Re: [Tagging] Are different definitions for same key/value OK? – was: Re: Is tracktype=grade1 surface=compacted a valid combination?

2022-09-26 Thread stevea
On Sep 26, 2022, at 5:14 PM, Georg  wrote:
> Dear all,
> stevea wrote Mon Sep 26 2022 01:36:26 GMT+0200
> 
 Is tracktype=grade1 surface=compacted a valid combination?
>>> 
>>> while the EN wiki page https://wiki.openstreetmap.org/wiki/Key:tracktype
>>> does not explicitly exclude it [...] the DE wiki page >> 
>>> https://wiki.openstreetmap.org/wiki/DE:Key:tracktype tells (translated
>>> by me) "waterproof surface" and thus explicitly excludes the combination.
> 
>> As I said earlier, and as I've found in OSM across different countries / 
>> continents, there do seem to be and will seem to be "regional variations" 
>> like this.  The wiki using words like "usually" might encourage this, but it 
>> also "encompasses" it, as we're not very likely to get "perfection" with 
>> tracktype=* grades across the whole world.  Best practice seems to be 
>> denoting this in our respective wikis, which we appear to be on track doing 
>> so now.
> 
> noting regional or language specific variances of a definition in the
> wiki is OK for manual mapping, but
(and wrote a great deal more, which I appreciate, have read and here politely 
redact most of it).
> 
> Hence, I speak up for and strongly prefer to limit variances of
> definitions as much as possible, e.g. where simply not at all applicable
> because of local law.
> 
> How do you see it?

The short version of how I see it is "there will be regional variations" and 
"updating what regions know and do about this in our wiki is at least something 
(a good start, anyway)."

I'm certainly open and very much in a listening mode to hear how others think 
we might address the specifics of this (and related tags where similar 
"syntactic smearing" takes place due to regional differences).  I also believe 
that acknowledging that "we are not likely to achieve perfection" and 
"perfection is the enemy of the good" are useful talking points for us to agree 
upon.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


[Tagging] Are different definitions for same key/value OK? – was: Re: Is tracktype=grade1 surface=compacted a valid combination?

2022-09-26 Thread Georg

Dear all,

stevea wrote Mon Sep 26 2022 01:36:26 GMT+0200


Is tracktype=grade1 surface=compacted a valid combination?


while the EN wiki page https://wiki.openstreetmap.org/wiki/Key:tracktype
does not explicitly exclude it [...] the DE wiki page >> 
https://wiki.openstreetmap.org/wiki/DE:Key:tracktype tells (translated
by me) "waterproof surface" and thus explicitly excludes the combination.



As I said earlier, and as I've found in OSM across different countries / continents, there do seem to be and will seem 
to be "regional variations" like this.  The wiki using words like "usually" might encourage this, 
but it also "encompasses" it, as we're not very likely to get "perfection" with tracktype=* grades 
across the whole world.  Best practice seems to be denoting this in our respective wikis, which we appear to be on 
track doing so now.


noting regional or language specific variances of a definition in the
wiki is OK for manual mapping, but
1) causes quite a lot of wiki maintenance effort: Every language needs
to contain all variances of all other languages because we can't expect
e.g. all US citizens to understand French, German and Italian
sufficiently well when they are touring Europe and driving the 3 hours
from Euro-Airport (in France) through German speaking part of
Switzerland to Italy.
2) makes it quite hard to create programs that work correctly and behave
user friendly. Example below.
3) discussions are much more prone to misunderstandings if the same
key/value is defined in different ways, like we recently saw for SAC
scale in scramble topic (may or must a T5 path contain scramble sections?).
4) because we simply don't have the resources to implement 1+2 really
100%, we end up with data that does not match one of the definitions.
And cleanup is really difficult, because you cannot tell which data was
entered due to which definition. Example below.


Example for 2): A seasoned US mapper travels central Europe and maps
there using Vespucci. The mapper's mobile is still in English, so
Vespucci cannot simply rely on "it's enough if German GUI covers German
variance". Instead, it  would now have to contain a piece of logic
telling that temporarily, the English preset texts need to be changed to
match German content (so all regional variations need to be existing in
all preset languages). Moreover, the mapper shall be made aware of this
switch, because a seasoned mapper won't read the texts all the time but
instead directly click on a value out of routine. Things get worse if
the mapper has still pending changes in US and jumps between changes in
DE and US, as Vespucci needs to switch presets depending on which edit
is shown to the user. Same applies to all other editors, all quality
assurance tools, all routers, many analysis tools,...

Example for 4): A highway=path is tagged as SAC T5. Does it contain a
scrambling section as required by SAC, or does it not, because the EN
variant of https://wiki.openstreetmap.org/wiki/Key:sac_scale did mention
for years that scrambling is optional for T5, so someone may have tagged
the path as T5 despite it does not contain a scrambling section, i.e. it
shall be tagged as T4 only. As a consequence, we'd need to re-check the
tagging for thousands of paths but we have AFAIK no effective means to
document which paths have been checked (e.g. MapRoulette does not help
much here).

Hence, I speak up for and strongly prefer to limit variances of
definitions as much as possible, e.g. where simply not at all applicable
because of local law.

How do you see it?

Best regards,
Georg

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging