[OSM-talk] [tagging] Feature Proposal - RFC - incline up down

Roy Wallace waldo000000 at gmail.com
Sat Aug 22 23:38:42 BST 2009


On Sat, Aug 22, 2009 at 8:35 PM, Morten Kjeldgaard<mok at bioxray.au.dk> wrote:
>
>> "hard-to-verify data" - I don't see why incline=* is any harder to
>> verify than ele=* - as you said yourself, if you have one you can
>> calculate/verify the other...
>
> The fact that there's a lot of unreliable and hard-to-verify data is no good
> argument for adding more.

What? The key question is if a tag is verifiable. Incline=* is just as
verifiable as ele=*. It's just in a different form. The "good
argument" for adding incline=* is that it is 1) easy to read off a
sign (say, source:incline=sign), 2) provides valuable information in
the meantime, while we wait for you to develop and import your ele=*
solution.

> Incline tagging is useless unless a consumer of this data can count on it
> being generally available. A driver might find herself on a steep incline
> when expecting a flat one, just because it wasn't tagged.

This is ridiculous. The absence of incline=* does not infer incline=0
- it infers that the incline is unknown/unspecified. Just as absence
of ele=* doesn't infer ele=0 - it infers that the elevation is
unknown/unspecified.

> IMHO it is much more productive to spend time working out some system that
> would allow us to compute inclines automatically on the whole dataset. That
> would give you the desired data all over the world and not just in your
> local area.

Sounds great - but in the meantime, people will continue to tag what
they see on the ground - especially on road signs - in any way they
see fit. Better that incline=* is used consistently for tagging
incline, in the meantime.




More information about the talk mailing list