Drawing Roundabouts - Google instructions are incorrect.

[vandelay]

Hello,

 

I am posting this because I have come across a great number of incorrectly drawn circular intersections.  Google instructions located at http://support.google.com/mapmaker/bin/answer.py?hl=en&answer=1208728 are incorrect when it states that a roundabout is the same as a traffic circle.  They are not the same and should not be drawn the same way.

http://www.alaskaroundabouts.com/mythfact1.html

http://www.co.washington.mn.us/info_for_residents/transportation_division/roundabout_u/

http://www.dot.ca.gov/dist3/projects/yolo16_kentucky/documents/RndTrafficCircle.pdf

https://www.dot.ny.gov/main/roundabouts/background

 

I am not an expert at using MapMaker yet, I am however a licensed Professional Engineer and certified Professional Traffic Operations Engineer that specializes in roundabout design, operation, and driver education.  I am posting this as a resource to others and to hopefully spur discussion and a change in the incorrect instructions that Google Map Maker currently provides at http://support.google.com/mapmaker/bin/answer.py?hl=en&answer=1208728.

 

Map makers, whether volunteers or Google staffers, should always consider the actual design and configuration of any circular intersection, whether it is a roundabout, traffic circle, or a rotary.  It is a mistake and potentially dangerous to try to draw all varieties of circular intersections following the same eight-point template which is often incorrect.  I welcome any comments to this discussion.

 

It is important to note that the same terms sometimes mean different things in different countries.  Drawing roundabouts incorrectly, especially multilane roundabouts, can lead to bad information and even crashes.  Most importantly is that for a roundabout, a normal four-legged roundabout in the US, Australia, New Zealand (and others) should be drawn with FOUR intersection points, not eight.  The four segments connecting these points should be classified as "Roundabout/Traffic Circle".  Below is an example of a correctly drawn US/NZ/Australian style roundabout. 

Correct drawing:

Four intersection points

Four segments creat the circle as a "bulged square"

Lines follow through (straight ahead) movements and lane centers.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Below is the same roundabout mapped incorrectly, though this is how they are often drawn in Google Maps.

INCORRECT:

Eight intersection points, eight circle segments.

Lines follow left turn movements rather than through movements and do not follow lane lines or lane centers.

Incorrectly suggests that entry and exit is a turning movement, which conflicts with actual traffic controls such as signs and pavement markings.

 

 

[delta fox]

Hi, 

Many of us are not experts in English and for us traffic circle and roundabout are just the same thing, especially in MapMaker since it is how Google defines it. Even English-speaking people have different advices, depending of countries, regions etc. (see Wikipedia). And when translated to any other languages (perhaps you don't know that MapMaker has an interface in many languages) such subtleties are far out of our reach...

Now, if you think Google is incorrect, please explain why. Nothing in the help link you quote tells that we should use 8 intersections or 4 (personally, I prefer 4 as in your upper picture - since it is lighter to edit later if there is a new imagery - but with a 'circular' circle as in your lower picture).

And priority rules vary very much from one country to another in the world, even in one country there may be circles with priority for incoming traffic or for traffic already on circles, or traffic lights, etc. Hopefully drivers still look at traffic signs when driving, not only to their GPS device ;-). You may enter traffic signs at each intersection (though it's not displayed on the map, and I don't know how GMM does use them), that's independent of how the circle is drawn.

[djboge]

I think this is an interesting discussion and a good point to bring up. Visually I think the above example is superior, but I believe the distinction between the two lies in the way they will generate directions. For example, say the road running from the bottom left to top right corner is Main St, and the other is Smith Ave. You are driving NE on Main and want to go SE on Smith.

With the bottom example, you'd get the following directions.

Head northeast on Main St.

At the traffic circle, take the first exit onto Smith Ave.

If this roundabout were drawn in the above example, the directions might say:

Head northeast on Main St.

Turn right on Smith Ave.

If you were staying straight on Main St, I believe both examples would result in the directions:

Head northeast on Main St. 

At the traffic circle, continue straight to stay on Main St.

I'm not sure what other distinctions are to be made between the two. I understand the theoretical difference but I'm not sure which way is better, not from a technically-correct standpoint, but from the standpoint of generating proper directions.

[delta fox]

I'm afraid that in both cases, driving aid device would say "at traffic circle, take second exit" rather than "continue straight" - this may be the same, but it is more difficult to figure out that the second exit is straight ahead - especially if there are small roads on the same roundabout that may be not counted on the map...

[vandelay]

In the example photo that I added, the signs and pavement markings would tell you that exiting is a "straight ahead" movement, and continuing around the roundabout is a left turn.

If you look at the "incorrect" drawing, a driver seeing such a configuration would conclude that they make a "right turn" to enter the circle and another "right turn" to exit the circle.  This is not true and it disagrees with the signs and pavement markings.  It can also lead to driver mistakes that cause crashes when entering or exiting.

You are both correct that the driving directions, as currently programmed, would say to "take the second exit".  But in the future the driving directions might not be programmed the same way.  Additionally, a driver needs to know as they approach that they will be making a left turn.  A driver approaching the roundabout in this picture MUST be in the left-hand lane to make a left turn.  To say that they should "take the second exit" implies that they should instead be in the right-hand lane.  If a driver enters in the right-hand lane and wishes to turn left, they will crash with vehicles that are next to them.

I want to clarify that this applies only to the spiral type of roundabout that exists in the US, Australia, and New Zealand.  Roundabouts tend to function the same way in the United Kingdom though they don't include striped lanes in the circle.  Roundabouts in Spain and Germany tend to be the opposite and these issues would not apply.  As Delta Fox correctly pointed out, the word "roundabout" is used to apply to different types of intersections depending on where you are in the world.  This is indeed a problem of language, but the underlying advice is still true - Regardless of what it's called, the mapping configuration should match the actual design rather than try to fit an incorrect template.

I'm glad to see that this is generating discussion, I hope there is more to come.

[vandelay]

@ djboje - Good points and thanks for participating in this discussion.

To your point, ideally a driving aid device would say "At the roundabout, turn left onto Smith Ave".  This is superior to saying "Take the second exit" because a driver approaching the roundabout MUST be in the left hand lane in order to turn left.  If they aren't in the proper lane when they reach the roundabout, they commit a traffic violation and may cause an accident.

If the driving device says "Take the second exit", a driver has no indication of which lane to be in.  Describing an "exit" implies that a driver must be in the outer lane in order to depart the intersection, which is not true for US/UK/Aussie/NZ designs.  It's also unclear to a driver as to whether or not the immediate right turn is counted as the first exit or omitted from the count.

[djboge]

> I'm afraid that in both cases, driving aid device would say "at traffic circle, take second exit" rather than "continue straight" 

> You are both correct that the driving directions, as currently programmed, would say to "take the second exit". 

