PA: US 222 at PA72_N and between PA 501 and US 30 in Lancaster

[Markkos1992]

This is going to be another fun one to think about.

US 222 NB follows PA 272 NB to US 30 EB.  Simple enough.

US 222 SB must follow the C/D Road to turn left onto PA 501 SB to get to PA 272 SB.

What I think I should do would require three points on US 30 here.  I'll explain more once I put it in the WPT editor.

[dfilpus]

This situation was much discussed back in CHM development. There always have been two options for situations like this:
1. Split the difference for the divided route.
2. Route the divided route on one of the directions.
Tim chose the latter for this situation.
It preserves the concurrency of US 222 north and PA 272. It misses the concurrency of US 222 South and PA 501. Splitting the difference would lose both concurrencies.

EDIT: Fixed typo to US 222.  -Mark

[Markkos1992]

Based on what I have temporarily done in my branch, I prefer to split the difference. (https://github.com/Markkos1992/HighwayData/pull/38)  PA 272's file would be unaffected minus the breaking of the concurrency.  US 30 already has US 222 referenced in both labels at PA 501 and PA 272.

For US 222:
PA501_S +PA501 http://www.openstreetmap.org/?lat=40.060649&lon=-76.306231 (lines up with PA 501 and PA 272)
US30/501 http://www.openstreetmap.org/?lat=40.068753&lon=-76.305472 (in center of US 30 between PA 501 and PA 272, purposedly does not line up with US 30 as I did not add a point here)
PA272Lan_N +US30_W http://www.openstreetmap.org/?lat=40.068398&lon=-76.303066 (lines up with US 30 and PA 272, US 30 concurrency would stay as marked now)

I see why Tim did what he did as SR 222 in RMS would follow the PA 272 concurrency.

[yakra]

This strikes me as a solution in search of a problem...
Way I see it, oftentimes in these situations where a one-way couplet forms a triangle, splitting the difference = cutting straight from one end of the couplet to the other = effectively following one direction. For an example of what I mean, see ME ME22 I-295 StJohnSt .
As such, I think the way it's currently handled in the HB is the best way to do things, with the one exception that it may be worthwhile to break the 222/272 concurrency between PA501 and US30, as it's possible to drive US222 SB without driving PA272. This would be done by moving the coords of PA PA272 US30/222 by 0.000001. This would break the Intersecting Routes feature in the HB, but at present that's simply a consequence of "Looks Intentional" NMPs, and something worthy of its own topic.
For a good head-explody case study in intentional NMPs and breaking false-positive multiplexes, go here, search for US6/MA138@BriSt and click on it in the table, and hit the + key or use the mouse wheel to zoom in.
Or, double-click near the middle (near 41.719440,-71.153808) of the bold-looking blue line south of there, heading toward MA79/MA138@US6_E and keep on pressing + top zoom in. Surprise! It's actually three blue lines, very closely overlaid on top of one another.

(https://github.com/Markkos1992/HighwayData/pull/38)

I didn't look around at the individual commits, as I couldn't tell what file each one was referencing without clicking on it.
Suggest commit messages such as "PA117: Recentering Along US 422", "PA934: Recentering Along US 422", etc.

PA272Lan_N +US30_W http://www.openstreetmap.org/?lat=40.068398&lon=-76.303066 (lines up with US 30 and PA 272, US 30 concurrency would stay as marked now)

The label would actually be PA272_LanN (PA272Lan_N would imply that there's a "PA272Lan" route in the HB), though I don't recommend adding it.

[Markkos1992]

For the commits, I was expecting you to look at files changed.  I'll note that moving forward since the routes are not put in the commits.

Would I be able to break the concurrency just by adding a shaping point on PA 272 in US 222's file?  I thought I saw where that was doable.

I'd rather keep all the routes in the interesecting route feature.

[yakra]

For the commits, I was expecting you to look at files changed.  I'll note that moving forward since the routes are not put in the commits.

Doy! I can do that, can't I? 

Would I be able to break the concurrency just by adding a shaping point on PA 272 in US 222's file?  I thought I saw where that was doable.

Yes.

I'd rather keep all the routes in the interesecting route feature.

This would be a "best of both worlds" approach.

[Markkos1992]

I did want to mention something I currently did involving PA 72 that may merit further discussion.

In my draft for the next pull request:

"Centered PA72_N point into center of the north end of the US 222/PA 272 one-way section. (intersection where PA 72 NB turns onto US 222/PA 272 SB, Shaping Point was added where the split occurs as PA 72 already has a point at Fruitville Pike)"

[yakra]

Oh, isn't Lancaster just fun?
Took me a minute to catch on to what's happening here. The new PA72_N makes sense for the center of the Duke/Prince couplet now that Lime St has already ended. If PA72 weren't a factor here, I'd be tempted to have this point at the couplet's end, at Duke & McGovern. But PA72 *is* a factor. Considering PA72's route here, and the benefit of gaining a graph connection, your solution here at PA72_N is a good one.

PA72 itself:
I wanted a better idea of what PA72 SB does here, and took a look. Wow, okay. (I don't propose truncating PA72, as there are a few examples of one-way routes in the HB, such as the west end of NY878 (including I-878), and ME11BMil.)
All I'd do different is to convert +X222272 to a visible point, as there is a valid intersection here.

[Markkos1992]

https://github.com/TravelMapping/HighwayData/pull/2235

I made two US222/272 labels here (US222/272_B and moved US222/272_N back to is original location making the shaping point visible).  I added a shaping point on US 222 to break the PA 272 concurrency as mentioned above.