[OSM-talk] Path rendering in the cycleway

[robin paulson]

Christoph Eckert wrote:
> Hi,
> 
>> I think there are two messages here, firstly we should be really careful
>> about voting for new features in the core tagging list unless they are
>> strictly necessary,
> 
> I think we shouldn't vote on tags at all. Instead, we should monitor what gets 
> used most by the mappers (see Tagwatch and the tool announced by  
> Schuyler Erle).

one of the problems with this, is that it's highly likely two mappers 
will develop two contrasting, but both valid methods of mapping the same 
object, and use them liberally. then some other mappers will follow one 
way, and some other mappers will follow the other way. then we have a 
jumbled data model.

not only does this make things difficult for the renderers, as gerv 
said, but also for data analysts. hopefully, this data will be used for 
something more than routing between places and drawing nice coloured 
maps. i would like to think that geographers, social scientists, 
environmental scientists, historians, all could make use of the data by 
dragging out useful stats. when there are a number of ways of mapping 
the same thing, it becomes increasingly difficult to keep up with what 
relates to what. give everybody a simple list of consistent, 
unambiguous, non-overlapping data types, and the data becomes so much 
simpler to use

i don't have much confidence that most people will use tagwatch to find 
out if a tagging scheme already exists. even with the present, partly 
centralised system, people are creating 
duplicated/overlapping/inconsistent schemes