Page 2 of 3

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Sun Aug 17, 2014 8:48 pm
by NorfolkMustard
I drove this roundabout today, coming south on the A404. Still not giving a numbered exit, voice or visual, guidance on approach.

In my view it would be fine for under-the-wood to be a separate exit. It's a surprise when you see it, but when you do then the exit number waze has given you will make sence

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Thu Jun 26, 2014 8:35 am
by Superfraggle
It's should be avoided, but there are legitimate cases such as here

https://www.waze.com/editor/?zoom=7&lat ... layers=901

Where you are not allowed to take the first exit, when coming west down the A2, but you can go round the roundabout.

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Thu Jun 26, 2014 9:35 am
by Timbones
Twister-UK wrote:I thought sharing roundabout nodes was still a no-no due to the risk of triggering a 0th exit error?
No it doesn't, it just messes up the node counting.

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Thu Jun 26, 2014 12:05 am
by Twister-UK
AndyPot2 wrote: I believe one solution would be to connect Under the Wood to the same node as the A308 Marlow Road. I think that will put the exit count back to 4 which should satisfy most Waze users. Looking at the road geometry I suspect that Under the Wood was originally connected to the same node as the A308 but has recently (quite correctly) been disconnected and that is why it is now mapping differently.
As the editor responsible for the recent change, here's my reasoning behind it...

In the previous layout, Under the Wood was connected to the southbound one-way segment of the A308, so anyone navigating to UtW would be given slightly dodgy directions to take the nth exit onto the A308 and then turn left, rather than simply taking the nth exit directly into UtW as the real road layout requires. This in itself wouldn't be that big a deal.


More of an issue for me, and the reason I made the change, is that when I'm negotiating a roundabout with directions to leave at the nth exit, I count the actual exits as I drive round rather than counting the nth exit on whatever signage is present leading up to the roundabout and then trying to match what the sign showed with what I see as I proceed around the roundabout. Sometimes the signage presented at each exit doesn't match up with the signage present on the approach, sometimes there isn't any exit signage at all, and sometimes I've simply forgotten what the approach signage said I ought to be seeing on the exit signage :oops:

Counting the actual exits on the other hand generally is pretty foolproof. Granted, there are a few exceptions where you might find an exit stub built in readiness for further roadbuilding that hasn't yet taken place (or never will), or a no-entry road joining the roundabout whos geometry gives the impression that it could also be an exit if you miss seeing the no entry signs or road markings. However, in my experience you're more likely to come a cropper relying on the signage matching up all around the roundabout than you are by counting the visible exits, and in this particular example the UtW exit is sufficiently clear as an exit in its own right that failing to include it in the exit count seems more wrong than leaving it out just to appease whichever muppet at the Highways Agency decided that UtW didn't even deserve the tiniest of stubs on any of the approach signage.

Note also that every alternative mapping source I've checked so far shows UtW as a seperate exit off the roundabout, and as several of those sources are used by other navaids/route planners then it's not just Wazers who get given exit counts that match up with the actual exits but not the approach signage. Not saying that just because everyone else maps it that way that it's the correct way to map it, but given the real world layout of the roundabout I think this is one of those cases where we ought to say stuff the signage, it's what's on the ground that counts.

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Thu Jun 26, 2014 6:30 am
by Twister-UK
ditchi56 wrote:
AndyPot2 wrote:one solution would be to connect Under the Wood to the same node as the A308
Here's an example of a roundabout with two roads connected to the same node, if you want to play and test the exit counts. I think it has the effect you want.
I thought sharing roundabout nodes was still a no-no due to the risk of triggering a 0th exit error?

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Thu Jun 26, 2014 12:56 pm
by Twister-UK
Timbones wrote:
Twister-UK wrote:I thought sharing roundabout nodes was still a no-no due to the risk of triggering a 0th exit error?
No it doesn't, it just messes up the node counting.
Good to know, thanks.

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Wed Jul 09, 2014 12:42 pm
by Twister-UK
I've now moved Under the Wood onto the same node as the A308 exit, which should restore the "correct" exit count in the client instructions without also preventing correct routing into or out of UtW (as was the case in the layout prior to my earlier edit). Let's see what sort of feedback this revised layout generates...

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Wed Jul 23, 2014 9:18 pm
by Twister-UK
OK, so something isn't right here :? Following the original merging of the Under the Wood and A308 exit nodes, the client was giving me 0th exit errors. I checked the roundabout and noticed one segment was defined A-B whereas all the others were B-A, so decided to rebuild the entire thing just to be sure the data was clean.

As there have been several tile updates since this last bit of tweaking, I'm assuming the roundabout now in use by the client and livemap is this latest revision and not the earlier, visually identical, version. Yet I'm still seeing the 0th exit error in the client and also in the route instructions returned by Tim's route checker script...

Help!!!

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Mon Aug 18, 2014 6:45 am
by Twister-UK
NorfolkMustard wrote:I drove this roundabout today, coming south on the A404. Still not giving a numbered exit, voice or visual, guidance on approach.

In my view it would be fine for under-the-wood to be a separate exit. It's a surprise when you see it, but when you do then the exit number waze has given you will make sence
Given that this was how I'd mapped the roundabout back at the start of this thread, I'd be only too happy to stick UtW on its own exit again, thus both matching the real-world geometry of the junction and clearing up this unexplained 0th exit error.

Except we'd then be back to square one as far as having to deal with countless URs from people who don't like the idea of the exit count in Waze not matching up with the exit count they've calculated from looking at the signs on the approach to the roundabout.

If we could resolve this 0th exit problem then I'd be happy - although it wouldn't be a wholly accurate rendition of the roundabout, it'd be the one which generates the least amount of URs.

Since the roundabout shouldn't be behaving like this, and since it IS behaving like this even after having been completely torn down and rebuilt with new segments, perhaps it's something the Waze team could look at?

Re: Is it Acceptable to Map a Roundabout Incorrectly?

Posted: Fri Sep 05, 2014 5:07 pm
by Twister-UK
I've been cultivating different test roundabout designs up in a secluded bit of the Shetlands, and I think I've got one which will give the desired results here - just need to wait for the next tile update to check the fine tuning I did this morning... If that looks good, I'll rebuild the A404 one to match, and hopefully that'll be the end of it.

Until some bright spark posts a UR complaining that there are actually 5 exits on the roundabout :lol: