[Tagging] Data redundancy with "ref" tag on ways vs relations

Frederik Ramm frederik at remote.org
Tue Jul 31 10:45:51 BST 2012


On 07/31/2012 09:31 AM, Paweł Paprota wrote:
>> No. We only create relations when the ref tag is not sufficient. We
>> don't recommend that relations be created for roads otherwise, and
>> anyone doing anything with the data should not expect relations to be
>> there.
> How would you define "sufficient" then?

The "ref" tag is sufficient whenever the membership of a way in a road 
can be correctly modeled by it.

The "ref" tag is insufficient when a way can be part of several roads at 
the same time.

> Is copying same "ref" value to
> thousands of ways in major national (or even long regional) roads a good
> use case for "ref"?

It is an accepted use case, and easier to edit than using relations.

> I think this "freedom of choice" and lack of clear
> guidelines is an illusion in this case since it creates inconsistency.

Consistency is not a goal in itself. This is often mistakenly assumed by 
newcomers to OSM. The magically expect OSM to be the same the world 
over, but ...

> Someone said earlier that OSM is not a technical project but a "social"
> one.

... people are not "consistent" the world over either.

> Well I would agree BUT for the reasons that Petr Morávek describes
> very well in this thread there needs to be a minimal level of consensus
> for tagging stemming from technical reasons - like this situation here
> for major roads.

The amount of consensus there "needs" to be is indeed rather small.

> And it is easy to achieve that - recommend that "the target is that we
> want to have well-maintained relations for major roads in Europe",

We don't.

> then
> motivate people with reporting tools and it works - Poland now is headed
> for a nicely cleaned up road network because people are trying to get
> the report "all green" and also are enjoying that they are doing
> meaningful work by remapping/fixing major roads which as it turned out
> from reporting were in not-so-good state.

I don't know where you get your idea that having a relation for every 
major road is "nice" and "cleaned up". Is it possible that you are using 
your tool to impose your personal view on a community that doesn't know 
any better?

> I'm not sure why some people get very defensive when someone suggests
> some basic (and reasonable - I think) guidelines in order to improve the
> data quality.

I don't know about others, but personally I often see a certain pattern. 
Newcomer - usually with an IT background and just a couple months OSM 
experience, usually with very little or even no "out in the field" 
mapping hours - barges in and tells us how OSM is in a bad shape and 
needs to be fixed. Person has not even bothered to think about the whole 
picture, person does not even know how relations came to be invented, 
how they are used and for what; person believes that on the strength of 
their superior intellect and their database modelling fu alone they know 
what's good for OSM. Words like "consistency" are often part of the 
argument and the importance of "consistency" is never questioned or its 
disadvantages understood, and often a very specific use case is 
mentioned or seen as the most important (e.g. "to make OSM usable for 
routing, the following must be done").

Your show, until now, falls mostly into this pattern (with the exception 
that before you decided to fix the data, you had decided to fix 
usability, which is also something that we often see). The fact that you 
have only been on these lists for four weeks doesn't automatically mean 
that you're wrong with everything you say, but if you had a little more 
experience then you would see more things for yourself and would require 
less talking to.

You need to understand that OSM in Poland is not the same as OSM in 
other countries, and does not necessarily have to be. You also need to 
understand that OSM data is edited by humans who will make mistakes, and 
that OSM tries to attract ever less-technical people to be mappers so 
the more complex your data the more likely it is to break.

(The typical newbie response to this is "well then the editors and API 
must make sure my data model cannot be broken. And yay, before you know 
it you're not only doing usability work and reshaping the data model, 
you're also working on three editors and a rails backend.)

> Of course communities can go on and maintain stuff their
> way but going for unified approach within EU for example could foster
> collaboration between countries - like now German people are actively
> fixing road network in Poland and using the reports. I see this going to
> the next level when mappers from EU countries start to build relations
> for Euro routes

This is actually something that I view with scepticism. The route 
relations specifically, but also the idea that mappers export their 
mapping style elsewhere.

OSM is mostly a project driven by local knowledge. A European mapper has 
no business imposing his personal style of relation mapping to countries 
he hasn't even been to.

> and look at the data in different countries. How is this
> bad?

Because countries with many mappers, like Germany, would thereby 
overrule smaller communities in other countries. But what works well for 
Germany does not necessarily work well for other places.

> The recommendation of using relations in this case is just to kick
> off the whole thing and define some base line for collaboration

The base line is that relations are an optional extra and not a required 
ingredient for roads mapping. Anything you want to "kick off" over and 
above that should be strictly in your local community.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"

More information about the Tagging mailing list