[Tagging] Feature Proposal - Voting - lanes General Extension
imagic.osm at gmail.com
Mon Mar 5 13:28:08 GMT 2012
You summarized the problems with proposals very well: actually no-one
cares at all.
* If a proposal is opposed, doesn't matter - it is tagged that way anyway.
* If it is approved, doesn't matter also - it is still tagged somehow
What I have seen in the last months regarding proposals lead me to the
conclusion, that they are a big waste of time. Nonetheless I want to -
at least try to - go the whole way; at least once.
Regarding your comments (which would better fit into the discussion page):
* Left-Right: this is the "renderer problem". You can either use
:left/:right or :forward/:backward to distinguish the driving
directions. Both have some drawbacks. The left-right-approach makes it
very easy for the renderer, but impossible for the router, because the
router misses the information of the driving direction. The
forward-backward-approach makes it easy for the router, but the
renderer doesn't know if the forward-lanes are right or left (and the
same for the backward-lanes). Both problems disappear as soon as the
information about left-hand/right-hand traffic is available.
I decided to go for the forward-backward-approach, because it better
fits my thinking of driving directions on lanes.
* Lanes type: This can easily be achieved with this tagging scheme:
(Please ignore the actual values in this example - it is just an
example). To maintain compatibility it would be better in this case to
I'm already using this scheme to tag the lanes on the streets in my
region and up to now haven't run into any serious problems. As the
fundamental idea behind this scheme is very simple and extendible I
assume it will stay this way.
Thanks for your feedback.
2012/3/5 Tobias Knerr <osm at tobias-knerr.de>:
> Martin Vonwald wrote:
>> The proposal
>> is now open for voting.
> Is voting on this really helpful right now? I believe that nobody can
> honestly know at this point whether the proposal's idea is actually good,
> whether it needs minor corrections, or whether we maybe even should try
> something completely different.
> Personally, I consider it most likely that the general idea is the right way
> to go, but that minor modifications would still improve it. However, I
> cannot tell with any certainty before I had the opportunity to gather some
> experience with it.
> For example, I'm currently experimenting with lane rendering in OSM2World.
> The experiences so far have confirmed some of my previously-held assumptions
> about the usefulness of certain lane tagging schemes, and made me question
> others.¹ Similarly, I expect that actually trying to apply a tagging scheme
> to large parts of my town would also result in some practical insight into
> what works and what doesn't. But that, again, takes time.
> So I cannot vote for or against this proposal today because I couldn't make
> a sufficiently informed decision. The vote might give some insight into how
> mappers think about value-sequence based lane tagging, but probably not much
> about this specific variant thereof - so no matter what the result of this
> vote will be, I will not consider myself bound by it. Hopefully, others will
> think the same way, and remain open for better solutions if they come up in
> the future.
> ¹ I found it very helpful for my use case to have the road's lanes split
> into a "left" and a "right" part. I also relied on the assumption that I
> would be able to clearly determine a lane's "type" (e.g. sidewalk, cycle
> lane, parking lane, bus lane, ...) somehow.
> Tagging mailing list
> Tagging at openstreetmap.org
More information about the Tagging