[Tagging] importance=* tag (for transportation etc)

Alexander Matheisen AlexanderMatheisen at ish.de
Fri Mar 25 12:19:27 UTC 2016


Am Freitag, den 25.03.2016, 11:26 +0000 schrieb Andy Mabbett:
> On 20 March 2016 at 00:12, Alexander Matheisen
> <AlexanderMatheisen at ish.de> wrote:
> 
> > If you have a look at the highway=* tagging: This scheme is
> > subjective,
> > but there is no alternative.
> 
> Poppycock.

Why?

> > As the person who created that station importance draft, I will
> > focus
> > on stations, but for other features like mountain peaks the
> > situation
> > should be similar: It is not possible to calculate the importance
> > of a
> > station just by some values.
> 
> How do you calculate it, then? Rolling dice?

I do not use any calculation based on measures?! My proposal (seems
that you did not read it) uses a list of characteristic criterias for
each category.

The scheme I propose can be compared to the place=* scheme. The tagging
of places is also a kind of importance ranking. According to the wiki
it is not based on population and it can be decided by the user to get
better results:

"By way of example, the charter city of Alameda in California
 (population 76,000) is tagged as a town, due to its proximity to the 
much larger and better known cities of San Francisco (populations 
805,000) and Oakland (population 309,000)." (
http://wiki.openstreetmap.org/wiki/Tag:place%3Dcity)

Also see this description taken from 
http://wiki.openstreetmap.org/wiki/Tag:place%3Dtown:

"Use place=town to identify an important urban centre that is larger
than a place=village, smaller than a place=city, and not a
place=suburb. Towns normally have a good range of shops and facilities
which are used by people from nearby villages."

This definition also mentions "important", so where is the difference
to the importance proposal? And why is importance=* so problematic from
the view of some mappers, while we are also using a similar scheme for
places?

> > I also see problems in getting some of the proposed values. For
> > example, the amount of passengers or trains per time is difficult
> > to
> > measure for a mapper and is not easy to be checked by other
> > mappers.
> 
> Please explain what measure you are using, that is *more easily
> checked* by other mappers.

If you think that the number of passengers or trains is easy to map in
a larger scale, then please describe how you would map these values.

For example see the number of passengers. In most cases railroad
companies do not publish these information except for some big station.
That means that a mapper would have to go to the station, stand near
the entrance and count how many people go in and out. For larger
stations you need more than one mapper to count this.

And also if you are able to count the number of passengers, there are
more problems: This value depends on time, so it would be necessary to
do this more than once at different times, different days, etc. to
calculate an average value. Do you think this is praticable?

Another suggestion here was the number of trains. This can be done more
easily, and if timetable information in standard formats are available,
this value can be collected automatically. But the problem is, that
this value does not indicate the importance of a station. A suburb
station may have trains departing every 2 minutes while a central
station may have trains departing every 10 minutes. But that does not
indicate that the suburb station is more important.

> > I also see the problem that calculating the importance by a complex
> > algorithm might be very intransparent.
> 
> Please explain what measure you are using, that is *more
> transparent*.

As I said, I do not use any calculation based on measures. My proposal
requires that a mapper classifies a station to one station category. It
is more transparent because everyone can see that a station is tagged
with a certain category. So a user easily understands why e.g. a
station is rendered in a certain zoom level or was recognized as more
important in the ranking of search results.

If you just calculate the importance from a list of measurable values,
you may get good results with a complex algorithm that recognizes many
aspects. But then it is very difficult for a mapper to understand why
station A was ranked more important than station B, and it is also
difficult to influence the ranking if it is wrong. That is what I mean
with transparency.


Regards
Alex
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: This is a digitally signed message part
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20160325/83effa52/attachment.sig>


More information about the Tagging mailing list