Page 18 of 57

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Sat Feb 13, 2016 8:08 pm
by DwarfLord
Looks like we've converged! One last word (literally) I'd like to add. It is common for editors to use the Dirt Road / 4X4 Trail type for private roads (i.e. gated and on private property) if they happen to be unpaved, instead of using the Private Road type. The more I think about this, the more I think it is not optimal. But I don't want to add a whole new paragraph about it -- just a single word, in red:
proposal wrote:The Dirt Road / 4X4 Trail type has the unique property that Waze users may ask not to be routed over it. Users may ask to avoid it for all through routing with the settings option "Dirt roads - Don't allow", or to avoid it for through routing longer than 300 m (984 ft) with the option "Dirt roads - Avoid long ones".

Because of this property, this type typically represents public side roads that some fraction of local drivers habitually avoid due to surface quality. In metropolitan or other urbanized regions, this generally means unpaved (dirt, gravel, crushed rock) roads, or roads in uncommonly poor condition by local standards. In other areas, roads with unpaved surfaces may be essential routes and thus necessarily set to other types such as Street, Primary Street, or even higher, as if they were paved.

Check your state page for details on whether your state follows unique guidelines for dirt roads, or contact your regional coordinator for further guidance.
Hopefully this looks good to everyone at this point!

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Sun Feb 21, 2016 1:40 am
by DwarfLord
The update to the Dirt Road / 4X4 Trail section is complete. I'm a bit puzzled by all the non-breaking spaces in the title/link but I guess it has to do with formatting in the colored lozenge. I didn't change anything from what was there in that respect.

Thanks everybody for your support :)

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Sun Feb 21, 2016 3:50 pm
by DwarfLord
voludu2 wrote:Most of the section on "Walking Trail" needs to be removed because of the change in how Walking Trail behaves.
I would disagree (and not just because it was such an ordeal to write that section in the first place that I don't want to see it dumped).

The two main points of the WT section are (helpfully boldfaced in the article):

1. Connected or not, the Walking Trail type should never be used where effects on local routing are not desired.

2. Editors should not map Walking Trails, or any other road type, for the sole purpose of encouraging non-driving Wazers.

I have seen nothing to suggest that either of these is no longer true.

Yes, Waze has indeed changed things. Apparently many editors somehow thought that the WME's "Walking Trail" road type was meant to represent -- can you believe it? -- walking trails. Strange, but true. So, if I have the story straight, Waze felt the best way to protect drivers from editors misusing the Walking Trail type to create walking trails was not to change the name, but rather to lobotomize the routing behavior. So now, when editors see "Walking Trail" and mistakenly think it should be used for walking trails, they won't do as much damage. Yay!

Meanwhile we've lost the one tool we had for efficient routing to destinations that can be accessed from multiple roadheads, like long pedestrian malls or railway stations where one can cross the tracks on foot but not by car.

Unfortunately, despite having lost one of their most useful features, Walking Trails are still dangerous. In the SF Bay Area we continue to see deleterious effects on routing from unattached WT. I sometimes call these floaters the "ghost in the machine", because of the weird things that happen when they are around.

Likewise, as far as I know, cyclists and pedestrians who use Waze near drivable roads still corrupt the database, and we still want to do everything we can to discourage that application.

So I believe the main points of the article are still absolutely valid and relevant and would disagree with the perspective that most of the article needs to be removed.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Sun Feb 21, 2016 11:10 pm
by DwarfLord
voludu2 wrote:Here are things that I think need to be removed:
* The longish section on applications
* The example image showing exactly how routing is affected -- it doesn't work exactly that way any more.

It seems to me that unattached and otherwise inaccessible (red arrow or TBSR) segments of all kinds cause bizarre routing problems (see the routing mess at Secaucus Junction). Do detached walking trail bits cause different effects than other routable and non-routable road types?
My understanding is that WT are still useful for things like what we have in the applications as long as we make the WT twice as long as used to need to be. In other words, last I knew, Waze's new WT routing takes you to the junction with the WT provided the destination is closer than half of the total length of the WT segment. So I suppose the new text should say to make WT twice as long as before. How to do that if the actual WT on the ground is not that long, I don't know, extend the WT into spirals or celtic knotwork or something.

True, Waze has busted WT's use for routing into a destination from either end. As far as I know they don't care, so that wonderful capability is probably never coming back. So yeah, that application should go.

However we are still using WT with addresses in order to route people. Is there something wrong with that application? It seems to work, for the moment at least.

Last I knew, detached WT really do indeed still have negative effects on routing rather like the cartoon diagram that people want to get rid of. I've seen it, quite recently in fact. Maybe nobody else is seeing this but I sure am. I'm not sure why the interest in toning down the warnings about WTs, because I think they are still evil. I swear to you I've seen it.

How detached WT compares with other detached routable road types, I don't know. Maybe they're all bad, but that doesn't mean we should tone down the warning about WT, if anything that means we should make the warning prominent in the global wiki.

I'm sorry if I sound frustrated, it's not with you or the editing community, it's with Waze. It is hard to write guidance when the developers apparently have no interest in either product stability or communication.

