I have some comments on the page.

We need to label and define the critical angle. Something along the lines of "Turn Angle - the turn angle of sN is the angle between continuing straight on s-in through the junction and the departure direction of sN from the junction. The angle between the linear segment between the last geometry node of s-in and the junction and the linear segment between the junction and the first geometry node of sN is the supplementary angle to the turn angle of sN."

The Best Continuation concept still retains many of the flaws of the original page.

• "One of the segments leaving the junction will be considered the 'Best Continuation'. I think this is misleading, I would change it to "may be considered."
• If would define Best Continuation as follows: "If s-out has a turn angle of less than 45 degrees then it will be considered the best continuation unless another sN has better segment name and segment type match with s-in as compared to s-out. Segment name is more important than segment type."
• The chart for Best Continuation carries forward some of the flaws of the original page. Basically, the requirements for Best Continuation are does s-out have a turn angle less than 45 degrees followed by this section of the flow chart:

  Best Continuation.png

  (54.5 KiB) Downloaded 714 times

For the list of conditions I would change 1 as follows:
1) If the junction has only 2 segments, the instruction is "CONTINUE".

I'd leave out the T junction information and the discussion of the original text's discussion of the behavior of locked nodes.

In the flow chart I'd add the U-turn instruction to the normal roundabout instructions:

I like the changes. For the angle definition, I'd remove "turn" in the last line of the "in techinical terms" paragraph.

I like the turn graphic, but the first question I usually get when trying to explain this is "What happens if the roads aren't straight?" I'd be inclined to show what is E 5th St in the graphic as a curved road, maybe even with the blue selection highlighting showing the geometry nodes.

In the best continuation chart, you have S-out is Best Continuation and S-out is Not Best Continuation reversed from what I was showing. (Although If the angle is not <45 is correctly going to S-out is Not Best Continuation.)

kentsmith9 wrote:I found some conflicting information from Ianhouse in the UK in this thread. His study on Time-based tolls indicated he was getting (undesired) turn instructions when all road types and names were the same with one S-in and two S-out, UNLESS they overlapped exactly at the two S-out points. The testing was in December on the World server. Any chance they would have different routing than we have?

I posted over on the other thread as well. I was looking back at some notes for instruction testing that I did awhile ago (before last December) which confirmed that I did get instructions when all segments at a split were identical. So I don't think this is different in the infrastructures.

I agree with PesachZ that the chart is consistent with Ianhouse results. You would go through the chart as follows for a freeway split with all segments having the same name where you take the right branch: 01 19 02 04 13 05 06 07 08 20 25 15 16 - Keep Right. If you didn't have the overlap on the CCZ segments then if the time restriction is the segment to the right and you are routed through when the restriction is not in place then you would go through the chart as 01 19 02 04 13 05 06 07 08 25 - Keep Right.

So I'm now concluding that the overlaps are necessary for the CCZ segments and are consistent with the revisions to the turn instructions page.

PesachZ wrote:City names don't seem to affect this algorithm as was demonstrated by an intersection in Miami Fl.

Thanks for reverse engineering this. Does this mean that the primary names must have the same city for the alt matches to work? Or does this mean that city matching is no longer a requirement?

Makes sense. Thanks.

pumrum wrote:This ramp is a perfect example. State Route 20 West splits off from Special Route 401, which continues north as a freeway to the airport. Unfortunately, with this latest change, drivers are no longer getting an "exit right" instruction which defies all logic and common sense. There really should be a ramp exclusion from the TTS logic.

While I agree that I don't understand the intention behind this logic, I'm wondering if there is any disadvantage to giving this segment an alternate name of "State Rte 401 W." Would that not return the TTS to providing an "Exit" instruction for the ramp. I understand that this isn't going to be entirely intuitive for new editors that haven't studied PesachZ's work. But isn't that at least a solution to give the routing and instructions that we want at this point?

PesachZ wrote:The easiest instructions to new editors, when the wrong instructions are being given at a junction and everything is named correct, may be to set up a wayfinder.
Wayfinders stubs should be set without any alternate names.

Hold that thought. Waze staff is currently trying to convince me that ramp wayfinder stubs adversely affect routing. I'm not convinced yet. This is in connection with the routing issue raised here. The routing team doesn't understand why this segment exists. (So at the very least, the routing algorithms have been developed without contemplating what effects these ramp wayfinder segments may have on routing.)

PesachZ wrote:

CBenson wrote:Hold that thought. Waze staff is currently trying to convince me that ramp wayfinder stubs adversely affect routing. I'm not convinced yet. This is in connection with the routing issue raised here. The routing team doesn't understand why this segment exists. (So at the very least, the routing algorithms have been developed without contemplating what effects these ramp wayfinder segments may have on routing.)

Well first off Im not sure I'd believe they cause untoward routing effects.

So when I changed the wayfinder segment referenced in the other thread from ramp type to freeway type (thus destroying its function to trigger the keep left instruction) waze did route over the freeway again.

Yes, I remember this issue now. That chart never updates on the page for me. Doesn't matter if I'm logged in or using someone else computer. Displayed on the page for me is the 1:30 11 June 2014 version. If I click on the image, I get the file page with showing the current version. If I remember correctly, its been this way for me since June 2014. I've forgotten recently to make sure I click through to the file page to see the current version.

https://wiki.waze.com/wiki/File:Turn_al ... wchart.png