Page 184 of 236

Re: Validation

PostPosted: Tue Feb 04, 2014 5:24 pm
by sketch
dbraughlr wrote:Rule Walking Trail elevation = -5 is arbitrary. I did not see where anyone came forward to explain why he cares.

I did, two pages ago. "Some communities have been using Walking Trails at elevation -5 to signify railroads as a matter of editor policy."

So while no, it shouldn't be enabled for all regions, there are places where it is a useful check.

I can't fathom any reason an actual walking trail would be set to elevation -5.

Re: Validation

PostPosted: Tue Feb 04, 2014 7:13 pm
by sketch
dbraughlr wrote:That isn't an explanation of the error. Of what is this a useful check?

Railroads are not currently displayed in the client. That's why editors in some areas started marking railroads as "walking trails", so they'd be displayed in the client. But now we are told by staff that railroads will soon be displayed in the client, so all those -5 walking trails should be changed to railroads. Did I leave some of that out?

As a matter of editor policy, all non-drivable road types (including railroads, runways, boardwalks, stairs, and walking trails) have elevation set to -5 per established convention. -5 indicates that the segment is not to be attached to the drivable roadways "to prevent false system reporting that think the roads should be connected.".

I don't know of this convention, other than with railroads. Perhaps it is not so widespread as you think. In my area, pedestrian boardwalks are marked using normal elevation rules.

Re: [Script] WME Validator 0.6.0 (BETA) / 04.02.2014

PostPosted: Tue Feb 04, 2014 7:52 pm
by sketch
Where?

Road Types (USA) does not mention elevation for non-drivable roads other than (1) it must be different from any road it crosses, and (2) it should be -5 for railroads.

https://www.waze.com/wiki/Road_Types_(USA)#Non-drivable

Re: elevation -5 must be used exclusively for railroads

PostPosted: Tue Feb 04, 2014 11:14 pm
by sketch
dbraughlr wrote:Yes, I quoted from the Railroad section. Of course, -5 is guaranteed to be different from any road it crosses.

OK, and I don't doubt that it isn't guaranteed. But that doesn't mean it's the rule. You're extrapolating a new rule from two existing rules. "Set every undrivable road to a different level than every road it crosses" and "set every railroad to -5" do not combine to mean "set every undrivable road to -5", because not every undrivable road is a railroad.

Walking trails at elevation -5 to represent railroads, on the other hand, was a rule (in some places), and is now an obsolete rule. Hence the check (which should be limited to some places).

Re: [Script] WME Validator 0.6.1 (BETA) / 05.02.2014

PostPosted: Wed Feb 05, 2014 10:21 pm
by sketch
^ (here, in response to Alan/Timbones)

Consider someone going to one of the buildings near the gate, after the gate is closed. If Waze doesn't ignore restricted roads entirely, having the restriction on the whole segment would make Waze indifferent to which direction it came into the segment, and would be just as likely to send you through the closed gate as it would be to send you through the complex (the "possible" way). If Waze does ignore restricted roads entirely, it'll send you to R806 instead of North Rd (inside the complex), which is unhelpful because those buildings are inside the fence.

Re: [Script] WME Validator 0.6.2 (BETA) / 07.02.2014

PostPosted: Sat Feb 08, 2014 3:23 am
by sketch
dbraughlr wrote:I think it should not report any unless one of the junction angles is 0. I see the error when the junction angles are 90 degrees or more.

I think you misunderstand the need for the error. It's not just to show hidden segments, it's because of known navigation errors when segments share end-nodes.

Re: parallel paths to the same endpoints

PostPosted: Mon Feb 10, 2014 2:12 am
by sketch
dbraughlr wrote:Let's consider first a one-way terminal loop.

What is the known navigation error in this situation?
How do you correct it?

I'm 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 isn't 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 don't have time to dig any of them up right now.

Re: [Script] WME Validator 0.7.1 (BETA) / 20.02.2014

PostPosted: Thu Feb 20, 2014 7:40 pm
by sketch
Is it possible to make a report that includes every street segment with a particular name? For instance, we name U turn segments "U turn" in the New Orleans area, but many have been named "U-Turn" and "U Turn". Trying to standardize all of them is a heck of a task, but being able to search "U-Turn" in Validator would make a big difference.

Toolbox helps to some extent, but having to comb at Zoom 4 to load street type segments takes much, much longer than a Zoom 0 or 1 scan.

Re: [Script] WME Validator 0.7.1 (BETA) / 20.02.2014

PostPosted: Thu Feb 20, 2014 10:54 pm
by sketch
Toolbox does not "scan" at all; it can only see what I can edit at a given zoom. The scan feature is Validator's great strength over other plugins, and it would be uniquely suited to this particular task. Not to mention it'd be a trivial addition to Validator, whereas adding a scanner to Toolbox would be a big change.


Sent from my iPhone using Tapatalk 2

Re: [Script] WME Validator 0.7.0 (BETA) / 19.02.2014

PostPosted: Fri Feb 21, 2014 5:13 pm
by sketch
Regarding "(Closed)" / "(Construction Zone)" / "CONST ZN":

The "CONST ZN" suffix was a relic from a time before text-to-speech. It's pronounced "Const Zee Enn", which makes no sense. Adding the parentheses adds a pause, and now it says what it means, and it's easier to read, too. Also, when applied to closed roads, "CONST ZN" isn't really descriptive. I wouldn't expect the standard to change again unless Waze gives us a construction zone flag with a little digging man icon. One change to the Wiki in 3-4 years isn't too bad ;)

As such, the Validator should show results for anything containing "Construction Zone", "Closed", or "CONST ZN". The first two as a reminder, and the last so it can be changed.

berestovskyy wrote:
sketch wrote:Is it possible to make a report that includes every street segment with a particular name? For instance, we name U turn segments "U turn" in the New Orleans area, but many have been named "U-Turn" and "U Turn". Trying to standardize all of them is a heck of a task, but being able to search "U-Turn" in Validator would make a big difference.

You mean a check which reports all of those errors (i.e. "U-Turn", "U Turn" and so on)?

That would be nice for my purposes, but I don't know if "U turn" is used at all outside my state, or even metropolitan area. So I don't know if it'd be an appropriate addition to the entire script for the US. What I was thinking was that I could put in a name (whether "U-Turn" or "* North *" or whatever else), run the report, and get all the streets matching that name in the report.