[OSM-talk] Tagging Scheme Recommendations: highway=path, footway, trail?
Andy Allan
gravitystorm at gmail.com
Tue Aug 31 11:16:03 BST 2010
On Mon, Aug 30, 2010 at 2:53 PM, Steve Bennett <stevagewp at gmail.com> wrote:
> On Sat, Aug 28, 2010 at 2:40 AM, Andy Allan <gravitystorm at gmail.com> wrote:
>> Then mark the reasons it's not suitable. We have this same discussion
>> with cycling (in fact, Peter Miller had an entire presentation on this
>> issue at SOTM09 - he just suggested the wrong solution :-) ). One
>> persons "unsuitable for motorcycles" is another person's fun and
>> games. So if the problem is that there are steps, then mark the steps.
>> If the problem is that there's a massive chasm with a log over it,
>> then mark bridge=yes width=0.25m surface=log maxweight=150kg (or
>> similar!). Mark the stepping stones as stepping stones.
>>
>> In short, mark the facts that lead you to think it's not suitable, and
>> leave the judgement to the producers of the map as to what they think
>> is appropriate for their particular audience.
>
> This solution sounds appealing, but is totally impractical. Recording
> the information you cite is orders of magnitude more work than
> recording a simple yes/no.
It might be harder, but it's also better. Recording the attributes of
the path is the right thing to do, making sweeping generalisations
whilst mapping is easy but wrong.
> Moreover, even with all the information you suggest tagging, I
> honestly don't even know what the end user would do with it all.
> Something somewhere has to boil it down to a yes/no. Your GPS isn't
> going to deal with it, so the logic has to be up stream. By far the
> best person to make a judgment call is the person who mapped it.
Ah, see here's the issue. The best person to judge the features of the
real-world situation is certainly the mapper on the ground. The
absolute worst person to judge how that data is going to be used is
... the mapper on the ground. When they are walking along a path and
want to add information to OSM they have no idea if it's going to be
used for a cycling map, for a route planner, for a wheelchair, for a
horse rider, for a forestry worker, for a firefighter, for a local
council official, for a birdwatcher, or for whoever. They have
absolutely no idea what the data will be used for, and more
importantly, they probably don't have the expert knowledge needed to
assess the criteria for most of these activities.
So the mapper should stick to the stuff only they can do - describe
what's there in front of them. All the decision making should be done
by the people processing the data, since they much better know what
it's going to be used for. And the people doing the processing can
have different takes on what makes a path "unrideable" - which
combinations of features would make that path no passable to the
people the product is intended for - and then end users can find a
particular variation that suits them well. For example, if we only
mark "cycling=suitable" and forget about the width tags, how would I
make bikes-with-kiddie-trailers-maps.org? In order to make new,
interesting - and unexpected - outputs, we need to reign in any
assumptions about what the data will be used for, and stick to the
facts.
Cheers,
Andy
More information about the talk
mailing list