usaca: California State Highways

[oscar]

Jim can speak to resource constraints, but now that our project is expanding internationally to at least five continents, the size of the database and associated processing burdens could still be issues. Even without Tim, there has certainly still been nudging to keep point density under control, including specifically for some of the largest route files in usaca.

Besides, the one-mile "rule" is not hard and fast. If there are closely-spaced intersections with numbered routes, they both go in. Same if we need closely-spaced points for shaping. And point density in densely-populated urban areas tends to go up, as it was even under Tim's management. But if we don't add some intersection with an unnumbered road of non-obvious importance (not even for shaping, which is the reason for many otherwise totally unimportant waypoints), and there's a point within a mile, the user who would like that intersection added usually will lose a mile or less in the user stats and maps. No big deal, IMO. That happens to me all the time.

[Jim]

No resource constraints at this point.  I think we just want to keep routes from getting too overly dense, so things are not too cluttered on the maps and tables.  That said, if there are a few points likely to be useful that happen to be close together, I'd be in favor of including all of them.

[mapcat]

One way to solve the problem of too many unnecessary waypoints would be to replace some of the visible points with invisible ones, and reserve the visible points for truly useful junctions.

[oscar]

One way to solve the problem of too many unnecessary waypoints would be to replace some of the visible points with invisible ones, and reserve the visible points for truly useful junctions.

That runs counter to what I'd understood was the current SOP, to replace hidden waypoints with visible ones where possible (which I've done with some routes covered in neroute2's point requests). If we're going to have a waypoint anyway, such as for shaping purposes, making it visible would make it available to users with little extra cost, even if the potential benefit is uncertain at best.

[michih]

if there are a few points likely to be useful that happen to be close together, I'd be in favor of including all of them.

Seconded!

[oscar]

if there are a few points likely to be useful that happen to be close together, I'd be in favor of including all of them.

Seconded!

OTOH, requested points for intersections with unnumbered routes, less than 0.1 mile from existing waypoints (there were two of them here), seem generally not to be worth the extra work. I think "not worth the extra work" would sometimes also apply to requested points not quite so close to existing points, though different team members might draw the line differently WRT the systems they maintain/develop.

I did add two points on CA 82 for connectors to nearby freeways, that were less than a half-mile from existing points. I've tried to include such points as a general rule, but didn't do so for that route when I initially edited the file.

[mapcat]

That runs counter to what I'd understood was the current SOP, to replace hidden waypoints with visible ones where possible (which I've done with some routes covered in neroute2's point requests). If we're going to have a waypoint anyway, such as for shaping purposes, making it visible would make it available to users with little extra cost, even if the potential benefit is uncertain at best.

AFAIK, that's been hinted at in the forum, but if there's ever been a clear directive to favor visible shaping points over hidden ones, I've not read it. It seems kind of pointless to clutter the files with named points at culs-de-sac and paper roads that no one will ever use.

[Duke87]

AFAIK, that's been hinted at in the forum, but if there's ever been a clear directive to favor visible shaping points over hidden ones, I've not read it. It seems kind of pointless to clutter the files with named points at culs-de-sac and paper roads that no one will ever use.

How is this clutter, though? You are making a point either way. The total number of points in the file is not impacted by such a policy.

If there is a cross street of some sort at the approximate location you are making a point anyway, you may as well make it a visible point because even though the likelihood of it being used may be low, it takes no extra effort to make it visible and no extra bandwidth to allow it to be visible.

That said I would only apply this to actual intersections. So, cul-de-sacs would be fair game, and so would dirt roads... but "paper roads" would not be.


Meanwhile don't dismiss the usefulness of such points - maybe someone has only driven the route as far as that cul-de-sac because their second cousin they went to visit that one time lives on it. Maybe they turned around at the intersection with that dirt road because it happened to be right over the line of  a county they wanted to clinch, or because it was at that point that they realized that had missed a turn and needed to go back the other way.

I'm certainly in favor of having "make a point visible if there is a nearby intersection to use" be official guidance if it isn't already.

[oscar]

^ I agree with the above. But the one kind of "clutter" from including minor visible waypoints, that would be reduced if they were hidden, is in the waypoint list shown in the HB.

In addition to people who used a minor intersection to go to their destinations, or to turn around, there will be people who go through that intersection, do a U-turn elsewhere (or break down, or go off-road, or turn around at another minor intersection), but they can use the labeled intersection to claim some more mileage in their list files.

In instances where I've U-turned right after passing a county line sign, while I can't ask for a county line point (unless there is a legitimate county line road intersection there), I have asked for a point at a nearby intersection that other users might find helpful too, so I could claim some of the mileage and my maps would hint at how I snagged that county.