I didn't say that it would tell you take the second exit (assuming we're talking about the third example I gave).

As long as you're staying on the same road, directions (tried it on Maps, MM, and my phone) will all say "at the traffic circle, continue straight..." If you are traveling straight-ish but onto a different road, it will indeed say "take the nth exit onto..."

Here are three examples from my area demonstrating this:

http://g.co/maps/suwjk

http://g.co/maps/a7y6t

http://g.co/maps/qa2sc

Another example from my phone: http://i.imgur.com/Y0GCk.png

In my opinion this is the correct behavior. If I'm already ON Main St, why should I "take the second exit" onto Main St.? With the current implementation of directions, it doesn't seem that this type of perhaps misleading direction will be given.

[vandelay]

"Now, if you think Google is incorrect, please explain why."

Google is incorrect because it suggests drawing the circle as one continuous roadway.  This is not true for US/Australia/New Zealand designs.  It is often true for roundabouts in Spain and Germany because they are designed differently, and it is also true for traffic circles which are not the same as roundabouts.

More specifically, Google's advice tells map makers to create a common/shared segment between an entrance and an exit, when normally such a shared segment does not exist.  To make a 90 degree right turn at the roundabout used in my example, the driver makes ONE right turn, not two.

To make a left turn, according to all the signs and pavement markings, a driver first goes straight ahead to enter the roundabout, and then makes a left turn, and then a straight-ahead movement to exit the intersection.

By the way, I am working on an effort to establish official definitions of "roundabout" and "traffic circle" that will influence this debate in the long run.  But trying to get a worldwide community of experts to agree on terminology is a long process and can generate controversy.  One recent example is astronomers seeking to refine the definition of "planet" which made many people upset when Pluto was no longer covered by such a term...

[vandelay]

To clarify, let's say that a right turn at this roundabout would be the "first exit", a straight-ahead movement would be the "second exit", and a left turn would be the "third exit".

The incorrect drawing suggests that entering and exits are right turns.  However the signs and arrows on the pavement refer to a left turn as a left turn.  So it's important to refer to it as a left turn, or a driver will not know that the "third exit" refers to a left turn.  They need to know that they will be making a left turn as they approach, so that they can choose the required lane.

The terminology is less important than ensuring that the lines match the actual arrangement of the lanes, the signs, and the required behavior.  In the incorrect drawing, it suggests there is a shared segment between each entrance and exit, which is not true in the US/UK/Aussie/NZ type of roundabout design.

[djboge]

I think your use of the phrase "left turn" can be misleading in this discussion - theoretically you're not ever making a left turn (as we think of hard turns) - you're either going around the circle (technically turning left) or "exiting" to your right. The left turn arrows on the pavement I don't believe indicate a left TURN per se, moreso that the marked lane continues around the roundabout in a counter-clockwise direction.

Really the only difference between the "correct" example of a roundabout and a regular + intersection of two divided roads is the addition of the "roundabout" segment usage to the middle "square" of segments. I'm not sure how adding this usage affects the directions generated going through that intersection, and since direction updates are slow, its not really possible to experiment (unless you find a roundabout that needs marking).

In this example, if you were traveling north (northeast) on Main, and wanted to go west (northwest) on Smith, I'm not sure which of these directions would be generated:

a. Head northeast on Main St. At the traffic circle, take the second (or would it say third?) exit onto Smith Ave.

b. Head northeast on Main St. Turn left on Smith Ave.

At any rate, I think we're really splitting terminology hairs here. In general, people know how to navigate roundabouts and pay attention to the signs in the area. It's not as if we're trying to program a robot to drive correctly based on Maps directions.

I'm still not sure what the SPECIFIC problem with drawing roundabouts in the "incorrect" way is, other than the fact that its incorrect IN THEORY.

[vandelay]

All three of your examples tell a driver to go straight, but your third example (Geiger Grade) actually shows making a right turn.  This is a great example of the problem.  The pavement markings in the photo and signs clearly show that a driver is making a right turn, not going straight.

If the roundabout were drawn correctly (instead of being drawn as a traffic circle), the directions would tell the driver to "turn right", which agrees with the signs and markings at this location.

[vandelay]

@djboje

"theoretically you're not ever making a left turn (as we think of hard turns) - you're either going around the circle (technically turning left) or "exiting" to your right."

As a professional engineer, I strongly disagree with the above statement (though I don't blame you, it's a commonly held misconception).  In every legal sense, you ARE making a left turn, this is why the signs and markings designate it as such.  And you do not "exit" to your right unless you are actually turning right, for example northbound to eastbound.  Otherwise, you are making a straight ahead movement when you exit, and your wheels don't even turn to the right until after you have exited the intersection.

[vandelay]

@ Delta Fox

"(personally, I prefer 4 as in your upper picture - since it is lighter to edit later if there is a new imagery - but with a 'circular' circle as in your lower picture)."

Although it's nice to draw curves smoothly, it actually should NOT be drawn as a circle.  Yes, a driver making a left turn makes it as a smooth movement, but corner radii exist at EVERY intersection and we do not draw every turn radius at any standard intersection.  It is simply an issue of the shape of the pavement so that drivers don't strike a curb/kerb as they turn left.

[djboge]

But that's just the thing, you're NOT turning right. Just because there's a right arrow painted on the ground doesn't make it a right turn. You're going straight. You might sort of gently turn your wheel to the right to follow the road, but you're on Geiger Grade, you want to keep going on Geiger Grade - it doesn't make any sense for it to say, "continue east on Geiger Grade. Turn right on Geiger Grade."

Anyway, I am not sure if this is a productive conversation. The idea of directions is to get people where they want to go - whether they do so meeting all specifications and definitions in the traffic engineer manual doesn't matter in my opinion. I would concede that you are TECHNICALLY correct, but I don't see how that affects the way that roundabouts are currently drawn on Maps.

(does it really say my screen name is "djboje"?)

[Geoffrey Perez]

I would think the generation of the directions through the roundabout depend on whether the roundabout is specifically named (i.e. West Circle) and/or carries the names of the streets it's on (Main Street), or is unnamed completely (the roundabout). If the it did carry the name of one of the streets, straight directions (from Main Street to Main Street) may potentially omit the roundabout. Perhaps the traffic signs that can be added when editing intersections will warn drivers to slow down for the roundabout, but not necessarily have to say "head straight to keep on Main Street".

Yes, the placement of the intersection points is significant as well for the terminologies that will be generated (slight right, right, sharp right, etc).

--
You received this message because you are subscribed to the "General Map Maker" Google group.
 
- To post a public reply in this same thread:
Reply to this e-mail and your note will soon be posted.
 
