AR: AR 11 and US 70 Minor Concerns

[Markkos1992]

AR 11:  I-40(193)>-I-40.

US 70:  A point should be added between AR 357 and US 79 to keep the line south of I-40.

[yakra]

US 70:  A point should be added between AR 357 and US 79 to keep the line south of I-40.

It really wants to be here unless I move the existing shaper on I-40. Doesn't seem to be a name for it though. Hidden point added.

AR 11:  I-40(193)>-I-40.

Oh man. I wrote a couple shell scripts and opened a bunch of pull requests... like 8 weeks before mapcat signed over AR. I'll run them again and see what I come up with...

extraneous_suffix.log

Code: [Select]

1 AR/usaar/ar.ar011.wpt:I-40(193) http://www.openstreetmap.org/?lat=34.819053&lon=-91.566753
2 AR/usaar/ar.ar018.wpt:I-555(45) http://www.openstreetmap.org/?lat=35.808939&lon=-90.720162
3 AR/usaar/ar.ar077.wpt:I-555(2) http://www.openstreetmap.org/?lat=35.411333&lon=-90.282683
4 AR/usaar/ar.ar089may.wpt:I-40(135) http://www.openstreetmap.org/?lat=34.970626&lon=-92.419058
5 AR/usaar/ar.ar091.wpt:US63(49) http://www.openstreetmap.org/?lat=35.857248&lon=-90.777583
6 AR/usaar/ar.ar229.wpt:I-30(114) http://www.openstreetmap.org/?lat=34.545007&lon=-92.631011
7 AR/usaus/ar.us063.wpt:I-555(45) +I-40(265) +I-40(276) +AR463_N +US49/1 http://www.openstreetmap.org/?lat=35.808939&lon=-90.720162The AR11 point loses the exit number suffix.
For I-555(45) on AR18 & US63, rather than change that now & then change it back when I-555 is eventually extended in the field, I'll kick the can down the road.
AR77 is FP due to a multiplex.
AR89May is FP: disambiguates an unnumbered entrance-only ramp labeled plain I-40.
Similarly, AR91's points will eventually become I-555/63 and plain US63, so going no-build until then.
On AR229, SouSt_E -> I-30(116).

UNIQUE_SLASH_SUFFIX.log

Code: [Select]

1 AR/usaar/ar.ar007.wpt:US63/167_S http://www.openstreetmap.org/?lat=33.136222&lon=-92.631912
2 AR/usaar/ar.ar175.wpt:US62/412_W http://www.openstreetmap.org/?lat=36.309307&lon=-91.504301AR7: Suffix removed.
AR175: Ack, a little indigestion here. I mean, US62/63 is also technically a US62/412 split, right? ... ...US62/412_W -> US62_W.

https://github.com/TravelMapping/HighwayData/pull/5145

[Markkos1992]

I wonder if your script may be useful whenever mapcat is ready to review his states again.  It may lower the list of point concerns I find a little bit in and of itself.

[yakra]

Certainly would be, yes.
Long-term, these items could become a datacheck. That'd be way faster than these bash scripts, and we could build in better guards against FPs.
Got a bunch of similar chicken scratches littered around. I come up with ideas faster than I can implement`em...

Trying these two out in DE & PA...

DE:
No UNIQUE_SLASH_SUFFIX results.
The one EXTRANEOUS_EXIT result is a perfect example of why marking FPs should be allowed if this were a datacheck:

Code: [Select]

1 DE/usaus/de.us202.wpt:DE141(1) +US40 +US13/40 http://www.openstreetmap.org/?lat=39.680441&lon=-75.590417
PA:
UNIQUE_SLASH_SUFFIX

Code: [Select]

1 PA/usapa/pa.pa982trkder.wpt:US22/119_W http://www.openstreetmap.org/?lat=40.398578&lon=-79.422159
2 PA/usaus/pa.us019.wpt:US6/322_W +US6_W http://www.openstreetmap.org/?lat=41.622615&lon=-80.203578EXTRANEOUS_EXIT

Code: [Select]

