Two-way Ramps vs. Routing [Split from WME Validator]

Moderators: Unholy, bextein

Re: two-way ramp rumor

Postby sketch » Wed Feb 19, 2014 8:32 pm

AlanOfTheBerg wrote:
dbraughlr wrote:When the on-ramp is restricted (closed), will Waze route the wrong way for a short distance over the off ramp to reach the highway?

In any situation like that, there is always the possibility that the routing server will incorrectly (and dangerously) select routing against a 1-way segment, through turn restrictions, etc. That's an unfortunate byproduct of the way the "penalty"-based system it uses. Nothing is impossible.

But is the penalty based on segment length at all? From what I understand, penalties are assessed on transition, not along a segment. It seems dbraughlr's worries are founded in a misunderstanding of the penalty system (or, I might be wrong).

SuperDave1426 wrote:I agree with dbraughlr - I really think this should be looked into more fully before making a change to the ramp direction check.

We've been using two-way ramp segments for years without incident. The inclusion as a check may comport with INTL rules and practice for ramps, but they do not comport with ours.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: two-way ramp rumor

Postby sketch » Wed Feb 19, 2014 8:59 pm

dbraughlr wrote:People say that the "same endpoints" error is real. I am on the side of excluding the validation until it is shown that it is a still real problem, especially since the solution seem to be to insert an unneeded node node that will have to be removed after the bug is solved.

It is a bug known to many. There have been multiple threads about it. I can't find them because they were started without an understanding of why Waze was giving such bizarre routes (because we figured it out using those threads). I'm trying to find them now, but I don't know if I will. If you're not willing to take a few champs' words for it, I don't know what else to tell you.

As for sketch's comments: I didn't start the rumor; so to suggest that it is based on my misunderstanding is entirely your misunderstanding. I'm just the messenger. You'll have to do your homework if you want to chide the originator of the rumor.

All I'm saying is that your understanding of the penalty system may be wrong. Mis-understanding. I didn't blame you for it.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: two-way ramp rumor

Postby sketch » Wed Feb 19, 2014 9:42 pm

CBenson wrote:I'm not sure what "the error" is. Seem to mostly be a display problem with the routes shown in URs as discussed for example here and here. It may actually effect the user when the endpoint of the route is on one of the two segments that connects the same two junctions.

It's more than just a display bug. I can't find it, but I remember a thread a few months ago about a one-segment road side parking area. The user in that case got an incorrect turn instruction.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: two-way ramp rumor

Postby sketch » Wed Feb 19, 2014 9:49 pm

CBenson wrote:I'm not sure what "the error" is. Seem to mostly be a display problem with the routes shown in URs as discussed for example here and here. It may actually effect the user when the endpoint of the route is on one of the two segments that connects the same two junctions.

Your post, the second post, in the first thread you linked shows a user that followed the "display bug" route rather than the TTS route. I don't think it's reasonable to consider something as important as the route line being displayed incorrectly is so minor as to be considered merely a "display bug". Especially when the user isn't using sound. Is the user more likely to follow the arrow in the corner, or the route line on the screen?
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: [Script] WME Validator 0.7.1 (BETA) / 20.02.2014

Postby sketch » Thu Feb 20, 2014 12:03 am

dbraughlr wrote:What basis do you have to suspect that my understanding might be wrong?

Four and a half years on Waze and its forums have given me a pretty good understanding of how it all works.

An unsubstantiated rumor is not a substantial basis for a Validator warning.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: Two-way Ramps vs. Routing [Split from WME Validator]

Postby sketch » Fri Feb 28, 2014 5:17 pm

Could you point me to an example of a place where this might be/would be/has been a problem? Not that I don't believe you, I just can't picture a situation where it could be.

I have seen problems with ramps leaving out important instructions, but it hasn't been with two-way ramps. Look at the UR here, keeping in mind that the two selected segments were Ramp type when I left it: https://www.waze.com/editor/?lon=-90.05 ... 6,57649175

Notice the lack of the "keep left" instruction onto the freeway.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: Two-way Ramps vs. Routing [Split from WME Validator]

Postby sketch » Sat Mar 01, 2014 2:47 am

AlanOfTheBerg wrote:It shows a stay left in livemap instructions.