- To post a note to the group on a *different* General Map Maker topic:
Visit http://goo.gl/DeX9Z and click on "New topic"
or
E-mail google-...@googlegroups.com
 
- To unsubscribe from this list:
Send email to
google-mapmak...@googlegroups.com
 
- To see additional useful options:
Visit http://goo.gl/GDnr and click on "Membership."
 
Thanks for using Google Map Maker! :-)

[Geoffrey Perez]

I guess it would make more sense to say "keep right to stay on Geiger Grade", wouldn't it? Reminds me of Google Maps directions at forks.

--

[Geoffrey Perez]

Also, I think sometimes in Google manually tweaks the directions on a case-by-case basis, where the placement of the intersection points would confuse the algorithm ("turn right on" vs "continue straight on" ---> "keep right to stay on"), or force a longer preferred route even when two parallel routes have the same priority (avoid separate carpool lanes/ramps to/from/on freeways), or take a shorter route with more turns to get on the freeway sooner: 

http://goo.gl/8CLIK

On Map Maker, the updated directions from the beginning take the closest ramp at Hawthorn Street, but the older Google Maps defaults to the ramp at India Street north of Glenwood Drive instead.

[djboge]

Sure, I could see it saying "keep right." I don't believe roundabouts have names, they're really only a type of intersection. A big traffic circle on the other hand could very well have a name.

[Flash]

I'll have to say that it really doesn't matter how you draw it, as it will get butchered afterwards by Google Reviewers that "know better" than yourself.

We have several new overpass/interchanges in my area, and provincial laws now require that such major reconfigurations use roundabouts unless there can it can be shown why it would be safer or necessary to do otherwise.  One near my house was finished a few months ago, though most of it has been configured and in use for almost a year. I have dozens of GPS traces and a 2011 aerial photo that I aligned in Google Earth and then drew the an overlay to use in Map Maker.  Just like the photos above, I had the roundabout in short, straight segments as this is how it is actually constructed and is how the lines on the road are painted in order to make traffic that is going "straight" through not have to turn much.  My overlay was very precise, tracing the exact centre lines of the roads.  I was not yet an RER when I drew it, so I had to wait for approvals often; plus there were corrupt segments that could not be moved and I had to wait for them to be deleted.  After 3 weeks I finally had an exact copy of my overlay into Google Maps.

Then the Google Reviewers that like to follow recent edits around moved in.  I did not know about them then, but they have hit any major new road that I have drawn since.  One would make a "tweak", then another made a different one, then another, etc.  I tried submitting feedback through the "Rate this edit" function, but they ignored me.  I tried fixing some of their changes, but yet another would come along and make new changes.  I tried using the feedback to ask for direction as to where I had gone wrong, I was ignored.  I asked how they could modify a route for which they did not have aerials, GPS traces, personal knowledge nor Street View; I was ignored.  Finally I have given up.  I'm sure in a few months it will finally morph back to configuration shown in the 4 year old aerial, one small edit at a time.

The sad part is the Street View car redrove the area 6 months ago, and the roads won't align with the pictures it took when they are published.

Load this overlay into your Map Maker and then zoom in on the southern roundabout.  I did have the roads precisely following the red traces at one time.  It seems minor, but the edits just keep slowly making it worse and worse.

http://goo.gl/ELDHX

[vandelay]

Agreed 100% with flash.  I have seen volunteer editors draw roundabouts with extreme accuracy, only to have Google reviewers come back and draw in something completely nonsensical in order to fit their desired template, even when satellite imagery shows that the original drawing was correct.

[vandelay]

Sorry for my typo, it shows up as "djboge", not "djboje".

 

As for this comment, I would reiterate that YES, you are turning right.  You are following road signs for the right turn.  There are countless cases of non-roundabout intersections where one must turn right in order to stay on a road with the same name, so the naming of the road is really not relevant.  What matters is the signs and pavement markings.  If that intersection had been built as a stop sign or traffic signal, you would clearly be turning right to stay on Geiger Grade Road.  The same is true with a roundabout.

 

This is very important for how drivers percieve roundabouts.  If a driver perceives it as an eight-legged circle with "exits", they will make mistakes that cause crashes.  When a driver makes a right turn, it's okay to enter alongside someone approaching from their left.  If a driver does this at a US/UK/NZ roundabout, they are very likely to crash because the car coming from their left is likely to be exiting.

 

But the main point isn't nomenclature, and language is fickle.  What's important is that the mapping matches the orientation of the lanes.  It can be easy to "fix" poorly mapped roundabouts by dragging the intersection points so they're essentially on top of one another, but I don't think that this is good mapping practice either.  If it looks like two intersection points are actually a single intersection, they usually are.

[Geoffrey Perez]

If only directions in Google Maps Navigation were like the ones from other GPS manufacturers with lane assist.

Yet, how would we be able to implement that in Google Map Maker when the number of lanes in nearly all road segments don't include turn lanes or are flat out wrong, as well as which lane is meant for which upcoming road? Having this would be useful for multi-lane roundabouts, forks and other intersections.

--

[vandelay]

Using the Geiger Grade roundabout as an example, here is how it would look if you draw the lines by centering them on the through (straight-ahead) lanes.  As with any other intersection, the line would be centered on the lane lines or lane centers, ignoring turn-only lanes.  So where there are two through lanes, the line follows the line between the lanes.  When there is only one through lane, the line follows the center of the lane.  If there were three through lanes, the line would follow the center of the middle lane.  Starting in the through lane of each entrance, the lines follow the path legally designated as the through (straight ahead) movement.

 

Drawing it this way is proper in every legal and engineering sense, and it will not mislead drivers.  What I cannot vouch for is how a driving device would interpret the directions, but I can be certain that it would not give any incorrect advice.  It just might not mention that the intersection happens to be a roundabout.

 

[Geoffrey Perez]

No worries on that last part. If the road segments that make up the roundabout have the segment attribute "Roundabout / Traffic Circle", it will be recognized as such. Segment usage is mentioned here: http://goo.gl/NGHJ7

Note that when updating anything on Google Map Maker involving directions, the directions won't propagate onto Google Maps until after a few months have passed. This is unlike the roads themselves that when published may show up displayed on the main Maps website in at least a few minutes.

--

[vandelay]

Here are three examples of "roundabouts" in Spain (http://g.co/maps/5tw64 and http://g.co/maps/2844f and http://g.co/maps/vcjd3).  Note that they are, functionally speaking, a completely different type of circular intersection than "roundabouts" in Australia or the USA.

 

In this Spanish example, unlike Australia, you must be in the outer lane of a circle road to exit, and you can pass an exit in the outer lane.  You can enter alongside traffic approaching from your left.  You do not need to choose any particular lane before you enter.  You may not exit from the "inside" lane of the circle.  Therefore, it is appropriate to draw it per the Google instructions, because they work well for this type of intersection.  This intersection is indeed numerous one way roads intersecting with a circular road.

