[OSM-talk] Road Widths, Stubs and Priority/Giveway.

[Interlug]

On 1/10/07, Ben Robbins <ben_robbins_ at hotmail.com> wrote:

>
> 1)
>
> Road widths:  This has been discussed beforeish, Im not really fussed how
> its tagged, but any suggestion would be appreiciated, otherwise I shall tag
> them as highwaytype=type1/2/3/4.*.   This is a way of splitting up the
> varing width variations of the small roads that are otherwise tagged the
> same (unclassified).
>
> *Additional to highway=unclassified, (and streetlight=yes, catseyes=yes if
> nessesery)
>
> 1=Road with Room for Normal viecles going in each direction to drive past
> each other normally
> 2=Roads where cars can pass but will usually slow down, and/or move to one
> side of the road.  Usually they lack road markings in the UK
> 3=Roads where 1 of the cars would need to drive up onto to grass/sand/mud on
> the side of the road
> 4=Roads where the road edge stops easy passing and 1 of the 2 cars would
> have to reverse to a passing spot.  Walls/enbanked edges/ditches/sinking
> sand etcetc along the edge.

Is this new?
Is this the same thing that you were talking about earlier?  I thought
you were talking about additional classifications for tracks only, now
I see unclassified mentioned above.  Did I miss that the first time
around?

If you're talking about tracks
highway="motorway" variations can be classified with the existing
lanes="#" tag.  Can lanes= and surface= give you what you seek?  Does
highway="track", lanes="0.5", surface="grass and gravel" tell you that
passing requires caution?

Why 1/2/3/4 ?
If these tracks can't be distinguished to your satisfaction with
highway="track", lanes="0.5", surface="grass and gravel" or similar
why would you choose to classify them with a code number (1/2/3/4)
rather than something more descriptive?  I don't know what you would
consider acceptable for a descriptive value but how about
"pass_with_caution" "use_passing_areas" or
"if_there_is_oncoming_traffic_you're_finished".

On a personal note, Ben.  I would support your proposal if it made
sense to me but I just don't understand it.  It is not a matter of
saying to myself "I don't think I'd use it, so I won't vote for it."
I don't understand what it is for.  When I try to imagine how I would
use your proposal above, I can not think of a group of public roads
that need to be distinguished in the way that I imagine you are
suggesting.

Is this proposal intended to distinguish between different tracks on
private property?  To show differing amounts of use on a farm or
something?  Help me understand.

> 2)
>
> Stubs**:  I've seen proposals for similar things, and this must have been
> discussed before, just becuase it seems such an obvious flaw.   When there
> is a road or anything that is known to be there, but is unmapped it would be
> very helpful to stick a symbol there saying that the road end on the map is
> not the road end in reality.  I'm not suggesting it should render on the
> main map renders, but either show up in an editor, or be rendered on a notes
> render that can be referencesd by people travelling somewhere and wishing to
> map.  Stubs could varie for what they mark.  footpath stubs/road stubs
> etcetc.  Any ideas?

Agreed.  I'd like to see something for this, too.  I recall earlier
discussions suggesting that a way could be marked with three closely
spaced nodes at the end to show that it continues ...  I also recall
that somebody suggested "tbc"for "to be continued" or "here be
dragons" for areas that are unexplored.

I think that rendering these areas with a "no entry" sign is
confusing.  We do expect to see "no entry" signs as we travel and they
mean "Don't come this way."  Your suggestion would change the meaning
to "Nobody else has come this way, but you should, if you are
curious".    How about a sign with a questionmark on it, or a dragon
(for "here be dragons") or a cat walking off a cliff (for "curiousity
killed the cat") ?

If we are marking "unexplored territory" we should have a way to mark
nodes, ways and areas that need further attention from our mapping
colleagues, or as reminders to ourselves.

> 3)
>
> If there is a cross road, then usually (not in towns) 1 of the roads would
> continue, while the other would giveway to it.  I'm currently concerned that
> a map using osm data would keep saying to go straight on when its
> unnesesery.   If the way goes straight threw the junction then that would be
> ok, but usually If a road is 100 miles long (for example), then it wouldnt
> just be 1 way, for editing convenience, and for lack of confusion with a
> great varaty of tags not applying to the whole way.
> My solution would be to be able to layble statuslayers for a road.  So if
> there is a cross road and 2 roads are '2', and 2 are '1', then it is clear
> that the 2's give way to the 1's.  Alternatively, if there is a  '1', '1',
> '1', '1', then its clear that all give way to each other.  The same for '1',
> '2', '3', '4'.  I'm shore theres better solutions...would anyone care to
> explain/think of one?.

You are talking about right of way at an interchange, aren't you?  I
think that this topic is similar to turn restrictions in that both
topics are valuable to routing software and both strike me as more
complicated the more I look at them.  In my limited experience adding
this sort of priority at the way-level won't work, it must be
considered at the intersection level.

Even at the intersection level, right of way can become complicated
enough to make my head spin.  I've seen secondary to tertiary
crossings with a traffic signal during the day.  At night the lights
change over to stop-sign-like operation (flashing amber for the
secondary, flashing red for the tertiary).  So now we have to address
right of way that changes with time of day.  It becomes even more
complicated if no turns are permited during rush hours.

I don't know how to resolve this.