[mapcat]

How is this clutter, though? You are making a point either way. The total number of points in the file is not impacted by such a policy.

Like Oscar said, every visible point makes the waypoint list showing in the HB longer, and adds a waypoint symbol to the map.

If there is a cross street of some sort at the approximate location you are making a point anyway, you may as well make it a visible point because even though the likelihood of it being used may be low, it takes no extra effort to make it visible and no extra bandwidth to allow it to be visible.

Of course it takes extra effort. I don't know what basemap you use in the editor, but I typically start with OSM Default. To add an invisible waypoint for shaping, you find the place where the point makes the shape conform to the limits and it's done. To add a visible point at a random unimportant road, you need to open GMSV at that location to check to see if (A) there is actually a road there and (B) if its name shows up on a street sign. If there is a road but no sign, you need to find some other way to confirm the name. Google's name for it isn't any more likely to be correct than OSM's. And then you need to see if you've already used that name somewhere else in the file, and if so, either modify one of them or move one to a different road that might the job. Which puts you right back to checking GMSV.

Meanwhile don't dismiss the usefulness of such points - maybe someone has only driven the route as far as that cul-de-sac because their second cousin they went to visit that one time lives on it. Maybe they turned around at the intersection with that dirt road because it happened to be right over the line of  a county they wanted to clinch, or because it was at that point that they realized that had missed a turn and needed to go back the other way.

And anyone who has ever asked me to create one for them has gotten their wish, as far as I can recall. Although CHM guidelines specifically stated that we NOT locate a point at a place just because it's personally relevant (your second cousin's house), I don't have a problem with this. I added one at a county line just the other day for Oscar, in fact. It wasn't needed for any shaping purpose whatsoever, but since I knew it would be used, it deserves to be in the file.

[bejacob]

I added one at a county line just the other day for Oscar, in fact. It wasn't needed for any shaping purpose whatsoever, but since I knew it would be used, it deserves to be in the file.

Seems like a reasonable approach. I've occasionally requested an extra waypoint (usually with a reason why I thought it necessary) and always had it added. In cases where I might have driven partway between 2 existing waypoints, I usually go with the next one back (so I don't mistakenly clinch a segment which I've only partially driven).

[compdude787]

That runs counter to what I'd understood was the current SOP, to replace hidden waypoints with visible ones where possible (which I've done with some routes covered in neroute2's point requests). If we're going to have a waypoint anyway, such as for shaping purposes, making it visible would make it available to users with little extra cost, even if the potential benefit is uncertain at best.

AFAIK, that's been hinted at in the forum, but if there's ever been a clear directive to favor visible shaping points over hidden ones, I've not read it. It seems kind of pointless to clutter the files with named points at culs-de-sac and paper roads that no one will ever use.

I agree with this 100%. Whoever created the waypoint files for WA state routes did this quite extensively, and it's something that I've found to be quite annoying and I am planning to clean it up when I get time.

[yakra]

http://cmap.m-plex.com/tools/manual_includepts.php

Prefer an intersection to act as a shaping point location wherever possible. Shaping points that coincide with intersections should be added as normal, visible waypoints labeled in the usual way.

My takeaway from the manual is:
• DON'T place a point just because it leads to your second cousin's house.
• DO place a point if it's needed to keep the route's shape within tolerance.
• When doing so, prefer an intersection over a hidden point.
From here, it follows that these points may occasionally be of use by travelers, as Duke87 noted.

County lines:
It's safe to say our target audience is roadgeeks who like collecting stuff. There will be some natural overlap here with county collectors. Points at/near county lines thus may be a little more likely to be useful. If requested, I may add them in, but want to have some other justification as well; I still try to be conservative here. When Oscar recently requested two points in canabs, I added one, but not the other as it was rather close to an existing point.

WA cleanup:
I haven't looked at stuff in Washington in too great detail, but I'd say that,
If a route is "over-shaped", with more points than necessary to stay within tolerance, then trimming out unneeded shaping points may be advisable.
If points are necessary to keep within lateral tolerance, leave them in.

[neroute2]

There is a signed CA 49 Historic on Gold Strike Road from San Andreas north. Signs:
south end
north end

[oscar]

There is a signed CA 49 Historic on Gold Strike Road from San Andreas north. Signs:
south end
north end

Thanks. This has been mentioned, as an apparent old CA 49 alignment, in the middle of a very long trip report on the AARoads forum. I haven't yet found other mentions, including on cahighways.org.

I'm inclined to add it to the HB, similar to the business loops from state routes already in the HB (as part of usaca, there is no separate system for the business loops). The San Andreas CA 49 historic route appears not to be state-maintained, but then in California neither are most of the signed business routes.