 

In the US or Australia, the center island is still circular but the traffic rules and lanes are completely different.  US and Australian roundabouts are designed as a series of overlapping one-way roads.  If you pass an exit in the outer lane, you have committed an illegal maneuver and may cause a crash.  You will crash if you try to enter alongside someone approaching from the left.  You are normally allowed to exit from the inside lane.  You must choose the proper lane before entering.

 

Locals would call all of these "roundabouts" even though they are completely different types of circular intersections.  Therefore, one must consider the actual design when drawing it.  It is irresponsible and incorrect to try to map Australian designs to fit a Spanish template when they are completely different.

 

Here are some USA, Australia, and New Zealand roundabouts, for comparison.  http://g.co/maps/hp7v5 (USA), http://g.co/maps/g6z6p and http://g.co/maps/5hbfe (Australia), and http://g.co/maps/d5ch4 (New Zealand).  These are all mapped INCORRECTLY in that they do not agree with the lane design and traffic regulations.

 

[vandelay]

I'll offer another example, because it's important to remember that roundabouts aren't always round.  This one is located on Keystone Parkway in Carmel, Indiana, USA.  It is currenly mapped improperly, showing more intersection points than actually legally and funtionally exist.  Below is how it should have been drawn, with only four intersection points.

 

There are countless variations of intersections - What I'm objecting to is the idea that there is a one-size fits all solution of mapping all roundabouts to follow the Spanish model.  The map maker must consider which lanes are designated as through lanes and map them accordingly.  They can then select all the innermost segments and label them as "Roundabout/Traffic Circle" if desired.

 

 

[vandelay]

Here is the Airport Roundabout in Dublin, Ireland, if it were drawn correctly.  Again, the lines follow the lanes designated as through (straight ahead) movements, so the map and drawings agree with all traffic regulations and pavement markings. This one, like many others, is currently drawn incorrectly to resemble an eight-intersectiontraffic circle rather than a four-intersection roundabout.

 

 

[vandelay]

Sometimes it is much easier to visualize in a perspective view rather than a completely vertical one.  Below is an image of a two-lane roundabout in Woodbury, Minnesota, USA.  The first image shows how it SHOULD be mapped, according to the signs and striping and traffic rules.  It can be clearly seen in this image that this roundabout consists of four intersecting lines.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The image below that shows how it is currently configured in Google Maps, drawn incorectly as an eight-legged traffic circle.  It can be clearly seen in the perspective view that this is incorrect and fails to match the lane configuration.  Lanes lead out of the roundabout, not around it.  This incorrect mapping would suggest that entry is a double right turn and that one can enter alongside circulating traffic, which would cause a crash.

 

 

[delta fox]

you know what ? Create a Vandelay Map Maker and you will decide what you want as correct and incorrect. Now, if you are speaking of Google Map Maker, I understand that you don't agree with them, but I still understand nothing to why the actual way is incorrect, so I will stay with it.

[djboge]

I think that comment was uncalled for.

I completely agree that there is a distinction between a roundabout and a traffic circle from an engineering standpoint.

However, I don't believe that there is a problem with the way many, MANY, roundabouts (in the US) are drawn, although they are, indeed, technically drawn as traffic circles. They still achieve the goal of providing meaningful directions to someone driving a route that includes that roundabout. When driving, people don't follow the directions like a robot, they read signs and roadway markings and everything else too, and successfully navigate many roundabouts every day.

I don't see any reason that they can't be drawn the "correct" way either, and I did look around for a bit, trying to find an unmapped roundabout that I could try the new way on, just to see how it works, but I struck out. 

I guess the bottom line is that yes, he has a valid point, but its not as if someone is going to uproot every drawn roundabout in the US and redraw them correctly. I also see the desire to have the help documentation clarified, to note the difference between a roundabout and traffic circle, which is definitely a possibility, if they so choose. If not, then so be it. The instructions and such may be technically incorrect, but this is not nearly as critical of a problem as this thread has made it out to be. It's not as if drivers are barreling into each other at roundabouts every day because they're drawn a certain way on Maps.

[vandelay]

This comment is very immature and not constructive.

 

I am sorry if you do not understand the particulars, I have tried to list many examples to illustrate the differences.  This is not just "vandelay" speaking either.  I provided numerous links to different road agencies across the USA that also clearly state that a roundabout and a traffic circle, in this country, are not the same thing.  They are arranged differently and they function completely differently.

 

The fact is that there are many different types of circular intersections in different parts of the world, and they should be drawn to match the actual conditions.  Google's instructions are written to match only one such condition, and this is a disservice to anyone who uses the service.  We can do better.

[vandelay]

You are correct in that I am aware of no instance in which the arrangement in Google Map Maker has directly led to a crash.  That doesn't mean that we can't improve.

 

I do know, however, as an engineer, that when drivers get in crashes at roundabouts, it is because they assume that it is arranged like a traffic circle.  As engineers, we need to do better to make the differences clear to drivers.  Google Maps is one venue that can help to affect public perceptions because so many people use the service.  We're not helping by planting the wrong idea in people's heads any time they see a roundabout in Google Maps.

 

Assuming in the future that Google makes their mapping information available to makers of driving devices, this becomes even more important.  One such scenario is if a device maker programs the device to say "turn right at the roundabout" when the driver is actually supposed to be in the lane marked for left turns, because the device interprets the linework as being a set of right turns.

[vandelay]

I will also add that this is not strictly a USA issue.  A "roundabout" in Spain is configured differently from a "roundabout" in Australia.  A "roundabout" in Germany is different from a "roundabout in New Zealand.  Some are spirals, some are radial.  The difference is important.

[vandelay]

I found an unmapped one you can try: http://g.co/maps/5hj72 .  I'm fixing one just to the north, as best I can, but it will probably involve moving the existing incorrect intersection points so that they're nearly on top of each other.

This is an American style roundabout that should consist of four crossing points.  Those points are pretty clear from the wheel tracks in the satellite image, or you can use the actual lane centers.  Please let us know how it goes.

[djboge]

But moving the intersection points on top of each other won't solve the problem - they'll still be interpreted the same as a traffic circle, right?

http://goo.gl/7lDGC

[delta fox]

At least you now admit that Google instructions are valid in some cases ;-) 

The main difference between roundabouts in Germany and New Zealand is that one is clockwise, the other counterclockwise. I hope this is marked correctly on Google Maps...

[Geoffrey Perez]

How about this one in Mare Island, Vallejo, CA USA? http://goo.gl/F7E8G I guess a right turn from Sundance Avenue onto Azuar Avenue absolutely needs two intersection points in this case.

