[Talk-transit] local_ref problem around Anerley in NAPTAN
richard.mann.westoxford at googlemail.com
Wed Sep 2 17:07:34 BST 2009
I'm not sure I like the idea of a "2" way on top of a shorter "2a" and "2b"
way; hence my instinctive preference for nodes in the complicated
situations. It can also be unclear where one subplatform starts and ends
(especially where the split does't reflect a signalling berth, as is common
in Germany, for instance). However, I can't really see the harm in using
ways in the simple situation, and equally the full platform face name could
be a way, but I'd make subplatform locations nodes rather than ways.
The length of a platform could equally well be recorded as a length= tag on
a node; the info is on the NR website if you know where to look - and have
permission to use it.
And the fact that it may not _yet_ render is - ahem - not relevant.
On Wed, Sep 2, 2009 at 4:42 PM, Peter Miller <peter.miller at itoworld.com>wrote:
> On 2 Sep 2009, at 16:27, Richard Mann wrote:
> 2) I don't like the idea of ways for platforms, except possibly for the
> limited case where you've got one platform on each side. It's just not
> extendable. They should be areas. Sublettering for parts of platforms should
> probably be on nodes, representing the point on the platform that's the
> midpoint for boarding a train that stops at that platform (it will be in the
> timetable system as "2a", and a notional router ought to direct you to that
> point). If a platform is split into 2a and 2b, you probably need three nodes
> - 2a/2b and 2 (for trains that take up the full length).
> Personally I find linear ways pretty satisfactory for platforms, which
> often have no more width than a footpath after all (which are also tagged as
> linear features)/ Possibly we should use areas for larger platforms (ie the
> paved/tarmac area) with highway=pedestrian;area=yes and then add
> railway=platform ways to the edges of the area as required. Sub platforms
> can also be linear ways for their actual extent (I don't like using nodes
> for sub-platforms because they do have an extent which can be measured and
> is sometimes be important). For a platform that serves two tracks, one of
> either side then an area should be used with the two different sides having
> appropriate linear 'platforms' associated with them. I am not sure how to
> represent a set of steps coming down to a point in the middle of an area
> though. One reason to use linear ways for now is because we already have the
> tools to build, render and route models that use them. Areas are fine with
> side accesses, but not top and bottom accesses.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-transit