[OSM-talk] [tagging] Road crossings proposal - status?

[Dave Stubbs]

On Mon, May 12, 2008 at 5:21 PM, Ben Laenen <benlaenen at gmail.com> wrote:
> On Monday 12 May 2008, Andy Allan wrote:
>  > Oh, I can think of a way. Yep, I can definitely think of some
>  > shorthand tags for the most common crossing types. Trouble is, as
>  > soon as I mention it, everyone starts uncontrollably ranting.
>
>  But that's the problem right? That no-one outside the UK understands the
>  tags without carefully looking them up. Suppose some language in a
>  strange country calls oneway residential roads which allows bicycles to
>  go both ways a "penguinroad", would you also support a
>  tag "highway=penguin"?
>


Guys, there's several ways to deal with this:

1) Put "How do I tag a Toucan crossing?" on the wiki and expect users
to type in about 4 tags to define each crossing. Similar for "How do I
tag a Penguin Crossing?". Tell all those lazy mappers to develop some
work ethic dammit.

2) Add presets to the editors to supply a "Toucan Crossing" option for
users in the UK which correctly adds the tags then reinterprets them
and presents them as a toucan crossing. Make sure people in the
Antarctic see "Penguin Crossing" instead. There's about 3 editors in
constant use at the moment... I'm sure patches are welcome.

3) Use crossing=toucan, crossing=penguin and have the data consumers
who are interested know to look for either. Make them look for the 4
other tags too if they are provided instead. Document on the wiki what
each one actually means to make it easier for people writing these
applications.

4) Write yourself a tag normaliser script that preprocesses OSM tags
to come out with some normal form that you define. For bonus marks
make this feature recogniser use a user definable transform file.
Update a reference transform regularly from documentation on the wiki,
or find some cool way of keeping this updated. This way the normalised
OSM file can be used by any data consumer without having to edit that
consumer for different crossing names.

5) Convince the OSM community of the need for One True Tagset, and
install the feature recogniser in the API with the ultimate tagging
system preloaded.


At the moment the anti-toucan camp is mostly proposing 1). With a
little bit of 2) and 5) thrown in with from what I see very little
intention of actually doing anything.
The pro-toucan camp is partially implementing 3) for as much as they care to do.

Some of us really couldn't care less either way. Frankly, please stop
talking about it -- you're not getting anywhere.

Dave