On Sat, Jan 28, 2012 at 1:49 PM, General Map Maker on behalf of delta fox <google-...@googlegroups.com> wrote:

At least you now admit that Google instructions are valid in some cases ;-) 

The main difference between roundabouts in Germany and New Zealand is that one is clockwise, the other counterclockwise. I hope this is marked correctly on Google Maps...

--

[vandelay]

They are certainly workable in some cases.  The part that is incorrect is that they say that a roundabout is the same as a traffic circle - That particular statement is always incorrect.  Roundabouts are not the same as traffic circles, and even "rounadbout" means different things in different places.

They all feature a circular center island, but they do NOT all feature a circular roadway.

[vandelay]

You are absolutely right.  This one's not a typical American roundabout in that it has five exit points and a larger than normal center island.  As an engineer I would have designed the southeast part of this one a little bit different but since it's a one-lane intersection it's not a big issue.  I just made a few adjustments to this one to make it more accurate, have a look and see what you think.

So yes, this one has some combined segments and more intersection points than most four-legged roundabouts.  But that's a good thing.  What I'm hoping to encourage is this kind of critical thinking, rather than having Google enforce an eight-legged Spanish/German template when it is not accurate for most US/NZ/Australian roundabouts.

[delta fox]

it was better before... two problems with your edits : 

- the turn segments (like this one) are displayed as if there was a turn to the right, then a sharp turn to the left where it connects to the road. That's not realistic. Previous drawing was better. If you really want to fit the imagery as close as possible, the intersection between the two turn segment has to be much further, away from the traffic circle, with a very small angle.

- instead re-drawing roads with a single intersection where it is possible, you moved a second intersection at the same place, with a ridiculously small segment between. It looks the same, but it may create data error. I think most guiding system will correct themselves, but the result is a mess and it will be very difficult to edit if there is a imagery shift or whatever.

As often, the better is enemy of the good ;-)

[vandelay]

This particular intersection is a red herring anyway, because it more closely resembles a Spanish roundabout.  It is a distraction from issue because Google's instructions work well for this particular junction.  But that is not usually the case for most American roundabouts, nor for other countries that use spiral designs (Australia, New Zealand, etc).

 

I made a small edit to address your concern.  I never intended to place those intersection points exactly on top of each other, I agree that doing so is not good mapping practice.

[vandelay]

You are correct.  I think it's better because it more closely resembles reality, but no it does not solve the problem.  Unfortunately Google does not yet provide an easy way to combine two intersections points into one.  The only way is to draw in new links, which triggers the Review process, which results in weeks of waiting after which Google staff redraws everything (incorrectly) anyway, often worse than before.

[delta fox]

even better when redoing to connect to existing intersection

BTW, Spain is not an example to follow for roundabouts, they have a lot that are really crazy like this one...

[vandelay]

You must have higher editing priveleges than I do.  Any time that I have tried to draw a new segment to correct a drawing (as you have done here), it triggers a review process and it never gets approved.  It sits there, untouched and unapproved for weeks, and then someone from Google redraws the whole thing anyway.  I would very much like to fix certain roundabouts by drawing in new links to connect to existing intersections, but I haven't had much luck yet getting those through the review process.

 

I'm not sure that this one should have been drawn as a single intersection though.  The arrow painted on the pavement suggests that this one is more like the Spanish design, where exiting the circle is considered a turn and remaining in the circle is a "straight ahead" movement.  But again, this is not typical of American roundabouts.

[djboge]

Trying to redraw a roundabout from one method to another is pretty much a lost cause. You're likely to corrupt one or more segments, besides having to wait for approvals. Besides, there's hundreds if not thousands of them out there drawn the "incorrect" way.

If you find a regular intersection and the imagery shows a roundabout - by all means redraw it however you want. Once you get to that baseline trust level where you can add roads, you'll only have to wait for approval on the final attribute change of the roundabout segments.

[delta fox]

Le jeudi 2 février 2012 00:07:28 UTC+1, vandelay a écrit :

You must have higher editing priveleges than I do.  Any time that I have tried to draw a new segment to correct a drawing (as you have done here), it triggers a review process and it never gets approved.  It sits there, untouched and unapproved for weeks, and then someone from Google redraws the whole thing anyway. 

Until recently, it was possible to delete or create most roads without moderation, whereas the slightest intersection move got pending moderation...

Anyway, you should use the best method, and if the edit remains pending too long, ask moderation in Review edits requests forum. 

 

[vandelay]

I've found some unmapped ones and drawn them correctly, per the satellite image, and by some luck managed to have all edits published immediately rather than having to wait for review.  Few things are more disheartening though than when a Google Reviewer who clearly has no local knoweledge comes along after everything has been drawn correctly, and then they draw in something that is completely incorrect.

They draw nice circles, I'll give them that.  They can draw a much more perfect circular shape than I can.  But they get the intersections all wrong, create misleading information, and it fails to line up with the satellite image.

[vandelay]

I need to vent a little bit now.  I've spent a great many hours drawing in some roundabouts that were not previously mapped at all.  Using photo overlays and actual project plans (which I have access to), I would draw them correctly from the start with four intersection points.  Unfortunately, Google reviewers then come in and completely redraw it within days, always incorrectly.  Once they add extra intersection points, they can't be removed.  The best I can do at that point is to drag those intersection points ridiculously close together.

 

It's sad that while some edits can go unapproved for weeks or even months despite being easily verified, reviewers apparently have time to go around redrawing and corrrupting things that are already correct.  I'm happy to volunteer my time, local knowledge, and engineering knowledge to improve maps especially when the safety of users is at stake, but what's the point if it's just going to be redrawn and corrupted within days?

 

Okay I'm done with my venting, but I might also be done making edits.

[vandelay]

Amen again to everything you said.  I recently drew in a previously unmapped roundabout, using an overlay, drawn with extreme accuracy.  Once drawn, I coded the inner segments as "roundabout/traffic circle" which I knew would trigger a review.  To my delight, it was approved.  Then within hours a different Google reviewer came in and completely destroyed everything.  I wish I could say that this was the first time that had happened, but that's definitely not the case.  What's the point of making edits if they're going to redraw it to be completely incorrect and even dangerous to users?

[Willy Hrachovina]

I am surprised nobody from Google has said anything about this topic yet. This is an important distinction and is likely something that needs to be addressed in the 

Help Guides, as well as with the reviewers once those changes are made. I know this kind of thing is frustrating, as I have done other kinds of changes that would get immediately reverted or changed to something else in a few days or so. However, if it is something that you think may not be right, bring it up to Google or start a discussion like you did here, and things might get changed.

