User talk:Boppet

From OpenStreetMap Wiki
Jump to: navigation, search


Hi Boppet! Thank you for your effort to make the pages more accessible to newbies! :-) I hope you agree with my follow-up changes (reverted your changes in minor parts) at Tags. --Aseerel4c26 (talk) 20:41, 11 January 2014 (UTC)

To be honest, I wasn't wholly in agreement with your changes, so I waited a while before replying. To me, the reversions seemed to place doctrines like "one key, one value" above the reality of actual tagging. Millions of ways in North America and in parts of Europe like Switzerland with "overlapping" highway route numbers have multiple values in a single ref tag separated by semi colons. To maintain that these are "single values" is for me unrealistic. I believe that the wiki's job is to explain and guide, and not to fight lost battles.
I was also a little surprised by your notes that said some deletions were "not a sentence." For me, the English wiki is often hard to read because it reads as if written in German (or should I say that it sounds to be with German sentence structures written?). Now, of course, I understand why this has happened, and my desire is to make the wiki more intelligible to non-German speakers. English sentences are much more flexible than those of French or German. In English, no authorities claim the right to define rules of grammar, control vocabulary or change the rules of spelling. Just about the only rule of English sentences is that there must be a verb; and even this rule is often "more honoured in the breach than in the observance". Jesus wept is said to be the shortest sentence in the Bible. It seems to me to be potentially applicable here. ;)
I'd be happy to work with you on a "better English" project, should you so wish. --Peter Davies (talk) 17:54, 22 January 2014 (UTC)
thank you for your reply, I will get to it soon! In the meantime, sorry for apparently making you quite pissed. :-( That was not my intention. --Aseerel4c26 (talk) 22:02, 22 January 2014 (UTC)
Hey, no problemǃ I know that we all have the same end goals ... even if we sometimes see different "ways" "forward". Peter Davies (talk) 23:32, 22 January 2014 (UTC)
Since that topic issue belongs to the talk page of the topic, I copied it to there: Talk:Tags. I hope this is okay for you. I also tried to reply/comment. Looking forward to your reaction.
It is quite normal in a wiki that you are not fully in agreement with someone's changes. This is how texts evolve and get better even with partly differing opinions. We also could simply revert back to the state of before our edits and discuss then until we find a consensus.
Thank you for your "same end goals" comment! Very good, I agree. --Aseerel4c26 (talk) 00:25, 24 January 2014 (UTC)


Thanks for the movement on highway=stop. I agree with using direction=*, but let's use way-dependent directions and not force routing software to determine the way direction and compare that to a specified cardinal direction. Rationale: stop nodes representing physical signage/marking locations are always part of one, and only one way.

My country doesn't have 4-way stops, but it seems to me that tagging such a thing with 5 nodes violates per Tag:highway=stop#All-way_stop violates One feature, one OSM element. A highway=stop node represents a _stop sign_ (and its associated line), and not the concept of _stopping nearby at some other OSM object_. Therefore I'd suggest recommending people choose one of two models:

  1. EITHER tag the central junction node as highway=stop and leave it at that. Do not add a direction. Because it is directionless and at a junction, it by definition represents an all-way stop with appropriate signage and markings on all approaches.
  2. OR make a node tagged highway=stop on each way near the junction node but not on any junction, ideally where the stop line physically is in reality. Supply a direction if needed for disambiguation. If special traffic rules for all-way stops apply here, then to clarify that the stop is part of an all-way stop supplement the node with stop=all_way. Do not tag the central junction node.

How does that sound? The former is nice and simple, the latter allows detail without unhelpful repetition.

--achadwick (talk) 15:04, 16 January 2014 (UTC) (watching this page for replies for a few months)

Andrew: Your country is mine too, but I currently live in the USA. I'd already been having a few further thoughts about stop signs and give_way signs. I'd like to see if we can work on this together.
I agree that five nodes for one 4-way stop seems excessive. My preference, like yours, is for one junction node at an all-way stop that says "highway=stop". However, the previous wiki writer seemed to prefer to mark each sign, and I wanted to accommodate that viewpoint also. Now, with a quorum of two (you and me) I suppose we could decide to drop that option. :)
Thinking about "one feature, one OSM element", one could argue that there is one all-way stop priority-controlled junction, plus four stop signs. Maybe those who want to mark the signs could use traffic_sign=stop on the sign nodes? This way there is less ambiguity about whether highway=stop is a highway junction, or a traffic sign. This accommodates the desire to mark the signs (where it exists) as well as the desire to show the junction control.
It may not be essential to add "stop=all" on the junction node but it could help to be clear, especially for new users who start mapping and do not start by reading the wiki (probably the great majority being those who want to map things, not to read and write wiki pages?).
What seems harder to me is how to handle junctions with stop signs on only one or two approaches. I had reluctantly advocated marking the sign with "highway=stop" plus "direction=west" (etc.) and I think you advocate a similar approach using "direction=forward" or "direction=backward". As way directions are so easy to flip in iD, Potlach 2 and JOSM (users here having proposed to flip them by the hundred in connection with signing things like role=west), and because so far as I know the direction tag would not auto-correct, "direction=forward" would be very easy to break by accident. On the other hand, I know that while Americans always think in terms of cardinal directions, other nations with older (wiggly) road systems rarely do. So I'm uneasy about using signs to indicate junction priorities at all, because I don't really like "west" or "forward".
I also tried coding some local stop junctions (US cities being littered with stop signs at almost every corner) and found it tedious to be adding every individual sign. The intersection nodes already exist, but the sign nodes do not. It would be better if we could map the junction control and forget the signs entirely, in my view, except where enthusiasts want to add them.
What I had then wondered is that at minor road stops, we could still map the junction node highway=stop or highway=give_way or highway=priority_junction, to show that the intersection is controlled as a priority junction (I'm a traffic engineer, sorry) and then find a way of showing which approaches do not stop. Usually in OSM it's already clear which is the higher level roadway from highway=trunk/principal/secondary/tertiary/residential, etc.
Where there is ambiguity, in mainland Europe priority is shown using the yellow diamond sign indicating "priority road." We could assign priority road status to the ways that don't stop if it's not already clear from their higher status in the hierarchy of ways. Priority_road status would trump the junction control highway=stop node. We could also add "stop=minor" at the junction node to distinguish it from "stop=all" in countries with 4-way stops. But outside of the US, Canada and South Africa, we already know that every priority junction only requires the minor road(s) to stop or give_way.
So even the 2-way stops could be signed without recourse to mapping the individual signs, marking the junction control on the intersection node, and the priority_road status on the ways. Another reason why I'd personally rather avoid mapping the signs is that signs go missing (or were never present) and mappers might leave them out, even though a stop line or give way line (or simple common sense) still makes it clear who has the priority.
My hope is that by mapping junction priorities, an app could eventually warn drivers who are about to run a stop or give_way sign without slowing down, thus helping reduce the 700,000 priority junction crashes per year here in the USA. The unsigned priority junctions are among the most dangerous, so we should not adopt a system that encourages a mapper to remove the stop/give_way tag just because the sign is missing. At this point, the map would become as deficient as the road sign system itself.
In the UK, where every stop sign must be individually approved by DfT (the Secretary of State for Transport, no less), you could occasionally have a stop sign on one "blind" minor road approach and give_way on another (with better sight lines). I'm currently not sure how to indicate this using only junction control node mapping plus European priority road designation on the main road ways.
Some people want to do all of this with relations, but my hope was to keep it simple. Back to you?
Cheers, Peter Peter Davies (talk) 02:34, 17 January 2014 (UTC)