[Tagging] Barrier=berm

[Andy Townsend]

On 28/11/2019 09:59, Paul Allen wrote:
>
>
> On Thu, 28 Nov 2019 at 00:23, Graeme Fitzpatrick 
> <graemefitz1 at gmail.com <mailto:graemefitz1 at gmail.com>> wrote:
>
>
>     Question of my own - is there any particular reason that a berm
>     couldn't just be rendered the same as a wall?
>
>
> That question prompts another question.
>
> Why render it as a wall?  Since a berm is a type of embankment, why 
> not render it as an
> embankment?

I've rendered "2-sided embankments" at e.g. 
https://map.atownsend.org.uk/maps/map/map.html#zoom=17&lat=54.024306&lon=-1.02141 
as _their own thing_ for a while now.  That example is at the end of a 
firing range, 
https://map.atownsend.org.uk/maps/map/map.html#zoom=15&lat=53.87324&lon=-1.23959 
shows some for flood defences.  In that second image to the northwest 
you can see an embankment with a path on it.

OSM Carto, as far as I'm aware

  * Doesn't have a concept of 2-sided embankments or a rendering for them
  * Doesn't have the concept of "on an embankment" being a modifier for
    highways / railways in a similar way to "bridge" and "tunnel"


> Either way, if you render it the same as an existing object, and it 
> serves the
> same purpose as an existing object, the carto people are likely to 
> veto it under their "no
> synonyms" rule.

I wouldn't argue that a new tag for berm is "needed", because people 
have found ways to tag these features already (such as "embankment=yes" 
and various flood defence tags), but it could be argued that using one 
tag for these features makes things clearer.  Separately to that, even 
if you say "render X like Y" it doesn't mean that X is a synonym of Y.  
There are plenty of those in all renderings already.


> Even if you persuade the carto people to render berms, it will go on 
> their long to-do list and
> may take a long time to appear.

The usual answer here is "pull requests welcome"...

Sometimes people might not want to do that because they know it wouldn't 
be accepted (if it makes more use of lua processing than the OSM Carto 
folks are happy with, for example).  I suspect that wouldn't make sense 
to submit a pull request to OSM Carto for bus guideway handling to match 
the way I do it because it'd depend on lua changing a bus guideway to be 
a type of railway.  That "busway as railway" handling is why

https://map.atownsend.org.uk/maps/map/map.html#zoom=17&lat=52.315008&lon=-0.05611

show a bridge but

https://www.openstreetmap.org/#map=18/52.31481/-0.05622

does not.


>
> You also have the problem of having to inspect a lot of existing 
> embankments to see
> if some of them should be retagged as berms.  And the problem of 
> mappers, perhaps
> newbies, wondering what the difference between the two is.
>
> Others have proposed a berm=* subtag to differentiate types of 
> berms.   Why not, instead,
> use a subtag for embankments?

A two-sided embankment is fundamentally different to a one-sided one.  
The renderer would need to split them out into a different feature 
anyway to render them, so it's fairly irrelevant to it how they are 
tagged (other than the extra complication in "select" statements if 
people insist on sub-tagging features that mean something else).

Best Regards,

Andy


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20191128/7a70a890/attachment-0001.html>