I personally don't mind if they come back and make the circle more round, for illustrative purposes. However, I think that some of what is going on might have to do with how they illustrate divided roadways, and if the approaches have an island, they wish to illustrate it with two parallel lines. Just a thought, and while I can understand this can have a significant impact with how directions are generated, bringing it up here for discussion is the best way to get something changed.

[KaliJay]

I'm not sure those short segments really fix the problem. It's like you're just trying to hide the problem by moving the intersections closer together. Personally, I'd keep fighting it. I know the end result is going to be a mess because inevitably some of the edits will result in rendering glitches, but you can't just give up! Send feedback to Google about the reviewers making changes that aren't correct. 

[delta fox]

It would help if you provide links to where are the roundabouts you are speaking about.

[vandelay]

Below is one such recent edit. It was previously unmapped and did not reflect the recent roundabout construction.  I drew it using an overlay and drew it correctly with four intersection points.  When everything was finished and lined up, I selected the four segments that surround the circular island and change the segment usage to "roundabout/traffic circle" which I knew would trigger a review.  I added a note to the comments pleading with the reviewers not to redraw it.  After about 2 days, Google Reviewer Phani approved it.  Unfortunately, 2 hours after that, Google Reviewer "Karunasri" ignored my comments and redrew the whole thing as a eight-point traffic circle, which is completely incorrect.  Soooo frustrating.

[vandelay]

I agree completely.  I don't like moving those segments so close together, and it's definitely not the "Right" way to do it.  Unfortunately it seems to be the only way the Google currently tolerates because they insist on separate intersection points for the entrance and exit, when in reality they are a single intersection.  For what it's worth, I never try to put them exactly on top of each other, I always leave at least a token space in the right direction.  I wish Google provided a way to merge two intesection points into one.  Seems like that would be an easy thing to program.

[vandelay]

Thanks and I agree.  Hopefully this discussion will take root at Google.  To be fair, the instructions, in my view, don't say that roundabouts MUST be drawn like that; rather I think the instructions were trying to say here's how you can join roadways with a circular or oval loop when you need to.  Unfortunately, in an American, Australian, Irish, or New Zealand context, that's not what a roundabout is.  That's what a traffic circle is.  The terminology isn't important, but what's important is that they are not the same type of intersection.

 

On that note, I should add that I don't mind if they make the curves more smooth, but at a roundabout the "circle" should not be round because there is no single circular roadway.  Often times, the center island isn't even truly circular, especially if the roundabout has any exclusive turn lanes.  There will always, and should be, a slight corner point where the entrances meet the exit, as shown in the green-lined example in my initial post.

[Nels]

Hi Vandelay!

Just want to let everyone know, that we have been reading this thread.

We are taking the suggestions and investigating/discussing them! 

Thanks for all the in-depth analysis.

-Nels

[Flash]

That's great to know, Nels.  I am actually taking advantage of my day off and my exams being over, and am in the middle of an explanatory email regarding this issue to my guide.  Is there no longer a need for that?

[vandelay]

Thank you Nels, that is reassuring to know.  Again today (18 hours ago), another corrected roundabout was redrawn by a reviewer, incorrectly.

[Nels]

Hi guys!

@Flash: I think it would be good to have a compressed version of this thread sent to your guide. Please always feel free to send issues like this to your guides! They're here to help!

@Vandelay: A fix to this type of situation will take time, but rest assured we are working on it!

[KaliJay]

The roundabouts you are mapping, from a functional point of view, are no different from a fat, round intersection between two divided roads. Marking them with the roundabout/traffic circle attribute is just asking for trouble (at least as long as GMM doesn't make a distinction). I would simply re-orient the roads, leave the roads inside the intersection labeled normally (not unnamed like a traffic circle) and wait for a new policy from Google. 

I redid the one on Bailey Road for you and left it without the roundabout segment usage. I wasn't sure of the exact geometry since the imagery is still out-of-date, but made an approximation based on the edit history.

Does anyone see any reason this type of roundabout needs a segment usage at all? Does it add anything to the intersection? We don't want Google Maps to say "enter the traffic circle" ahead or "take the 3rd exit" or such...which it will do if we specify the segment usage. Directions for the intersection on Bailey Road are still based on the traffic circle so even after changing it the directions are wrong...and will remain wrong for several weeks. :(

[Flash]

The problem is they are different from intersections in that they don't have lights or stop signs; instead you must yeild to the people already in the circle.  You can get in the circle and turn continuous circles and everyone entering must yield to you; I've done this to get a GPS trace :)  And they are, at least in my area, marked as roundabouts.  You need the roundabout marking so that someone isn't confused by directions; they see the roundabout signs, they see they roundabout, thus Directions should say "At the roundabout...".  The really big difference between them and European traffic circles is that they generally have approaches that make a sharp turn to slow you down as you enter, they usually are not perfect circles, and that if you take the first right then you barely brush the edge of the circle.

I'm wondering if perhaps we do want to have eight intersections, and then place two of them extremely close together.  It would be up to Google, but here's my thoughts:

If we just do four and you are taking the first right, you might get confused by "Turn right at the intersection".  It might be technically right, but you can see a traffic circle and it is not mentioned.

If there are eight, with every two really close together (almost touching), then it would look like it is designed in Maps, but Directions could then say "At the roundabout, take an immediate right".  That little section of roundabout between the two intersections would be there to trigger the directions to say it so it makes sense to drivers.

As Nels asked, I'm writing this up for my RER Guide to be passed on further; what are the expert opinions on what I'm proposing?

[vandelay]

The circular island is a useful landmark, but it shouldn't be necessary to say "at the roundabout" in driving directions any more so than it would say "at the big oak tree".  They are simply landmarks.  Saying "turn right" is more correct in a legal and functional sense than saying "take the first exit" (assuming right-hand driving).  Saying "turn left" is more correct in a legal and functional sense than saying "take the third exit".  "Exit" implies that you need to be in the outer lane, but you are not supposed to change lanes in a roundabout and you must be in the correct lane before you enter.

 

When you say you can get in the circle and turn continuous circles, that's not always necessarily true for a roundabout the way that it is for a traffic circle.  Some don't even have continuous circles.  There are several of these "peanut" roundabouts in Carmel, Indiana, USA along Keystone Parkway.

 

For what it's worth, I am a traffic engineer that spends a great deal of time dealing with this topic, and I have presented at international conferences to other engineers regarding these distinctions.  That said, we shouldn't get hung up on terminology because there are different languages and terminology sometimes varies by countr even when the language is the same.  What matters is that at one type of circular junction, it's okay to enter alongside circulating traffic and exiting is a right turn.  At the other type, exiting is a straight-ahead movement, and entering alonside circulating traffic will get you into a crash.  The only way to ensure consistent and correct directions is to draw it with four intersection points.