In which direction, westbound? Because Jason increased the angle to the current one while they were still ramps, tested it, and it still didn't give the "stay left" instruction. I tested it too, I think.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: Two-way Ramps vs. Routing [Split from WME Validator]

Postby sketch » Wed Jan 20, 2016 3:00 pm

GooberKing wrote:To put a twist on this topic, we've come across an interesting situation in CT that I think would work better as a two-way ramp. This ramp is currently mapped as two one-way ramps, despite the road not meeting requirements for a split road. I'm proposing we redo this as a two-way ramp, but make it with two connected segments: One named "to SR-68" and connected to N Colony Rd, and the other named "to US-5 / to SR-15" connected to Church St.

I think this would make the road look better on the map and make it easier to maintain, as it reduces segments and geometry nodes. I realize Validator will get an aneurysm over it, but is there any technical reason why this would not work? Or is this just one of those "That's just not how things are done" situations?

I don't see any reason why this won't work, and I'm sure it's working fine. But it's also a bit of a hack, and in situations like this where the previous iteration was also likely working just fine and wasn't really any more difficult to maintain, well, all you're doing by reconfiguring it is losing a bunch of historical speed data.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: Two-way Ramps vs. Routing [Split from WME Validator]

Postby sketch » Thu Jan 21, 2016 3:16 pm

Sure, I understand that. There are different definitions of "hack", but I think a better one is "using nonstandard editing practices to accomplish something that standard editing practices cannot accomplish".

Here, maybe each way is somewhat of a "hack", considering that the former iteration of it used two parallel segments for an undivided roadway. But the other way at least looked normal, in such a way that there's no chance a passing editor would consider it some sort of mistake, and such that it didn't throw any Validator flags (not that that in particular is evidence of whether an edit is okay, but it does signal passing editors that something might be wrong, whether or not something is wrong).

There's also the fact that Waze 4.x shows the name of the current road on screen, and that can be potentially quite confusing for someone on the second half of that ramp, thinking, "Wait, I just got off SR-68!?"

My point, however, is that the amount of work that went into reconfiguring this interchange far exceeds the amount of work it would have been to maintain it in the previous configuration ("if it ain't broke, don't fix it"), especially when you have to start explaining it to every passing editor.

Simplifying the map is not an end in itself. Our goal is not to make the map "easier" to edit. We are here to make the app work the best it can with the map we have, and secondary to that is making the map understandable to other editors, not necessarily simple. In some respects, in some areas, the Waze map has been or was oversimplified for a long time, and that comes at the expense of ease of editing. Why do you think so many noobs have taken an undivided road and built them into divided roads? Because it's reality. Because it's not easy to understand why someone would draw a clearly divided roadway as a single segment. Because the editing rules that came about because of that are not as easy to fathom as they would be if the segments were simply drawn in ("why all these turn restrictions?!").

Why did we do it that way? This powerful spectre of "simplicity" that has haunted us for half a decade, making us edit the maps in ways that are more difficult to understand (against our secondary goal of making the map understandable to editors), forcing us to "reach our destinations" when we are still on the wrong side of the road (against our primary goal of making the app work well)...

Maybe this reconfiguration is marginally "simpler" (OK, two fewer nodes on the connected segments, though you still have the same number of ramps...), but it's more difficult for a passing editor to understand, and it has negative effects on navigation in the client (historical data is lost, plus confusing current road name display in 4.x). So, where does that leave it?
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

Re: Two-way Ramps vs. Routing [Split from WME Validator]

Postby sketch » Thu Jan 21, 2016 5:59 pm

CBenson wrote:So does it do it for ramps segments? That would be a significant change as we do generally edit with the idea that ramp names won't display on the client map. Kind of surprised there haven't been complaints.

It did as recently as a few weeks ago, and I believe it still does, though I can't say certainly right now. I did complain, but it seemed to fall mostly on deaf ears.
sketch
Waze Global Champs
Waze Global Champs
 
Posts: 6143
Joined: Sat Aug 08, 2009 6:13 pm
Location: New Orleans, LA
Has thanked: 1599 times
Been thanked: 2146 times

PreviousNext

Return to Navigation & Routing

Who is online

Users browsing this forum: Bing [Bot]