My bad on the overall wording of the article, then. In context, especially with the changes, it's fine. The reason I say "typically paved" is that throughout most of the country the preferred roads will be paved—typically, not always. We can change it to something like "typically but not always paved," perhaps.

There are certainly reasons to avoid roads with poor surfaces whether paved or not, but there are also perfectly valid reasons to avoid unpaved roads altogether. Dirt and gravel alike cause similar problems particular to loose road surfaces—things fly around, whether rocks or dirt or dust, and get all up in your undercarriage and also all up on your car. Maybe you have to pick up a client for business and the 5 minutes the gravel road saves you doesn't counterbalance the 10 minutes you have to spend at the car wash. Or maybe you just like your paintjob.

Point being that there are valid reasons to want to avoid roads with a poor surface and valid reasons to want to avoid unpaved roads. Many of the concerns with dirt roads which are intrinsic to the dirt surface are also present with other types of unpaved roads, whereas "poor surface" is not necessarily intrinsic to any particular type of road, simply more likely with some than others (which is why hard surfaces exist). I see two reasonably valid interpretations of the 'dirt road' type beyond just dirt, and the latter (1) is a more logical connection, to me, and importantly (2) does not require any value judgments or arbitrary decisions on a road-to-road basis.

I understand that the situation is quite different out west, but the purpose of this page is to provide the general rule, which states/regions can modify if necessary—because not every state will set its own standard. The better general standard is the one which is not subject to individual interpretation. Modify as you need, but don't necessarily play that up in the general article.

I think that captures it very nicely. Well done, and thank you.

Something like that might be okay, although that is what this section tries to accomplish already:

Waze's definition of "highways" includes all these, but it also includes other roads that are not necessarily a part of any numbered highway system. In other words, think of the dictionary definition of "highway" ("a main road, esp. one connecting major towns or cities") rather than the legal definition of the term.

That's fine, and I'm fine with that. I didn't mean to say "that is already in there and that is enough," just wanted to point that out in case you missed it. I do think some language like this should be added, and I like your proposal, but I would like to think of the right place(s) for it in the article.

jdeyoung, I'm not sure if this really clarifies it much.

twilde's suggestion is more to the point—the idea is to allay new editors' protests that "these aren't highways!!!" in town. I think it's a good start.

As for a "when to violate FC" section, I agree something more is needed—in particular, a section on "editing for continuity". I meant to write this section a year ago, but real life happened and here we are. As it stands, "continuity" is mentioned exactly once in the Road types/USA article. This is shameful on my part, because it's so important.

"Continuity" is the principle that roads should be typed for their clearly-intended use even where the FC map may not explicitly classify a particular segment. This is not a large-scale thing; this kind of thing occurs within the context of an interchange or intersection, generally. Continuity is the same principle we have used for years in determining the type of an AGC or roundabout. But we have not explicitly written guidance for this context.

This example is based on something I saw on a real FC map somewhere. I think it was in New York.

Here is, for purposes of discussion, an FC map:

http://i.imgur.com/wu7arYal.png

Direct use of hybrid FC would lead you to this:

http://i.imgur.com/9ji0xQgl.png

But there's a problem. You can tell by the signage on the ramps that this so-called "local road" is intended to carry traffic between I-3333 and US-4567. So you need to edit for continuity, so that MH-level traffic can move as intended between MH and Freeway, thus:

http://i.imgur.com/DGOZAwRl.png

Again, this is not a new concept, it is how we have been doing AGC types for a very long time. We just need to be careful that we are using it at the interchange level as well. And so we need to put it in the wiki...

t0, you are not wrong in saying that violation of hybrid FC should typically have a firm basis. That said, there are situations like these where not perfectly matching hybrid FC is not really a violation of hybrid FC, but simply an enablement thereof.

I'm not sure what you're proposing here. You don't pick a color in the road type drop-down when you're editing the map, you pick a road type by name, some of those names are "Major Highway" and "Minor Highway", and sometimes those names are found on roads that "aren't highways." Hence the confusion.

In other words, the confusion entirely centers around the road "highway" as seen in the editor. We can't remove the word "highway" from the road type drop-down, so it'll always be there, we can't get past that. So we need to hit the issue head-on and explain in no uncertain terms "not every 'Highway' is a 'highway' the way you're thinking of it".

I'm OK with removing the dictionary definition of highway, but only if it is replaced with something else more to the point that the "M___r Highway" road types in Waze often end up being used for roads that are not numbered highways. Maybe in a different section of the article, maybe as an infobox, I don't know, but it's a common hangup that needs to be addressed.

Sort of like,

"Yes, all Principal Arterials are at least Major Highway, and all Minor Arterials are at least Minor Highway, even if they're not what you might call 'highways'."

but less rude.

Kartografer wrote:As for FC continuity, can we be specific about which roads get upgraded to provide connections? I like sketch's illustration, but what criteria determine whether a road like this that connects an interchange to another highway gets upgraded to the same type of the highway. Must it have signage to that highway and be under a certain length?

That should be Major Highway. Not to say that you will never ever get a route that way, of course. But you probably won't be routed that way for routes longer than a couple dozen miles, so on such routes where this course might otherwise be the best way to go, you will get a route, but you won't get the optimal route.

Signage is the most important thing to look for—the best "tell", if you will. You need those segments to get between I-70 and SR-49, and it's clear the DOT has intended that bit of road to be used that way. As for distance, well, it depends on the situation. Usually when you see something like this the portion of road that needs to be bumped up for continuity is only a few blocks long, maybe up to a mile or so, but maybe not. They can maybe be a couple miles in rural areas, so I'm not really comfortable proposing a distance cap on continuity editing. But you need something affirmatively indicating that that is the intended purpose of that road; it can't be based on subjective feeling, like "this is a shortcut I know," it must be something more, something objective. Signage is a good indicator of that.

DwarfLord wrote: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.

What's going to happen is the road type is going to be deprecated and eventually removed, and everything in it is going to be turned into a Street with the Unpaved check box checked. That's Waze's current plan, anyway.

What are we supposed to do? I don't know, I guess nothing yet. I have not tested whether "Avoid dirt roads" in the app will avoid things flagged as "unpaved" that are not set to the "unpaved" road type. So we can't recommend that, say, unpaved state highways are set to Minor Highway with Unpaved checked, even though that's the endgame here.

Many are clamoring for the ability to leave "4x4" as a road type and to give the app the ability to use 4x4 as a vehicle type for routing, as many parts of the world contain both unpaved-but-maintained and unpaved-and-not-maintained roads, the latter of which can save a lot of time but do require a 4x4 to travel. As of yet, staff is not very receptive to this idea.

Whether or not they do implement that dream of ours, it's too soon to change substantive guidance re: the checkbox unless we do some routing tests on the checkbox.

4x4 would make most sense as a vehicle type (a la Taxi or HOV-3).

Brian, the issue at hand is essentially what we can/should tell editors to do. In a perfect world, during this transition, app & routing would honor both the road type and the check box, so we editors could start using the "unpaved" box alongside PS/mH types where appropriate for hybrid-FC reasons. But we can't do that until we can trust the box.