[vandelay]

I don't see a need to identify segment usage for roundabouts or for traffic circles.  But I try to stay within the rules whenever possible.

Thanks for the redo on that.  I can use my overlay to adjust the alignment to a more precise leve.  But unless there's a shift in GMM policy, it'll probably be redrawn as a traffic circle again by someone else.

 

And I agree that the segments around the circular island should not be unnamed.  It's entirely appropriate for them to carry road names or road number shields when applicable.

[KaliJay]

I respectfully disagree on the need for the Traffic Circle segment usage...at least with regards to how Google Maps provides directions today. 

The traffic circle segment usage is important in Google Maps if you want directions like this (see http://g.co/maps/hpjyu):

1. Head east on Orchard Walk

2. At the traffic circle, take the 3rd exit onto Blackstone Way

Without the segment usage we get crazy navigation directions (see http://g.co/maps/xdvvu):

1. Head west on Martin Meadow Way toward Gray Fox Circle

2. Turn right onto Gray Fox Circle/Martin Meadow Way

3. Turn left to stay on Gray Fox Circle/Martin Meadow Way

4. Turn left to stay on Gray Fox Circle/Martin Meadow Way

5. Turn right onto Gray Fox Circle

That's ugly. This intersection needs the traffic circle segment usage and the names removed. The ones vandelay is working on (e.g. Bailey Road) do not need this segment usage since the left turn is just a left turn. 

We need some type of lane assist in Google Maps...

Unfortunately it's a complicated situation that a lot of mappers and reviewers will have trouble dealing with. Luckily the segment usage will always require moderation so we don't have to school every mapper....just the RERs and GRs. :)

[KaliJay]

vandelay can correct me if I'm wrong here, and maybe you already know this, but just to make sure we're clear on the distinction here...

In the intersections vandelay is discussing, you cannot continue driving around in a circle if you enter the roundabout in the outer lane. Doing so will cause an accident since the inner lane is permitted to "go straight" at any point (thus exiting the roundabout). 

I actually have this crazy idea of drawing all of the lanes individually with the proper turn restrictions...but I'm sure that would freak everyone out....

[Flash]

It can vary, as vandelay says you must read the signs before you enter and be in the correct lane.  Sometimes the outer lane has the option of exiting or continuing around, and the inner lane must continue around.  Other times the outer lane must exit and the inner lane is allowed to exit or continue around.  These two scenarios can be within the same roundabout; it has to do with which roads are minor and which are major.  If I know the roundabout and I have the option of either lane to accomplish my route, I try to be in the lane where I won't get hit if someone doesn't pay attention.

[vandelay]

This isn't strictly a discussion of multi-lane intersections.  Even single-lane roundabouts are generally designed to function as four intersection points.  The arrows on the pavement next to the circular island stipulate that the driver has a choice between a straight-ahead movement (exiting) and a left turn (not exiting).  This is the USA standard in the Manual of Uniform Traffic Control Devices (http://mutcd.fhwa.dot.gov/pdfs/2009/pdf_index.htm) and is substantially similar for Australia and New Zealand, and I'm sure many other countries.

 

So in these situations, no one is forced to exit (though that's often the case for multi-lane roundabouts).  But if you don't go straight (exiting), you have legally made a left turn, according to the signs and pavement markings.

 

At single-lane roundabouts, this issue isn't terribly important.  You could draw them either way and, although Google's current way is incorrect, it's not going to cause anyone to crash.  But at a multi-lane roundabout it might cause a crash, and at a minimum it feeds into driver misperceptions.

[vandelay]

Absolutely correct.

[delta fox]

@Kalijay: you're right that this intersection needs 'Roundabout/traffic circle' segment usage. I don't think removing the names are needed, though; if there is the correct segment usage navigation will say 'take the 3rd exit at roundabout' or 'pass through x roundabouts' when the road name is the same before and after the roundabout, whether there are names or not on the roundabout segments (but this hard to check because of the months delay
in directions updating)

For all other aspects (4 or 8 intersections, perfect round drawing or not so round, etc.) I think it's a pure loss of time to discuss this in GMM, since it will have no effect at all on what navigation says. Yielding or not when entering/exiting) is important to avoid crashes, of course, but if drivers are looking into their GPSs instead of looking outside for traffic signs or painting, crashes will happen whatever there is in their GPSs :-( That's not a mapping question.

Remember that GMM is worldwide and cannot accommodate every local variation that only experts would understand (at best). Here are a few weird roundabout where I'm living : 

- a Squareabout (or traffic Square ?) which technically is a roundabout but it fails the Roundabout attribute in Maps: http://g.co/maps/q3uev

- twin roundabouts: http://g.co/maps/8jgzs or even smaller: http://g.co/maps/e25jp or here: http://g.co/maps/aptpe where instead of just saying pass through the roundabout, ir says:

1. Head southeast on Rue Paul Bellamy
		52 m
	2. At the roundabout, take the 1st exit
		10 m
	3. At the roundabout, take the 2nd exit onto Rue Paul Bellamy
		51 m

Rue Paul Bellamy

- place de l'Etoile in Paris http://g.co/maps/5s2b6 is an exception because traffic on the circle yields priority to incoming traffic - but my opinion it would need the 'traffic circle' segment usage too, because it's easier to count the exits that to read the road names at each exit !

[Nels]

Hi everyone!

Just wanted to let you know, we've taken all the suggestions in this thread, and made sure the proper teams see them. Thanks so much for your ideas and suggestions!

-Nels

[vandelay]

" The roundabouts you are mapping, from a functional point of view, are no different from a fat, round intersection between two divided roads."  --Exactly.  I could not have said it better myself.  That is the normal American design per the Manual on Uniform Traffic Control Devices (http://mutcd.fhwa.dot.gov, chapter 3).  

Is it possible to remove the traffic/circle roundabout attribute and simply draw them as fat, round intersections?  Any time I have tried to clear that attribute by selecting the segments and then changing it from "roundabout/traffic circle" to "---", I just get a message that says "there were no changes".

[Saikrishna Arcot]

You'll have to remove the Segment Usage each segment at a time, instead of using the Select a Segment tool.

[Flash]

One of my roundabouts that I drew as a four point intersection is being deconstructed into a 7 pointer right now (eighth road does not exist as it is a one way freeway entrance on the west side).  Because D-Map is involved and she is always polite, she contacted me through a different edit she was reviewing first and alerted me, then started and asked if I minded the changes.  When I protest, she politely states this is the way Google wants it and asks if I understand and mind.

[vandelay]

Interesting.  I have yet to receive a heads-up before one of my edits gets wrecked, roundabout or otherwise.  Hopefully this thread helps spur some changes to what seems to be interpreted as "the way Google wants it" since it's potentially hazardous, and at best misleading.

