[Page Update] Junction Style Guide

[PesachZ]

Images for the Wayfinder section are in progress. I've been a lot busier than I expected to be this summer.

For purposes of explanation, here's something from the MUTCD.

[ img ]

Same number of lanes exit as stay, so a wayfinder is called for.

Exit is on the right, and it's a freeway-freeway split, so ramp stubs are called for.

Left stub is named "I-50 W / LaSalle" because it's the continuation.

Right stub is named "Exit 55: I-79 N / Daly" because that's how exits are named.

In Waze it looks like this:

[ img ]

https://www.waze.com/editor/?lon=-90.21 ... 25&env=usa

BTW, the images and guidance in the MUTCD are explicitly in the public domain, so I may just use their images rather than try to find our own examples of everything, if possible.

what would you do in the same configuration, but where there was no BGS to the left, or the sign simply said "<I-50 W". I would imagine the stub name should then be "I-50 W", but then it wouldn't be different than s-in, so I thought it should be "to I-50 W" which will be spoken as "keep left at eye fifty west". (However it will display at the top of the screen "to I-50 W", which as I explained earlier I don't think is wrong either. Being that staying left at this fork takes you to the continuation of I-50 W.

[PesachZ]

what would you do

If only the final sign at split lacks the control city, but others approaching the exit include the control city, I go ahead and include the control city. Otherwise, I put a space after the cardinal direction ("I-50 W ") to make the name different. However, I can't find any I've done with the space that still exist on the map.

That is interesting, but so is using the alternate name field. We should test these and put one in the wiki as a guidance.

[PesachZ]

As a side note only the first "to" in the name is suppressed. A ramp for example names "to I-295 N / to I-95 N / New Haven" will be spoken as "keep right at eye 295 north, to eye 95 north, New Haven". If you need the word "to" to be spoken it can probably be accomplished (I haven't tested) by naming the segment "to to I-80 W".

Not only does it look wrong, it sounds wrong: "Stay right at to I-80 W".

it doesn't say the "to" it says "Stay right at I-80 W".

EDIT: These are actual sound files from the client which followed "stay to the right" they do include the word "to".
https://www.dropbox.com/s/v39kn3u9oyq2l3z/1402388758-314670-5.mp3
https://www.dropbox.com/s/h86o9s89th092a5/1402388899-949772-1.mp3

[PesachZ]

Alternatively, you could just add the control city anyway, if there is one. (The sign that used to just have an I-75 shield now has I-75 shields flanking the word "Detroit", so that made it easier.)

What about the city name field? If that changes does it force enough difference for it to be announced?

Maybe we can make a list of options available starting with the most likely and most favored for this case. Sounds like options include:

1. Add the control city name if it appears on the BGS.
2. Create another stub (before or after?) the stub for the continuation without any street or city name information.
3. (I personally never had a problem with "to" in the visual instructions. I am on a freeway and I come to a split. I want to know which way "to" continue with my current freeway. Seemed totally logical to me.)
4. Others I missed or forgot?

5. add a space after the stub name
6. remove the compass cardinal from the stub name
7. use the alternate name field

Sent using Tapatalk for Android 4.4.2

[PesachZ]

[quote="PesachZ"]As a side note only the first "to" in the name is suppressed. ... If you need the word "to" to be spoken it can probably be accomplished (I haven't tested) by naming the segment "to to I-80 W".

Not only does it look wrong, it sounds wrong: "Stay right at to I-80 W".

it doesn't say the "to" it says "Stay right at I-80 W".

You suggested writing "to to I-80 W" to get "to I-80 W" spoken. Now you claim that it still doesn't say the "to". But the BGS says "to" and Waze should say "to".[/quote]

I retract that last statement, I was confused what you were replying to. In the simple case of a BGS saying

Code: Select all

To
I-80 West

the instruction would be odd as you mentioned. In cases where the BGS says

Code: Select all

I-295 N
To
I-95 N I-678 N

the ramp could be " to I-295 N / To I-95 N / I-678 N" for an instruction "stay to the right at I 295 north, to I 95 north, I 678 north"

Sent using Tapatalk for Android 4.4.2

[PesachZ]

Geometry

All junctions should be laid out to follow a few basic forms to ensure better prediction of navigation behavior. The table provides guidance for how to set the angles of junctions between segments to achieve the desired navigation instruction listed.

Table
Departure Angle | range | etc..
0 | 0-21 | ...
33 | 22-44 | ...
90 | 45 - 150 | ...

Departure Angle = The deviation from straight ahead (0) in degrees.
Range = The range of angles that will result in that condition.
Road Type = The road type connected to the junction. Higher road types take precedence.
Traffic Side = Left-hand Traffic, Right-hand Traffic, or both
Navigation Instruction = The instruction to be given through the client app.
Stay away from the ends of the ranges (22 and 45) as those are tipping points between the settings that are hard to determine visually on the map. Instead stick with the centerpoints (0, 33, 90+) shown in the departure angles column for better predictability.

