sketch wrote:Aren't there known navigation-instruction problems with the collector-distributor cloverleaf setup recommended by the guide? As in, the keep-left/keep-right at the middle junction point where the ramps meet are not given with restricted turns like that? We should let detour prevention take care of it.

This has been in the back of my mind since we added the detour prevention explanation to the wiki. My thoughts are threefold.

1) Yes, I believe that the keep-right at the middle junction point is not given where the left option is restricted to prevent through routing. My understanding was that these instructions used to be given, but then stopped. There was thus some question whether we should really be editing to make the instruction come back or whether waze should just bring the instruction back. I've heard nothing recently that indicates waze will bring the instruction back.

2) The detour prevention mechanism is only effective if the distributor/collector lanes don't exceed 5 kms. Is there any utility for these methods for very long distributor/collector lanes?

3) Waze did just confirm that one example of a routing off/on the freeway was due to the fact that the detour prevention is just a penalty that was overcome by very heavy traffic. Of course, we don't want waze to route off/on to avoid very heavy traffic (that's kind of the point). So the question becomes: is it worthwhile to employ both methods to increase the penalty as much as possible?

I agree with your other points regarding ramp junction placements on the freeway.

I agree. The points were essentially what I thought need to be considered before changing the wiki and were my justification for the changes you propose not having been made when introducing the detour prevention to the wiki.

With regard to your counterpoint 2), I guess that raises the question whether we should have a wiki section that addresses large complex multiple road interchanges with long collector/distributor roads. My gut feeling is that that they should be few and should be handled by experienced editors so that wiki guidance is not essential.

With regard to your counterpoint 3), I take it that means that we should continue to recommend the restrictions that don't affect routing instructions, like the restriction for ramp to ramp routing on the basic diamond interchange.

So should we simply delete the Complex/Distributor Collector section as being rendered obsolete by the big detour prevention mechanism?

I'm not sure what to do with the "bottom" cloverleaf ramp setup. It seems to me to be an unnecessary alternative where a single recommended method would be better for consistencies sake. However, I remember there being some editors that preferred this option, but I can't remember why.

Should we try to get Weeezer's input? I'm hoping that springtime is treating him better than the winter did.

If you are planning on revising the text regarding Exits, I would like some confirmation on some of the Exit geometry statements.

"The entering segment and one exiting segment must be one of the three Highway/Freeway types."
I don't believe that it is a requirement for "one exiting segment" to "be one of the three Highway/Freeway types" to get an exit instruction. In any event, this in directly contradictory with the description of using ramps for wayfinders, which states that if both exiting segments are ramp type you get a keep left and exit right.

"The Highway/Freeway exiting segment must have close to a zero degree departure angle from the entering segment."
I don't believe that this is requirement. First, as noted above, you don't even need a "Highway/Freeway" exiting segment. Second, I believe that you will get the exit right instruction as long as the ramp is not the best continuation. That only means there has to be another segment with a departure angle of less than 45 degrees.

"The other exiting segment must be of the type Ramp"
This is also not true. Exiting to a primary street or street segment will also give an exit instruction.

"The Ramp exiting segment must have a departure angle of between 20 and 30 degrees from the entering segment."
This is not true either. Any angle less than 45 degrees will work. This includes zero degrees. Although, the exit instruction is not given if the ramp segment overlaps with other exiting segment (in other words, both exiting segments can't be zero degrees).

"When those conditions are met, the navigation will present an "Exit Right/Left" instruction when the ramp is to be used . . ."
You only get an "Exit Right" in right-side traffic jurisdictions and an "Exit Left" in left-side traffic jurisdictions.

sketch wrote:Correct. To get an "exit" instruction, the previous segment must be a highway/freeway type, and the "exiting segment" must be (1) to the right and (2) either ramp type or {...}.

{...}: I can't remember if it has to be primary street/street, or if it just has to be a lower type than the previous segment. In other words, does MH --(25°)--> mH give "exit" or "stay"? I feel like this used to be in the Wiki but it wasn't explicit in the relevant article.

I'm fairly certain that it's "primary street/street." I believe that F--(25°)-->MH, F--(25°)-->mH, MH--(25°)-->mH all give stay instructions. I found some old discussions with Alan about this, but all the example links are the old style and when I could find the intersection, the road types had been edited. However, I believe that this is consistent with what Dror originally posted here.

I really don't understand why we map the Lehi, DDI that way. It works fine, but everyone seems to think that you must split the roads for a DDI. It don't understand why. If the road is not otherwise split, the DDI really doesn't have to change anything on the map. I'd prefer at least one unsplit example of a DDI (if there is one).

daknife wrote:Because if you don't split it, you get bad turn instructions. As I said, I've mapped all five of the existing ones in Utah and tried to figure out a non split method, it does not work. I tried it. With the way the traffic crosses over. You have to have split lanes, turn restrictions alone (remember they are only penalties, not absolute) are not sufficient to ensure proper routing and correct instructions.

What bad instruction do they give? Where does the route go bad? I don't doubt you, I just don't understand.

Ok, I willing to buy that and defer to those that think the split helps understanding the interchange. In fact I have in practice. There is one kind of half DDI around me here. When it was opened I mapped it unsplit, drove it and seemed to work fine. But within a week or so someone changed to split. This being before I could lock things at high level, and as you can see the aerials haven't update yet, it got changed back to match the aerial. I would change it back to the unsplit DDI and then someone else would come along and split it. This happened multiple times. I've never changed it from split to unsplit because it works split and clearly some editors prefer to see it split. To me the split seems like adding all 8 turn lanes where a split road crosses a split road. But given the opinions here, I guess this is a circumstance where the extra complication does provide a benefit to users that I haven't previously understood.

I have come across another flaw with the old collector/distributor guidance. If you have a full four-leaf clover interchange and you restrict only the straight through movement on the ramp to prevent routing back to the freeway, then waze can still route around all four leaves and back to the highway. You would think this would never be faster, but it seems to have happened.

Well if the name of the freeway is consistent, it should be prevented by the big detour prevention mechanism. Alternatively, if you have brought the leaves together to a single junction to prevent the straight through movement, you can also prevent the leaf-to-leaf movement. If that is done on one of the other leaf-to-leaf junctions that should also prevent this. But I would suggest relying on the detour prevention.