The instructions as written, however incorrect, never sounded like "policy" to me, but rather just an easy way to achieve a desired result when mapping certain types of traffic circles.  I don't know if this is the document that D-Map is referring to or if there is some underlying Google policy as well.  Either way, plenty of Google reviewers see it an absolute must-do, even when the intersection isn't circular.

I came across this one, which blew my mind.  Clearly just a wide intersection, and more square than round, yet Google maps shows false spaces between what's clearly a four-pointed intersection.  Directions would tell people to "take the first exit" rather than simply turning right, and at no point would the directions tell a driver to turn left.  Check out the turning signs in street view as well.

http://g.co/maps/8rp5p

[vandelay]

I'm not sure when it was changed, but I see that the original referenced Google support document, located at  http://support.google.com/mapmaker/bin/answer.py?hl=en&answer=1208728, appears to have been changed.  Unfortunately, it has changed for the worse.  It now states "Each road connected to the roundabout should be connected separately to ensure proper directions routing. The images below show the correct way to connect roads to a roundabout" and it proceeds to show the image below.  Unfortunately for me, Google's new update reinforces the incorrect notion that a roundabout should consist of eight intersection points, even worse than before.  This continues to falsely suggest that entry and exit is a right turn.

[Flash]

I am being told by D-Map for my particular roundabout that they separate them like that as they want drivers to know they must yield at the roundabout.  I had said that if I was coming north and turned left, I never enter the roundabout.  She states that physically I am, and thus must yield to traffic to prevent accidents.  I am telling her that ALL traffic must yield when turning, it is the law.  I'm also pointing out that traffic remains separated in my particular roundabout even if a right turner fails to yield.

[vandelay]

Unfortunately D-Map's explanation doesn't hold any water.  It's true that entering drivers must yield, but the way Google says to draw them, you'd only have to yield to the outer lane because you'd be making a right turn, and that's just not true.

 

When passing through a roundabout, you're making a crossing movement, not a right turn, so you need to yield to BOTH lanes.  See the picture below.  D-Map says to draw everything like the intersection on the left, where it's okay to enter alongside.  But that's not the case at a roundabout because you'd crash (see the drawing on the right).

 

If you draw them as four-pointed intersections (rather than eight-pointed), then it's just like drawing an intersection between two one-way roads where one of the roads must yield.  This reflects the reality of a roundabout.  There is (normally) no common segment between an entrance and an exit.

 

 

[vandelay]

I'll also add the following, which is based on a drawing from the US Manual on Uniform Traffic Control Devices.  This mirrors Australian practice.

 

 

[vandelay]

It was good while it lasted - The Bailey/Radio roundabout has now been redrawn as an incorrect eight-pointer by shruthi.

[vandelay]

Below is an example of an actual engineering drawing that includes roundabouts - Note that in all cases, entry and exit exist as a single intersection (look at how the black lines intersect).  In each case, they function as a single intersection between two one-ways.  The most clear example in this drawing is on the northwest portion of the northern roundabout.

[Flash]

I'm sorry if I wasn't clear as to what D-Map meant.  She meant that they want the roundabout designation to be in place even if you are turning right immediately, so that you get a warning of it coming up and you yield to traffic already in the roundabout.

[vandelay]

Thanks for the clarification, but I'm still not sure that D-Map's position makes sense.  "Roundabout/Traffic Circle" is a segment attribute but it does not pertain to entry yield control.  There are roundabouts that are signalized (mostly in the UK, Ireland, and Australia) where entering traffic has the right of way when the light's green, and there are traffic circles in New Jersey where the entering traffic sometimes has the right-of-way (http://g.co/maps/egx5c), and there are rotaries where entering traffic can "merge" side-by-side with circulating traffic where neither one really yields.  All of these use the same "Roundabout/Traffic Circle" segment attribute and they are all mapped the same in Google Maps, despite the fact that they are all structured (and operate) completely differently.  So I don't see why using the Roundabout/Traffic Circle segment attribute would instruct a driver to yield.  If it does, then it's wrong in many of these other applications.

Google shows intersections between one-way roads all the time (in non-roundabout contexts) and Google Maps works fine for those applications, I'm not seeing why this situation would be different.

[assa...@gmail.com]

If entry yield control is important, isn't that what the circle traffic sign is for?

Sent via BlackBerry from T-Mobile

---

From: General Map Maker on behalf of vandelay <google-...@googlegroups.com>

Sender: General Map Maker on behalf of vandelay <google-...@googlegroups.com>

Date: Fri, 6 Apr 2012 11:05:58 -0700 (PDT)

To: <google-...@googlegroups.com>

ReplyTo: General Map Maker on behalf of vandelay <google-...@googlegroups.com>

Subject: Re: Drawing Roundabouts - Google instructions are incorrect.

--
You received this message because you are subscribed to the “General Map Maker” Google Group.
- To post a public response in this same thread, please reply to this e-mail.
- To find a list of Frequently Asked Questions, visit- http://goo.gl/yuVdY
- To post a new topic, visit- http://goo.gl/DeX9Z & select 'Post a question'
- To unsubscribe from this list, manage your subscriptions here: http://goo.gl/oGfnJ

Thanks for using Google Map Maker! =)

[vandelay]

I'm not sure I understand your comment - Yes, roundabouts feature yield at entry.  Yes, there is a "roundabout" traffic sign available in GMM.  I don't see why that sign couldn't be used at a correctly-drawn four-pointed roundabout.

 

There's also a yield sign available in GMM - This would make just as much sense, if not more so, than the roundabout sign when used at an intersection between two one-way roads.

 

Regardless of which signs are assigned (yield vs. roundabout), if any, I'm not grasping why it would be important to draw a space between each entrance and exit, when in reality such space does not actually exist.

[assa...@gmail.com]

I was referring to the intersection traffic sign for traffic circles in GMM. That should be enough for direction alerts for circles in Maps.

Sent via BlackBerry from T-Mobile

---

From: General Map Maker on behalf of vandelay <google-...@googlegroups.com>

Sender: General Map Maker on behalf of vandelay <google-...@googlegroups.com>

Date: Fri, 6 Apr 2012 11:55:45 -0700 (PDT)

To: <google-...@googlegroups.com>

ReplyTo: General Map Maker on behalf of vandelay <google-...@googlegroups.com>

Subject: Re: Drawing Roundabouts - Google instructions are incorrect.

--

[vandelay]

Here's a screenshot of how I would *recommend* doing it, as shown at one of the few correctly mapped roundabouts that I am aware of.  I think you could use the roundabout sign where the yield sign is, but I think the yield sign is more correct.  I'm not sure what instructions the Roundabout sign provides to the driver, but a yield sign is what actually exists in the field.