[Script] WME Validator v2021.3.5 (+ PLACES BETA)

[dbraughlr]

I would say, don't mark it as an error when:
- the non-driveable is a dead-end road.
...

We consider that someone has arrived at his destination when he reaches the limit of the drivable roadway. What is a scenario where we would attach a non-drivable road to a drivable road? Such a situation would lead to Waze attempting to route over the non-drivable road. What do you have in mind? Is this for walking directions or bicycles?

[dbraughlr]

So I'm not sure if Validator should suppress issues at the "not closed" road

A report for obsolete CONST ZN markers will be added.

I agree. I see two independent options to filter the report:

1. Omit any roadway with name containing "closed".
2. Omit any roadway with name not containing (at least) one of {"closed", "construction", "CONST"}.

(Neither, either, or both options may be selected.)

[dbraughlr]

When a USA street name starts with a compass indicator [ENSW], the next word should start with a capital letter. "W laRue St" should be flagged with error #94.

[dbraughlr]

... (lets say less than 30 grades)? A kind of warning, "Turn with too low angle, pls check"

I think that your proposed rule could be applied based on the lowest class of the roads forming the angle. Streets, dirt roads, and parking lot roads may have such sharp turn angles. Freeways never do.
Ramps junctions probably are where this type of error occurs most.

I consider going straight to be an angle (course deflection) of 0.
Turning left or right in a perpendicular intersection is an angle of 90°.
You are talking about a turn greater than 150°.

After you have checked a turn that is greater than 150° and see that it is fine, then what?
U-turns in bowtie intersections sometimes are very tight because of the need to have a nearly parallel interception angle to get the correct turn instructions.

[dbraughlr]

When a USA street name starts with a compass indicator [ENSW], the next word should start with a capital letter.

We can go further and check if every word in the street name starts with a capital letter. But at the moment #94 is checking just the beginning of the street name.

Negative. The rule applies only to the first letter of the first word. A compass indicator prefix is not the first word of a name.

[dbraughlr]

Would also be nice to have a way to search for all elevation not set to "Ground" in an area where there are no tunnels, bridges or overpasses.

It might be nice, but Waze has no way to know this. There is nothing to validate.

[dbraughlr]

Validator flags the segment with a 'turn too sharp' at the north end when that end has the restricted turn (yellow) 24/7, but passes it when the turn is forbidden (red).

Make it red. If someone makes it green, it will fail validation.

[dbraughlr]

I meant, search for segments that aren't set to ground level currently, but that would probably be another script!

Are you talking about Level Highlighter?

[dbraughlr]

... IMO it's better to create new more comprehensive check with a list of allowed lowercase words (to, a, the, de, la, etc) and report any other lowercase word no matter its position in the street name.

"De La Vina St" → pass.
"Norte de La Vina St" → pass.
"E del Paso Blvd" → error.
"Van de Vanter Ave" → pass.
"W de Vanter Ave" → error.
"S De la Cruz Blvd" → error.
"N Via de las Cruces Hwy" → error.

[dbraughlr]

In both cases to check if a segment intersects with another, we have to compare it with every other segment on the map = very slow. Sorry, it won't be implemented unless we found another solution

Comparisons can be greatly pruned by segregating roads into smaller map squares. Roads which are completely confined to a small square need be checked only other roads which intersect the square it occupies. No segment needs to be checked against all others. Most segments need to be checked against only other roads in the local area.

Create buckets to divide the map into equal sized squares, say 1 km².

For each segment on the map, determine which square(s) it crosses and add it (or a portion of it) to the bucket for those square(s). Repeat until all segments are distributed to the buckets.

For each bucket, compare each segment (or portions of a segment) against all those segments in that bucket.