1 PA/usapa/pa.pa028trkbro.wpt:I-80(78) http://www.openstreetmap.org/?lat=41.170544&lon=-79.097979
2 PA/usapa/pa.pa390trkhaw.wpt:I-84(30) http://www.openstreetmap.org/?lat=41.366680&lon=-75.127602
3 PA/usapa/pa.pa402trkhaw.wpt:I-84(34) http://www.openstreetmap.org/?lat=41.358016&lon=-75.055504
4 PA/usapa/pa.pa420alttrkpro.wpt:I-95(8) http://www.openstreetmap.org/?lat=39.870980&lon=-75.323312
5 PA/usapa/pa.pa423trkpoc.wpt:I-380(3) http://www.openstreetmap.org/?lat=41.106083&lon=-75.397754
6 PA/usapa/pa.pa435.wpt:I-380(13) +I-380 http://www.openstreetmap.org/?lat=41.230500&lon=-75.498012
7 PA/usapa/pa.pa715trksta.wpt:I-80(302) http://www.openstreetmap.org/?lat=40.997909&lon=-75.269694
8 PA/usapa/pa.pa940.wpt:I-80(273) +I-80 http://www.openstreetmap.org/?lat=41.055161&lon=-75.775999
9 PA/usausb/pa.us220altmil.wpt:I-80(158) http://www.openstreetmap.org/?lat=40.955854&lon=-77.769599
10 PA/usausb/pa.us220busbed.wpt:I-99(1) http://www.openstreetmap.org/?lat=40.056178&lon=-78.511104
11 PA/usaus/pa.us019.wpt:I-76(28) http://www.openstreetmap.org/?lat=40.678138&lon=-80.099666
12 PA/usaus/pa.us040.wpt:I-70(15) http://www.openstreetmap.org/?lat=40.162118&lon=-80.278097
13 PA/usaus/pa.us209.wpt:I-84(53) http://www.openstreetmap.org/?lat=41.357222&lon=-74.714059
14 PA/usaus/pa.us522.wpt:I-70(168) +I-70_W http://www.openstreetmap.org/?lat=39.757037&lon=-78.189228A lot of these could be FPs; I haven't looked at them in the HB. Only checked out the one Delaware example.

[yakra]

I changed up the EXTRANEOUS_EXIT script to exclude false positives like in the AR77 example -- it now attempts to account for multiplexes jumping from one Interstate to another, by looking forward & backward one line for another label containing parentheses.

This yields a false negative: PA US19 I-76(28) drops off the list of results because I-79(76) is adjacent, even though it's the only junction with I-76 on the route.
This could become either I-76 or I-76/79...

[Markkos1992]

Yeah, I am definitely considering the DE one as a FP.

I made the following changes in PA:

UNIQUE_SLASH_SUFFIX
pa.pa982trkder.wpt: US22/119_W>-US22/119
pa.us019.wpt: US6/322_W>-US6/98 (I liked including PA 98 in the label over going plain US6_W or debating whether I should go with US6/322 or consider US6/322 a FP.)

EXTRANEOUS_EXIT
pa.us019.wpt: I-76(28)>-I-76/79 (to match I-80/476 on PA 940)
pa.us040.wpt: I-70(15)>-I-70 (It is hard for me to believe that I-70 and US 40 only have one direct interchange with each other in PA.  The I-70 Exit 6 interchange is actually with Old National Pike.)
pa.us209.wpt: I-84(53)>-I-84 (probably used I-84(53) copied over from US 6)

The rest of these are FPs.  Most of them are due to concurrencies with interstates minus the following two exceptions:

pa.pa435.wpt:I-380(13)  (PA 435 has two interchanges with I-380.)
pa.pa940.wpt:I-80(273)  (PA 940 more or less has two interchanges with I-80.  The second involves a connector road that connects to I-476 to the north and I-80 to the south without a name.  I think it is best to leave this as-is.)

https://github.com/TravelMapping/HighwayData/pull/5151

EDIT:  I went with I-76/79 on US 19 before the last post was made.

[yakra]

UNIQUE_SLASH_SUFFIX
pa.pa982trkder.wpt: US22/119_W>-US22/119

Yup, that's what to do in these cases usually.

pa.us019.wpt: US6/322_W>-US6/98 (I liked including PA 98 in the label over going plain US6_W or debating whether I should go with US6/322 or consider US6/322 a FP.)

