PA: US 219 Freeway Opening Meyersdale to Somerset

[Markkos1992]

I have kinda decided at this point to just leave this unsolved for about three months.  If nothing else has changed by then (approximately March 1, 2019), I will just mark it as solved and any changes can be posted separately.

Part of me just wonders if there could be a separate contract involved with signing along the old US 219 corridor.  That is why I want to hold out on this despite not seeing any signage on the plans in ECMS.

[yakra]

One thing that would make sense for part of old US 219 would be to extend PA 653 southeast of Garrett, to connect to the new US 219 alignment.

Or PA601?

[Markkos1992]

https://github.com/TravelMapping/HighwayData/pull/2392

I edited the PA 653 file to change US 219 to Mason-Dixon Hwy (MasDixHwy) for the time being.

[yakra]

Shaping points along US219 are getting pretty thick. Best practice is to just what's needed to keep the road's centerline within the red boundary lines in wptedit. If necessary, a few more can be added for major mountain switchbacks and the like, to keep mileage accurate, or to differentiate from the traces of other nearby routes to keep things less confusing in mapview. I was able to get US219 down to just 20 hidden points, still within tolerance.

Replacing visible points with hidden points should also be avoided.
Quoth the manual:

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.

If a point is needed for shaping in a given area, keeping it as a visible point allows users to specify their travels even more precisely.

To pick on the BruRd -> X35A conversion specifically (I didn't look at the others), IMO the shaping is better with the visibile point at the original coords -- the trace north of there more evenly bisects the road's E-W diversions, with a smaller diversion overall on the west, and the trace south of there is closer in to the roadway.

[Markkos1992]

Part of my concern on mileage came from me removing temporarily a shaping point along I-579 that took out about 0.1 miles.  After that, I tried to limit the mileage concern and added some shaping points in areas where the route almost goes out of lateral.  Most everything else is out of intersecting roads not really going anywhere.

I'll relook at US 219 otherwise.

Part 1 in Repository: https://github.com/Markkos1992/HighwayData/pull/107/commits/7b729e1a66c0f324b32b37b727cc1dcb3d541e5a

(Removed all shaping points that reduced mileage by less than 0.05 miles.  Replaced +X20 with Barr Ave.  This puts me at 46 hidden points. 199.12 miles.)

After the relocation and before further changes, the mileage was 198.83 miles.  I guess my question is what gauge should be used for mileage loss by removing shaping points.  For the record, most of the shaping points (certainly in McKean County) seemed to already be there.)

[Duke87]

Keep in mind that the policy about being as stingy as practical with shaping points came from the old CHM days when there were much more constraint server/processing resources available compared to now, which made it necessary at the time to keep the number of points down.

I see nothing to be gained by expending effort reducing the quantity of shaping points when we're no longer at risk of overtaxing resources if we don't.

[oscar]

But also let's not overdo it on adding new shaping points, even if that would make minor or even noticeable improvements to distance accuracy, or repositioning existing shaping points that could be removed instead.

For example, the HB estimates HI 360's distance as 26.22 miles. The actual distance is 34.93 miles. With about 600 hairpins on that really curvaceous highway, it would take at least several hundred new shaping points (only the bare minimum 9 shaping points now, plus 14 visible waypoints) for the HB to show something resembling the actual distance. Not gonna happen.

[Markkos1992]

Part 2:  Removed most other unneeded shaping points (reduced down to 20). Replaced +X4 with Clark Rd (ClaRd). Replaced +X29 with Johnson Rd (JohRd). Replaced +X30 with Camp Corbly Rd (CampCorRd). Replaced +X33 with Old Grade Rd (OldGraRd_A). Replaced +X35A back with Brubaker Rd (BruRd). Replaced +X35B with Walburn Run Rd (WalRunRd).  (Down to 20 hidden points with 196.12 miles)

+X36A stays as I did not see justification to bring back EagSt as I doubt it would almost never get used. 

