Parking lot roads I add disappear

[mapcat]

3) Mapping parking lot roads (correctly!) will help eliminate issues like this:

I've been seeing a fair number of system generated errors where people leave waze running while going through the drive through (bank or food) or getting gas. I think that it's a very reasonable thing to do, and when I see it, I'm mapping in a parking lot, so we don't get more reports of it.

Are you really eliminating any issue at all when you're doing this? The "map problem" identified isn't a system-generated error; it's an error with the system generating it. Waze developers realize that when a driver goes off the road grid, there's a potential error, so they thought about it for 5 minutes and came up with a half-assed solution. The results of shifting the problem to editors are thousands of unimportant parking lot and private road segments littering the map, more work for the routing servers, and well-meaning editors drawing even more needless segments wherever there's space for them.

Solution: throw an error whenever a user goes off road for the first or second time at that location. If an editor marks the error "not identified" both times, assume the editor has some intelligence and stop identifying problems there.

[gettingthere]

Are you really eliminating any issue at all when you're doing this? The "map problem" identified isn't a system-generated error; it's an error with the system generating it. Waze developers realize that when a driver goes off the road grid, there's a potential error, so they thought about it for 5 minutes and came up with a half-assed solution. The results of shifting the problem to editors are thousands of unimportant parking lot and private road segments littering the map, more work for the routing servers, and well-meaning editors drawing even more needless segments wherever there's space for them.

Solution: throw an error whenever a user goes off road for the first or second time at that location. If an editor marks the error "not identified" both times, assume the editor has some intelligence and stop identifying problems there.

Very well stated.

We had some direct interaction with the Waze developer who was working on these system generated errors. Unfortunately we could not convince them that these algorithms still need work, and that that the mapping activity that is occurring do to these poorly implemented system generated errors is not necessary.

Same issue with the same error that occurs at many wide intersections where the error leads you to believe that you must drop in a right turning lane. The routing generally works perfectly without the turning lane being mapped.

Another point here is that Waze is not evaluating the map imagery when they generate these problems. They only have the GPS points to evaluate vs. mapped roads. That is another reason why a human needs to be involved to filter these, to compare the GPS points with the imagery.

Out of the hundreds or thousands of these 'Road Missing' System Generated messages that I have worked on, I could count on the fingers of two hands how many really needed a road to be added. Based on my results, if Waze doesn't want to make this more useful they should just turn this problem type off. At least for base mapped countries. Maybe it's of higher usefulness in countries where the map is being entirely build from scratch.

[FreeMan12]

Are you really eliminating any issue at all when you're doing this? The "map problem" identified isn't a system-generated error; it's an error with the system generating it. Waze developers realize that when a driver goes off the road grid, there's a potential error, so they thought about it for 5 minutes and came up with a half-assed solution. The results of shifting the problem to editors are thousands of unimportant parking lot and private road segments littering the map, more work for the routing servers, and well-meaning editors drawing even more needless segments wherever there's space for them.

Solution: throw an error whenever a user goes off road for the first or second time at that location. If an editor marks the error "not identified" both times, assume the editor has some intelligence and stop identifying problems there.

Very well stated.

Weird... This part of your comment only appeared once I hit "reply/quote"... (and, now that I've submitted my response, it appears... )

We had some direct interaction with the Waze developer who was working on these system generated errors. Unfortunately we could not convince them that these algorithms still need work, and that that the mapping activity that is occurring do to these poorly implemented system generated errors is not necessary.

Same issue with the same error that occurs at many wide intersections where the error leads you to believe that you must drop in a right turning lane. The routing generally works perfectly without the turning lane being mapped.

Another point here is that Waze is not evaluating the map imagery when they generate these problems. They only have the GPS points to evaluate vs. mapped roads. That is another reason why a human needs to be involved to filter these, to compare the GPS points with the imagery.