State route over a US route?
6 & 322 split together; this is the one place they do so. Recommend plain US6/322:
https://travelmapping.net/devel/manual/wayptlabels.php#avoidsuffix
https://travelmapping.net/devel/manual/wayptlabels.php#identicalmultiplex
https://travelmapping.net/devel/manual/wayptlabels.php#plexnosuffix

EXTRANEOUS_EXIT
pa.us019.wpt: I-76(28)>-I-76/79 (to match I-80/476 on PA 940)
pa.us040.wpt: I-70(15)>-I-70 (It is hard for me to believe that I-70 and US 40 only have one direct interchange with each other in PA.  The I-70 Exit 6 interchange is actually with Old National Pike.)
pa.us209.wpt: I-84(53)>-I-84 (probably used I-84(53) copied over from US 6)

Thumbs up

The rest of these are FPs.  Most of them are due to concurrencies with interstates minus the following two exceptions:

*Takes a look at the rest* Agreed, although...
PA390TrkHaw: PA390_S -> PA390
PA715TrkSTa: Gott im Himmel PENNDOT WUT
US220BusBed: The only non-plex here. A special case that happens sometimes. What would otherwise be I-99(3) becomes US220_N because parent route, arguably eliminating the need to disambiguate plain I-99 from anything else. Rather similar to the next examples; moving on to those...

pa.pa435.wpt:I-380(13)  (PA 435 has two interchanges with I-380.)
pa.pa940.wpt:I-80(273)  (PA 940 more or less has two interchanges with I-80.  The second involves a connector road that connects to I-476 to the north and I-80 to the south without a name.  I think it is best to leave this as-is.)

435: One interchange is with two concurrent interstates: 84/380. I-84/380 is appropriate here.
940: One interchange... (two trumpets? a double trumpet?) touch down at one point. I-80/476 is appropriate here.
So what of the other points, then? They could be plain I-## labels, with no other points at just that route to disambiguate from.
Less with Interstates & exit numbers and more with US/state routes & directional suffixes that I can remember, this was a convention Tim was pushing by the late days of CHM -- if KS666 intersects solo US37 at one point and US37/38 at another, they could be labeled US37 and US37/38, no need for directional suffixes. Not sure how widely this caught on though.
So yes, it's an option for Interstates, though as you may recall, one I find ugly enough to have avoided on MA12.
Don't know what I'd do if PA were my state, but it's possible I'd decide it's OK as-is, same as you've done.

[Markkos1992]

US 19:  I went with US6/322 over US6/98 as I should have before.  I also changed I-70(19) to I-70/79 in the meantime as it may have (or should have been) originally.  I think that I had gone with I-70(19) to try to prevent confusion with I-79(19) for some reason.

The PA 715 Truck Routes are different for both directions (SB and NB), hence the insanity.  I made the PA390TrkHaw change.

Otherwise, I think am sticking with no build for PA 435, PA 940, and US 220 BUS (Bedford) for now, but I am always willing to change my mind.  I am more leaning on not using a single label like I-99 if the route intersects it more than once because I think it may cause confusion, but that is just me.

[yakra]

US 19: ... I also changed I-70(19)

Weird. How did the script not flag this one? Worth noting that there's also a (19) in I-79(19), but that shouldn't matter... right?
Mumble mumble...
Edit: Oh OK I get it, it's working (imperfectly) as intended. Both the bits before and within the parens must be unique. If the bit within is not, that flags a bunch of 127(80) style labels. This gets complicated quickly...

to I-70/79 in the meantime as it may have (or should have been) originally.  I think that I had gone with I-70(19) to try to prevent confusion with I-79(19) for some reason.

The I-79(19) label dates back to CHM.

The PA 715 Truck Routes are different for both directions (SB and NB), hence the insanity.

LOL wow, I didn't even notice there was another one!

Otherwise, I think am sticking with no build for PA 435, PA 940, and US 220 BUS (Bedford) for now, but I am always willing to change my mind.  I am more leaning on not using a single label like I-99 if the route intersects it more than once because I think it may cause confusion, but that is just me.

Yep. I've felt that way historically, and so won't try to change your mind on this one. But I do love getting hung up pondering all the possibilities...