[OSM-talk] NCN refs - consistency

[Peter Miller]

Umm, lots to think about here.

In Ipswich I have used ncn_ref for national cycle routes, and used a ';'
between entries if there is a shared route, for example: ncn_ref=N1;N51

Sustrans also has regional routes (ncr) but we don't have any in Suffolk to
my knowledge. We do have regional routes, for example 'Suffolk Costal
Route'. I haven't encoded these but proposed using rcn_ref.

I also wanted to encode local cycle routes, and in the absence of any other
suggestions I have tagged these as lcn_ref (lcn_ref=4A, lnc_ref=12 etc).

We have routes where two national cycle routes share a segment with a local
cycle route, and also no-doubt a regional cycle route if I looked carefully.

When making a map it might be useful to be able to distinguish between a
local route and a national route, hence my use of 3 different tags.

To be clear I am very happy to discuss this, come to a conclusion and then
change my tagging, but would like to do it once rather than repeatedly :)

Btw, I was at a meeting at the UK's Departent for Transport discussing an
on-line cycle journey planner. The DfT is very committed to bottom-up
community data gathering by cyclists for cyclists on this one but identified
that that different projects in the country are currently using different
data standards. Ideally they would like their planner built on OS data, but
want to keep the field open at this stage and propose a mosaic approach
based on the same data model but possibly with different data sources and
collection methods.

The DfT is keen to progress this idea and is likely to set up a working
party to define a data schema for the additional information needed for this
purpose. I know many people will be less than ecstatic at the prospect but I
will be involved in this process and will attempt to ensure that OSM has a
voice. Incidentially the working party will be taking input from the Dutch
who are also building a national cycle journey planner which is interesting
given the AND developments.



Regards,




Peter




> ------------------------------
> 
> Message: 10
> Date: Wed, 11 Jul 2007 10:32:40 +0100
> From: David Earl <david at frankieandshadow.com>
> Subject: Re: [OSM-talk] [Talk-GB] NCN refs - consistency
> To: OSM <talk at openstreetmap.org>
> Message-ID: <4694A3B8.6040703 at frankieandshadow.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> On 11/07/2007 09:22, Richard Fairhurst wrote:
> > David Earl wrote:
> >
> >> Sustrans, who developed the network, uses 'NCN' as the abbreviation.
> >
> > Actually, sometimes they use NCR and RCR!
> >
> >
> http://www.sustrans.org.uk/webfiles/general/South_East_Annual_Report_05_06
> _print_version%5B1%5D.pdf
> >
> > I think it's been observed on the talk@ list that ncn_ref is quite an
> > insular tag, and cycle_ref would be better. Adding lcn_ref, too, would
> > seem to exacerbate the problem.
> >
> > But for the time being I just use NR47 and RR47. (Am not very good at
> > remembering to add route=cycle, though... slapped wrists.)
> >
> > The free text "proposed NR41" that Andy cited is also one of mine.
> > This is for a route which will be NR41 but isn't, yet. Any thoughts on
> > a better way to tag this?
> 
> I though you were asking about the tag value, not the tag name.
> 
> There's a problem with the same way being used by multiple routes in
> general, and how to link up with their identifiers. The route tag
> doesn't really work - has been commented on several times on this list.
> I wonder if we should try to clear that up rather than just address NCN.
> 
> Considertations: what if two NCN's share the same way (this occurs in
> Cambridge). General point: what if two or more bus routes share the same
> way, or three bus routes, two ncns and two highway designations all
> belong to the same way. What general mechanism could support this?
> 
> How about the following (don't worry about the actual tag names for now,
> it's the principle I am thinking about):
>    route=id1;id2;id3
> (where id1 etc are *any* strings of the mappers choice) and then
>    ref:id1=11
>    route:id1=ncn
> In this way we can know that there are three routes belonging to this
> way; that the first is a national cycle route and its reference is 11.
> Whether ref:id1=11 or ref:id1=NCN11 or whatever isn't too important:
> consumers can choose to extract only the number from the ref if they want.
> 
> And this is general. Consider my last example:
>    route=a;b;c;d;e;f;g
>    route:a=ncn
>    ref:a=11
>    route:b=ncn
>    ref:b=51
>    route:c=bus
>    ref:c=101 # i.e. route of bus number 101
>    ...
> 
> I'd be inclined to use more meaningful identifiers than a,b etc, but the
> point is they are only used to tie the multiple routes to their
> individual properties. e.g.
>    route=ncn11;ncn51;bus101;bus37
>    ref:ncn11=11
>    route:ncn11=ncn
> etc
> 
> David
> 
> 
> 
>