now I have been changing the bearing in NaPTAN field,
> but I think we need a proper bearing field. Possibly we should add a
> bearing field and populate it if the is different
>
> Situation: The name of the flag is not the name in NaPTAN
> Response: at it as alt_name
>
> more?....
>
>
> We can then run a script from time to time to compare the current
> state of the two datasets and create reports for either party of the
> discrepancies.
>
>
>
> Regards,
>
>
> Peter Miller
>
>
>
>
>
>> Ed
>>
>>
>>
>> _______________________________________________
>> Talk-transit mailing list
>> [email protected]
>> http://lists.openstreetmap.org/listinfo/talk-transit
>>
>
>
> _______________________________________________
> Talk-transit mailing list
> [email protected]
> http://lists.openstreetmap.org/listinfo/talk-transit
>
>
_______________________________________________
Talk-transit mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/talk-transit
_______________________________________________
Talk-transit mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/talk-transit
_______________________________________________
Talk-transit mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/talk-transit
