Render junction=roundabout like oneway=yes

[sb12]

According to the wiki junction=roundabout implies oneway=yes.
Therefore, I propose to render all highways with junction=roundabout and oneway!=no in the same way as oneway ways.

[matthijsmelissen]

See also #465 (the code there solves part of the problem here).

Not sure if we want to implement this, though. Opinions?

[matthijsmelissen]

@pnorman @gravitystorm What do you think?

[matthijsmelissen]

Any opinions?

[pnorman]

I'm skeptical about "implied" values, particularly solely based on wiki documentation. Leaving that aside, something to watch for is oneway arrows appearing at relatively low zooms where the roundabout itself is quite small, you do not want this. The arrows only add value when the zoom is high enough to see within the roundabout

[dieterdreist]

2014-05-28 2:59 GMT+02:00 Paul Norman notifications@github.com:

I'm skeptical about "implied" values, particularly solely based on wiki
documentation.

I think for roundabouts it is broadly accepted that the tag
junction=roundabout conveys also the oneway-implication. This is btw. also
how it works with traffic signs (roundabouts are all oneways, but there is
never a oneway-sign, at least not in the ones I know of). This is not valid
for all circular roads of course, but for true roundabouts it should be.

You see this reflected also in the current tools like JOSM validator etc.
not only in the wiki, and I know from discussions on the local ML that
people actively remove oneway=yes tags from roundabouts.

Leaving that aside, something to watch for is oneway arrows appearing at
relatively low zooms where the roundabout itself is quite small, you do not
want this. The arrows only add value when the zoom is high enough to see
within the roundabout

yes, but this is a different issue.

[matkoniecz]

Developers of iD editor consider that there is "assumed default of oneway=yes on things like junction=roundabout and highway=motorway"

See openstreetmap/iD#2220

See https://github.com/bhousel/iD/blob/master/js/id/core/oneway_tags.js for set of tags that according to iD editor implies onewayness.

[Rovastar]

I think it is safe to assume they are one way.
if on the crazy edge case counter example they are not they can be tagged one-way= no and it would not render

[matthijsmelissen]

So we seem to agree that roundabouts should be assumed to be oneway. Next question: does it make sense to render arrows on them?

[chrisfleming]

From a map point of view, I don't think it's necessary to render arrows on them, as most people will see a roundabout and know that it's one way. The only exception would be larger cases where it might not be so obvious if something is a roundabout or a 2 way road.

The other advantage would be that having the arrows on roundabouts would point people to fix cases where the roundabout direction is wrong.

[dieterdreist]

2014-05-29 13:01 GMT+02:00 Chris Fleming notifications@github.com:

From a map point of view, I don't think it's necessary to render arrows on
them, as most people will see a roundabout and know that it's one way. The
only exception would be larger cases where it might not be so obvious if
something is a roundabout or a 2 way road.

you cannot see from the rendering if something is a roundabout or not,
because we do not render any icon or style particular to roundabouts. You
will see a circle road and can only guess it could be a roundabout.

[gravitystorm]

I'm in favour of oneway arrows on roundabouts (those without oneway=no, of course), as per the original request.

[matthijsmelissen]

See also https://trac.openstreetmap.org/ticket/4130,

[Rovastar]

I was not clear in my above comments but I am in favor too.