I think you misunderstand the need for the error. Itâs not just to show hidden segments, itâs because of known navigation errors when segments share end-nodes.
what would be superb is if validator was able to suggest people to âdowngradeâ or âupgradeâ a road type based in road classification rules and Highlightâs script speed info .
This would be quite easy to do. But it is not something that we would want where roads are classified by function not speed.
Perhaps the branch to take here is to make an api for the Validator so that people can customize what validations they want. It should be easy to insert your function CustomValidations() into the array of functions that Validator calls for each segment. People could script and share all the âvalidationsâ they like.
Further plans
Guys, let me know if you are interested in the following features:
User interface internationalization - i.e. translation of buttons, checkboxes, tips etc. on your language. Would it be useful or English is OK?
Separating translations from the main script. The main script will be slimmer and faster, while you will be able maintain translations on your own. I suggest to extract the following features:
a) UI translations (optional, see 1)
b) translation of check titles, problem descriptions, howtos and links localization
c) enabling/disabling checks for the country
d) managing exceptions and check parameters
Let me know what you guys think before I start to implement the functionality.
I would also like to hear if the script startup time in v0.6.2 is still an issue for you.
I currently get an indicator that a median u-turn segment is too short when it is actually a one way cross street that intersects with a split road. No direct left turn is enabled from the split road onto the cross street, so itâs not technically a u-turn segment. Would it be possible to disable this when the street name of the âu-turnâ segment is different from the intersecting road (and not âno streetâ or âu-turnâ as well), when there are more than three segments intersecting at a junction, or some better solution? On tapatalk, so no permalink right now unfortunately.
Iâm not sure what the problem would be on a one-way terminal loop, if any, but knowing which way to go from a dead-end isnât really a problem.
Where I have seen numerous problem reports from segments which share endpoints is with simple parking lots and drive-thrus and simple freeway parking areas. I experienced the first personally in that same location. There have been a few threads on it; I donât have time to dig any of them up right now.
I was thinking about kind of âlanguage packâ, so local communities could maintain the pack on their own (i.e. add/correct translations, add/remove exceptions and enable/disable the checks).
Sorry, I would appreciate a permalink. Iâm looking into the code and the situation you describe is impossible :?
Yes, I think localizations would be helpful, as it would widen the user circle (not every editor is fluent in English). The translations can probably be handled by string lists, which could be fed to you by local persons or groups in each country. In particular, I like the idea of having localized links to wiki/howto references.
About 2c): I suggest to have a default enabled/disabled check list for each country, where each user can then individually edit his list. A reset-to-default functionality should allow fixing screwed-up configurations.
Here is another idea to be discussed (or not ): checking average speed on a segment (donât know if you can get it like the Color Highlights script do) and highlight those with excessive speed regarding the road.
I have URs that appears with weird detours being proposed and the explanation seems to be one or several segments with illogical speeds registered. The cause of this is segments that have been re-used and modified, sometimes generating a false average speed.
Like 150 km/h on a ramp (with the Freeway attached to it being limited to 90), and Waze tells the driver to exit the Freeway just to re-enter on it the next ramp (without any trafic jam or whatever on the Freeway).
Since you wonât be able to know what the correct speed should be for each segment, you could do a general check and mark all the segments with, for example, a speed over 150 km/h. Or 130 km/h but excluding Freeways. Or whatever seems logical in your country regarding a type of road.
The type of warning is to be determined (orange or red since it affects the routing).
Iâve seen numerous reports of problems with one-way terminal loops. The problem is that waze ignores the one-way property of the road, so will appear to route either way around the loop. You correct it by putting two extra junctions in the loop.
I have some small issue on the report of the area on the permalink.
At the end of the report says:
Summary
WME Validator has checked 635 segments and reported 2 errors (3â°) and 8 warnings (13â°).
3% of 635 errors is not 2 errors, and
13% of 635 errors is not 8 errors.
Percentage is being calculated with a big error (Reported percentage is 10 times bigger than real)
As you can see Validator has aloud me to reduce significantly the errors of the area (over 200 when started fixing it, I didn´t check percentage before starting).
I hope you can fix this small issue, it is not a bug, is just a small issue and has nothing to do with the proper functionality of the error detection.
The is a bug introduced in 0.6.1 with âfixed few bugs with new and deleted segmentsâ. At the moment Validator reports ~2 times less segments (and hence ~2 times less errors) during the map scan. No problem with the highlighting, i.e. it highlights all of the errors on the map.
The issue will be fixed in the next version. Sorry for the inconveniences.
Now that I think it through, I believe that it was most likely a connectivity issue with the cross street segments that was causing the âu-turnâ warning. The issues are reported in separate sections of the Validator report, but I clicked on the u-turn warning and likely fixed the glaring red connectivity issue at the same time. The u-turn problem was what stayed on the brain because it is more unique, whereas Iâve been fixing tons of connectivity problems recently.