[OSM-talk] Good practice, and should we rely on defaults?

[Mateusz Konieczny]

Apr 6, 2022, 20:22 by gdt at lexort.com:

>
> Yves via talk <talk at openstreetmap.org> writes:
>
>> I wonder if using defaults from an external database (let say we have
>> it) and analyse it with our boundaries is really more efficient than
>> tagging everything?
>>
>
> I don't understand how you can think that.   Simplifying, there are two
> paths forward:
>
>  A) Don't produce a file that maps regions to defaults and promulgate its
>  use.  If a property is not tagged, apply some global default from the
>  wiki.  Encourage everything to tag every property always.
>
>  B) Produce the file and use it.
>
> To talk about efficiency, we need to estimate both the total amount of
> effort, and the result in terms of accuracy of rendered information.  It
> seems obvious that B leads to much better results with less total
> effort.  Producing the file probably only takes 1000 hours and using it
> per rendered another 100 or so.  Tagging every road, done honorably,
> would take a very long time.
>
The problem is: what about places differing from assumed default?
And where applying such defaults produces incorrect data?

Without systematic tagging of info, including "obvious" values it
is much harder to systematically map and catch such cases.

In addition: note that there are more options than
"tag explicitly" vs "use defaults"

There is at least also 
"tag explicitly, if nothing is tagged guess based on reasonable assumptions".

This way data consumers have something (and most use this strategy,
see roads without oneway tag set) but things can be systematically mapped.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20220407/5484e09a/attachment.htm>