After doing extensive testing, documented in the new page https://wiki.waze.com/wiki/How_Waze_det ... _maneuvers it is obvious that some figures in this table are inaccurate. Would anyone mind if I updated this section to be compliant with our test results.

Namely

• Remove the first row 0-21° because you will get instructions at 0° if the conditions are correct, any segment that will give an instruction at less than a 43.68° will still give the same instruction at 0°.
• Change the second row to be 22° | 0°-43°
• Change sentence at end to say "Stay away from the ends of the ranges (44°) as those are tipping points between the settings that are hard to determine visually on the map. Instead stick with the centerpoints (22°, 90°+) shown in the departure angles column for better predictability."
• Add a note to the end "Avoid having the angle between any two diverging segments be less 15°, this makes it very difficult to select the turn arrows in WME, as well as in the client 'Report>Road Closure'. with a link to section in the JSG/diverging roads

EDIT:

• I would also add a note at the end that "Angles can be adjusted with a dogleg, using a geonode very close to the junction at the desired angle while zoomed in. This will give the desired instruction, but will not affect how the road displays in the client, and livemap."
• Add image of a dogleg segment.[ img ]

Sent using Tapatalk for Android 4.4.2

[PesachZ]

I would prefer to set the minimum recommended angle at 10°, still easy enough to edit and in accordance with this section of the Interchanges page.

Is it still is enough to select for a closure in the app.? There has been a lot of complaints in the forum about inability to select the correct arrow when trying to report the road closure while driving.

Maybe we should settle for 15° as it's between 10-20° mentioned in the interchanges page

Sent using Tapatalk for Android 4.4.2

[PesachZ]

I don't wanna be too rigid about it since many editors will be eyeballing that angle anyway. I don't know that it's worth editing everything on the map to accommodate the admittedly broken and ineffective closure tool.


Sent from my iPhone using Tapatalk 2

what if I copy the note from interchanges saying 10-20°?

Sent using Tapatalk for Android 4.4.2

[PesachZ]

That sounds good.


Sent from my iPhone using Tapatalk 2

I'm good if you want to make these discussed changes directly.

I made the changes. Please feel free to check them.

[PesachZ]

you'd have to also define "travel lane" to mean a lane that's been there a while rather than a lane that started just before the exit

Do-able: a travel lane could be defined as a lane that, at some point before nearing the exit, was neither striped nor signed as "exit only".

and that two exit only lanes isn't always enough, and that one exit only lane is sometimes enough.

The purpose of the wayfinder is to alert the driver to something that is both unexpected and may require action to continue, i.e. changing lanes.

If no travel lanes depart, no action is required. Therefore only situations involving the loss of one or more travel lanes should be considered for wayfinders. All such situations may require action but not all such situations are unexpected.

So now we come to the heart of it: how many travel lanes have to depart for a naive driver to consider the loss of lanes "unexpected"?

On the atypical exit side, exits are so rare that the loss of even one travel lane would be unexpected and the situation would qualify for a wayfinder.

On the typical side, maybe others feel differently, but I am unaccustomed to losing more than one travel lane at an exit no matter how wide the freeway. I've never driven the widest freeway in the world measured by main travel lanes -- the Ontario Highway 401 at 9 lanes each way -- but I imagine that on my first drive even on that road I would be surprised to lose two travel lanes at an exit.

So, that was my thinking in defining "non-obvious continuation" as the loss of at least two travel lanes on the typical side, or one on the atypical side. The only case where losing a single travel lane on the typical side would be unexpected, IMHO, would be if there had been only two travel lanes in that direction in the first place. But that case is covered by the "single lane continuing" clause.

This is for controlled ramps, highways and freeways. For uncontrolled highways I think losing even one travel lane on the typical side would constitute a non-obvious continuation.

So, another try:

At exits,

• the continuation path for a controlled-access ramp, highway or freeway will be considered "non-obvious" if it has least two fewer travel lanes in a given direction after the exit than before it on the typical exit side.
• The continuation path for an uncontrolled ramp or highway, or for an exit on the atypical side, will be considered "non-obvious" if it has at least one less travel lane after the exit than before it.
• (Added in edit) The continuation path for any highway will be considered "non-obvious" if signage preceding the exit suggests a split, for example with arrows for the continuation that point in a direction other than straight ahead.
• A continuation path comprised of a single lane will always be considered "non-obvious".

Where "travel lane" is defined as a lane that, at some point before nearing the exit, was neither striped nor signed as "exit only".

there are many locations in NYC where two lanes exit the freeway on the right, but one of those lanes is usually started just after the previous exit, the others here I can think of also only have only one travel lane continuing through so would qualify regardless.

Sent using Tapatalk for Android 4.4.2