[OSM-talk] Proliferation of path vs. footway

[Martin Simon]

2009/8/10 Tom Chance <tom at acrewoods.net>:

> It's important for OpenStreetMap to have some coherence.
>
> It's quite important that you and I both agree that "chair" refers to a
> piece of furniture on which we sit. Imagine if you used the word "chair" to
> refer to a small furry pet that meows and likes fish! We can't have a
> situation where - as others have pointed out - we have people using a
> particular tag in many different ways.

Agreed. The problem with path (the one I call a problem, too) is IMHO
that there are people who don't like this tagging scheme or think it's
unneccessary (which is not a problem for me), but instead of just not
using it and staying with footway/cycleway/bridleway, they think
"well, but its there and gets rendered, lets use it for something else
(e.g. "very narrow way in the forest") and change the wiki page".
*zap* - a small furry pet that meows and likes fish. ;-)

> It also helps if we stick to one way of describing any particular thing.
> It's lovely that in England we have "cow shed" and "byre" and many other
> phrases for the same object. But when you're writing a stylesheet for
> Mapnik, or trying to download an extract, or writing a routing algorithm,
> your task is made ten times more difficult if you have to keep adding lots
> of alternative ways of describing the same thing.

Yes, it helps, but it's IMHO better to come up with a new way to
describe something rather than changing the meaning of long-existing,
widely used and important tags...

> We don't need to force anybody to do anything, but here are some basic ways
> in we can encourage a more coherent approach:
>
> - discussions at SOTM or regional meetings
> - a well managed wiki (hah!)
> - stylesheets for Mapnik and Tiles at Home (both a bit out of hand, as Andy
> Allan says)
> - presets in Potlatch and JOSM
> - error checking tools
> - even bots that try to correct very minor errors like s/cahtolic/catholic/

Okay, no problem with that, as long as alternative ways of tagging
like highway=path are not treated as errors.

> I would support removing highway=path from Potlatch and JOSM and the Mapnik
> stylesheet until a wiki page is drawn up which unambiguously describes how
> it should be used. If it duplicates or replaces existing tags, that should
> be properly resolved.

That would be at least one step too far in my eyes. But a better wiki
page would be great.

> There's a big difference, Simon. Nobody had yet accepted any addressing
> schema, none of the community mechanisms I listed above properly supported
> any one approach, until the breakthrough. Now that approach is gradually
> being properly supported. It's a case of the anarchic approach working
> quite well, partly by luck.
>
> In this case, you have a tag which duplicates and possibly replaces
> existing tags; which nobody can agree on the definition for; and which is
> interpreted by different tools in different ways. That's a big step
> backwards.

Simon is actually the last name ;)
Yes, this is a bigger move than the breaktrough of the Karlsruhe
Schema, but for example post_code=12345 was already quite popular, so
the Karlsruhe guys used addr:post_code=12345. In my opinion a good
decision, because it clearly showed that objects tagged this way refer
to the Karlsruhe Schema, which is well documented in the wiki.

We had a similar situation when highway=path started, before different
groups made up thier own definitions, differing from the original
proposal...

-Martin