See commit here:  https://github.com/Markkos1992/HighwayData/pull/107/commits/0d69ab2edf42c2ef4a8f095ebc5c2e049e722e53

Full pull request:  https://github.com/TravelMapping/HighwayData/pull/2397

[Duke87]

For example, the HB estimates HI 360's distance as 26.22 miles. The actual distance is 34.93 miles. With about 600 hairpins on that really curvaceous highway, it would take at least several hundred new shaping points (only the bare minimum 9 shaping points now, plus 14 visible waypoints) for the HB to show something resembling the actual distance. Not gonna happen.

Eh - I blame no one for not wanting to go through the painstaking effort of adding all the extra points to do something like that. The way it is can be considered good enough under the circumstances. On the other hand, if someone were to decide to go and do it because they were just that bored or because it really really bothered their OCD that it wasn't done, I certainly wouldn't complain either. I'd be impressed and say A++.

[Markkos1992]

I added in-use alternate labels from the old US 219 to the endpoints of the new US 219 freeway section.

https://github.com/TravelMapping/HighwayData/pull/2400

[rickmastfan67]

I added in-use alternate labels from the old US 219 to the endpoints of the new US 219 freeway section.

https://github.com/TravelMapping/HighwayData/pull/2400

I wouldn't do that, as that might falsely give people credit for the new US-219 expressway segment, which is something we don't want to happen.  Let them notice the error in their error log and correct it on their own.

[Markkos1992]

I added in-use alternate labels from the old US 219 to the endpoints of the new US 219 freeway section.

https://github.com/TravelMapping/HighwayData/pull/2400

I wouldn't do that, as that might falsely give people credit for the new US-219 expressway segment, which is something we don't want to happen.  Let them notice the error in their error log and correct it on their own.

I made sure to search who used the old points in the User Data to make sure that this did not occur.  I based it on the respective list file entries I saw.  I know I saw one person that would still have an error because the entry started at US219BusMey_N and ended on the old US 219.

I did the same thing on I-95 a couple months ago in PA, which is why the "40" label has a bunch of alternate labels.  The difference there was that I did not even have to look at list files because anyone that had used a number beyond "40" would need to put in a entry for I-295 or would have still lost mileage.

[rickmastfan67]

I did the same thing on I-95 a couple months ago in PA, which is why the "40" label has a bunch of alternate labels.  The difference there was that I did not even have to look at list files because anyone that had used a number beyond "40" would need to put in a entry for I-295 or would have still lost mileage.

But what if somebody had only traveled former exits 40 to 46 (or any of the other former exits sans the NJ border) on 'old' I-95?  That could have caused a possible issue in the database (Jim, mind chiming in here on this possibility) with an entry of '0' for the mileage.

Also, what if said person had only traveled Exit 40 to the NJ border on the old I-95 & never had traveled on I-276 (PA Turnpike)?  Why should they be given credit for that?  This is one reason why breaking people's list files due to a reroute is completely acceptable, & necessary when it's something this big.

[yakra]

But what if somebody had only traveled former exits 40 to 46 (or any of the other former exits sans the NJ border) on 'old' I-95?  That could have caused a possible issue in the database (Jim, mind chiming in here on this possibility) with an entry of '0' for the mileage.

I believe zero-length, point-A-to-point-A, segments aren't a problem. Jim, CMIIW...

Also, what if said person had only traveled Exit 40 to the NJ border on the old I-95 & never had traveled on I-276 (PA Turnpike)?  Why should they be given credit for that?  This is one reason why breaking people's list files due to a reroute is completely acceptable, & necessary when it's something this big.

It's possible to search the UserData repo to see if anyone had PA I-95 40 NJ/PA or PA I-95 NJ/PA 40 .listed. I assume from reading Markkos1992's post that he did this.

[Markkos1992]

I am doing more work on I-95 and I-276, but the list files affected by alternate labels for US 219 were the following:

1.  Beerman- "PA US219 PA653 PA669"  PA669 is south of US219BusMey_N so PA653 is an alternate label for US219BusMey_N.

2.  JamesMD- "PA US219 CinRd US22"  US22 is much farther north of BerPlaRd so CinRd is an alternate label for BerPlaRd.