Out of the hundreds or thousands of these 'Road Missing' System Generated messages that I have worked on, I could count on the fingers of two hands how many really needed a road to be added. Based on my results, if Waze doesn't want to make this more useful they should just turn this problem type off. At least for base mapped countries. Maybe it's of higher usefulness in countries where the map is being entirely build from scratch.

Anyway...
I'm not here to disagree with mapcat's statement that this isn't the idea solution. My point, though is that this is what I've done:

It took me maybe 5 minutes, you can see (I've hit Shift-A to show not-allowed turns) that I've fixed up all the turn-ins and turn-outs (it's onto a 2-way, double yellow divided road), and we won't get any more issues here.

While this may not be the ideal, ultimate, theoretically perfect solution to the issue, it will avoid system generated issues until Waze updates the logic in their system to eliminate them.

[CBenson]

Are you really eliminating any issue at all when you're doing this? The "map problem" identified isn't a system-generated error; it's an error with the system generating it. Waze developers realize that when a driver goes off the road grid, there's a potential error, so they thought about it for 5 minutes and came up with a half-assed solution. The results of shifting the problem to editors are thousands of unimportant parking lot and private road segments littering the map, more work for the routing servers, and well-meaning editors drawing even more needless segments wherever there's space for them.

Solution: throw an error whenever a user goes off road for the first or second time at that location. If an editor marks the error "not identified" both times, assume the editor has some intelligence and stop identifying problems there.

I agree with the sentiment and your solution. In my experience, however, if you continually mark problems like this at the same location "not identified" eventually the ign team comes in a makes the change for you (sometimes in a manner that causes additional problems). It seems that this is not merely a problem with an automatic routine that wasn't fully thought through. The waze people seem to believe that many of the parking lot and right turn missing roads really are issues. I'm leery about ignoring these problems until the volunteer community and the paid editing community are on the same page.

[gettingthere]

I understand but I think that my and mapcats' points are that we shouldn't get those issues nor have to do any mapping to resolve issues that don't impact the routing. Or that this issue appears only once and we can make the determination that roads don't need to be added and the 'issue' can go away.

It's wasted effort by the volunteer editors with no/minimal added value to the map or the routing.

Yes, Waze thinks that they are providing value, but generally they are just wasting our time.

[FreeMan12]

<snip> the ign team <snip> the paid editing community <snip>

I've been around a while, sometimes more involved than others, but...

What is the "ign team" and are there paid editors?

[gettingthere]

Yes, Waze pays a company to provide map editors to supplement the volunteers to get the map 'fixed' quicker.

But if we were not wasting our time with these system generated road missing messages by adding Parking Lots maybe more real work that impacts the routing would get done! Add to that fixing the work of map editors that don't know our countries or roads and sometimes cause more issues than they fix.

[CBenson]

I understand but I think that my and mapcats' points are that we shouldn't get those issues nor have to do any mapping to resolve issues that don't impact the routing.

Sure, but I don't know the intricacies of the routing routine. I don't know which edits impact the routing. Certainly leaving a drive-thru or a gas station unmapped can affect the routing. I'm not saying it should, just that the waze solution to these issues seems to currently use the map editor.

[invented]

<snip> the ign team <snip> the paid editing community <snip>

I've been around a while, sometimes more involved than others, but...

What is the "ign team" and are there paid editors?

If you happen across the most horrible sh*t edits you've ever seen, click on the segment and 9 times out of 10 it will have an ign_[name](2) updater on it. Apparently they get paid, while those of us with a vested interest in creating the best local Waze experience we can, don't.

Parking lot segments that aren't part of malls shouldn't have to exist. They add busywork, have problems with navigation because consensus is that they shouldn't be named, and only serve to (probably) help alleviate phantom traffic jams. There has to be a better way!

[CBenson]

9 times out of 10 it will have an ign_[name](2) updater on it.

Well it used to be much worse when you would see ign_[name](3) and the segment was locked. Now I usually see ign_[name](1) which is a vast improvement.

There has to be a better way!

But until there is . . . .