Here's my perspective: as long as you keep the two main points of the WT article -- 1. We do not map WT if we don't want effects on routing, and 2. We do not map anything for the sole purpose of encouraging non-driving Wazers -- I am delighted to leave the rest to someone who is willing to try to sort this mess out.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Mon Feb 22, 2016 6:03 am
by DwarfLord
voludu2 wrote:The rest of what I've heard about the changes to WT routing makes me think they are not currently suitable for pedestrianised urban shopping districts, train stations with 2 parking lots, or rest areas in freeway medians with parking lots on both sides
Correct (as far as I know). They are still useful for routing to HNs on non-drivable streets, and to PP and RPP on non-drivable streets as long as they are closer than 1/2 the total segment length. Those are useful qualities, even if one does have to make the WT twice as long as before for the latter to work. But, unless they've changed things again, WT are no longer useful for efficient routing to destinations with multiple roadheads.
voludu2 wrote:I'm not sure if pedestrian boardwalk is currently better than WT for capturing walking-wazer traffic data next to roads.
My last understanding, unless things have changed again, is that PB is still much more inert than WT. This was my experience from a recent example in San Francisco. If the only purpose is to soak up data pollution from improper wazers, PB is still the better choice.
voludu2 wrote:I'm willing to give it a shot.
Go for it!

Your comments seem to suggest that you might have access to information that indicates that everything we determined about WT and PB is no longer correct and we must throw a great deal away and start from scratch. I know some things have changed but I'm not aware that everything has changed. If you do have inside information then I must step aside, because I don't! If not, I'd like to recommend we start with things we know for sure are wrong. It is a shame to throw away hard work if it's still accurate.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Mon Feb 22, 2016 7:00 pm
by DwarfLord
I have no further information on that either. Would be nice to know status and plans...oh well.

That other thread just referenced (Road Types (USA) - Walking Trails) may be the most appropriate place to continue the conversation about modifying this part of the article.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Sat May 14, 2016 4:20 pm
by DwarfLord
The text of the runway/taxiway section as it stands now was largely my choice of words, but based of course on the previous text and on public conversation about best practice. Another editor added the image later, and I noticed that it contradicted the text, but I needed to move on to other things at the time.

I would welcome and support modifications to the fly-in community map as discussed above to bring it into compliance with the text, followed by an update to the image in the wiki.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Sun Aug 07, 2016 3:21 pm
by DwarfLord
Late to this conversation, but as a general rule I heavily favor including some rationale. This is true for me in most contexts, because my brain seems to be wired for symbolic rather than literal thinking. But I think it's especially true in Waze, even for a wider audience.

One reason is that many elements of guidance in Waze are based on reverse-engineering and previous app or WME behavior. When better information arrives, or behavior changes, such guidance becomes obsolete. It may even become counterproductive.

Another reason is that literal guidance, while it is always crystal clear to its author, is not necessarily clear to the reader. I have seen the wiki misinterpreted many times, including by people who were absolutely adamant that their interpretation was correct, except it wasn't.

At such times, the more people who know why the guidance is what it is, the sooner the community can recognize and accept that it needs to be updated or clarified.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Tue Aug 16, 2016 3:33 pm
by DwarfLord
Did Waze simply open up their code and do a global substitute of "Unpaved" for "Dirt Road / 4X4 Trail"? Including both the road type in the drop-down menu and the checkbox? Or something more elaborate than that?

Speaking of which, I'm still unclear on just what that checkbox does, if anything. And even more unclear why, when Waze changed the name of the road type, they changed the name of the checkbox to exactly the same thing. Typically UI design leans away from marking multiple controls with the exact same words, but I'm sure Waze has a good reason.

Re: Road Types (USA) – comprehensive overhaul of drivable ro

Posted: Wed May 14, 2014 5:38 am
by Fredo-p
Just want to put my info on this subject regarding Arizona.

I have been in contact with the Arizona Dept. of Transportation (ADOT) FC Coordinator. So I have been relaying all the information that he has informed me of with the AZ editors. I have also gone through the ADOT FC manual and created a pdf that summed up the way ADOT classifies its roadways and why. After some debate, we came to the agreement to utilize what ADOT has provided after I explained the way ADOT defines each road type.
vectorspace wrote:(2) Rural areas of AZ currently seem biased to be less highways and more primary streets and streets. So the major arterial between distant cities can be primary streets. It seems under-differentiated.
This was a problem when it came to deciding how to implement the FC system into Arizona. The major issue was that in the ADOT manual, it stated some cities/towns don't follow ADOTS classification system. At first, this seemed like a nightmare. But when I made a chart of the listed cities that followed their own classification system it made sense. These municipalities followed their own rules to an extent. One major reason is due to the population size and roadway network running through them. This actually made things easier. Since these municipalities didn't see a need to recognize minor or major types, they grouped them as one (Arterial and Minor). Some didn't even qualify for arterial so they did not use this type.

The other discovery that we found was that many roadways that we once thought of as just a street were actually classified as a PS. This changed many areas of the Phoenix area and began to show how ADOT routes traffic based on traffic density.

Some of the classifications have cause us to think about changing them. However, the final result is that we are to follow the ADOT FC map and, only then, make changes as needed. Since there was/is so much information about how the roadways classification works, I took challenge to read all the manuals and help translate how to make it happen in AZ. The top editors like nnote, ply8808, and itseric, have helped tremendously in making it happen and giving their opinion on the matter.

How this ties in with other states is something that the top level editors from each state will have to figure out.

The biggest thing to take away from all of this that has been mentioned in other posts is the way the FC's are named and identified in Waze. Don't get fooled just because a type says highway. It's just a classification name. What is important to remember is what that FC name stands for. Focus more on how each state classifies Arterials, Collectors, and Local roadways. If the data that the state provides seems outdated, get in contact with them. If this still proves difficult, the last option is to take the states DOT manual and implement the changes yourself. This would take time but prove more accurate versus waiting.