[Talk-us] Proposal: Sunset ref=* on ways in, favor of relations

Kevin Kenny kkenny2 at nycap.rr.com
Mon Nov 9 14:39:50 UTC 2015

On 11/08/2015 09:04 PM, Paul Norman wrote:
> Phil's demo was an excellent proof of concept of pictorial shields 
> from route relations, but isn't something that can be reasonably 
> incorporated into a stylesheet as-is. 
> https://github.com/gravitystorm/openstreetmap-carto/issues/596 is the 
> OpenStreetMap Carto issue for shields from relations, 
> https://github.com/gravitystorm/openstreetmap-carto/issues/508 is the 
> one for pictorial shields.

Forgive my ignorance. I'm coming in as a consumer of the data and as a 
fairly inexperienced mapper. My biggest field mapping project so far has 
been http://www.openstreetmap.org/relation/4286650 - which pales to 
insignificance compared with the whole cities that some of you take on.

Nevertheless, I found that project interestingly related to the current 
discussion, because the trail in question is concurrent with named roads 
on portions of its route, and there are therefore ways that are 
intentionally not tagged with its name, lest, for instance, a highway 
router become confused. Having the relation available was the only ready 
way that I had to extract a single center line from the OSM data, which 
was then used to produce http://www.nptrail.org/?page_id=59 . It would 
have been challenging to assemble the ways into something coherent for 
measuring the distances without having the route relation.

In the issue on Github, you remark:
"I'm contemplating closing this. As above, there are currently serious 
data issues in the US that prevent route relations from being used for 
rendering shields. If someone wants to take up the data quality issue, I 
could leave it open or re-open in the future, but unless it changes, I 
just can't see any way to make use of the route relations."

That gives a horribly bleak outlook. I'm sure I'm misunderstanding what 
you say there, but it comes across as saying that the problem is forever 
unfixable, and that the hard work people have put in so far into 
generating route relations is all for naught, either because it remains 
incomplete or because the data model of route relations is fundamentally 
incompatible with shield rendering.

I do tend to think that, in the long run, deriving the shield from both 
the tagging on the way and on the relation is going to be problematic, 
simply because data maintained in two places are always going to be 
inconsistent. Maintaining the data exclusively on the way is going to be 
brittle, because of the issues of concurrent routes. Maintaining the 
data exclusively on the relation is going perhaps to be too difficult 
for novice mappers.

The result is that the data quality issues of which you speak will be 
always with us.

Is this an accurate summary of your position?

If so, I'd like to explore what has to change to break the log jam. My 
OSM experience has been limited to putting data in and taking data out, 
but I'm a pretty fair hand at SQL and GIS processing - your long query 
doesn't scare me, if the issue is programming. If the issue is that data 
are entered inconsistently, Maproulette has often been successful in the 
past at getting the data cleaned up. If the issue is fundamentally 
political, well, I'm probably not the man for that one. In John Adams's 
words, I'm "obnoxious and disliked."

73 de ke9tv/2, Kevin

More information about the Talk-us mailing list