Author Topic: andcs: Andorra Carreteras Secundarias  (Read 744 times)

0 Members and 1 Guest are viewing this topic.

Offline michih

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 3231
  • Last Login:September 16, 2021, 12:11:48 pm
andcs: Andorra Carreteras Secundarias
« on: June 27, 2021, 03:33:02 pm »
System is in preview. Please report issues here: https://travelmapping.net/hb/index.php?sys=andcs

Since the system is marked for peer-reviewing, I'll do the job :)

Offline michih

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 3231
  • Last Login:September 16, 2021, 12:11:48 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #1 on: June 28, 2021, 02:25:45 pm »
CS131:
Add shaping point at http://www.openstreetmap.org/?lat=42.443562&lon=1.535076 ?


CS144:
What's the source for the existance? OSM, GM and GSV don't indicate it.


CS200:
Eng is slightly off
CarEng is slightly off


CS240:
+X421580 is off
+X831391 is off


CS260:
CS261 is off


CS310:
Ori -> UrbOri ?
Remove +X850068
Remove +X342717
Remove all shaping points between CG3 and CamiPad


CS330:
Esc -> UrbGir ?


CS350:
Sor -> CarCapPeca


CS360:
Add shaping point at http://www.openstreetmap.org/?lat=42.581664&lon=1.517323 ?


CS370:
Remove ALL shaping points!


CS510:
Remove ALL shaping points!

Offline si404

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 1536
  • Last Login:September 02, 2021, 01:19:56 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #2 on: June 28, 2021, 03:46:51 pm »
CS144:
What's the source for the existance? OSM, GM and GSV don't indicate it.
WikiSara. Looking, I've found the 2016 Designation of Roads, which has the following entry:
Carretera de Mossers i de Civís (CS144)
Té el seu origen a la CS142, al cap del coll de la Gallina, i la seva finalització a la frontera amb Espanya.
Té una amplada de 10 metres distribuïts de la manera següent: (1,5 V + 3,5 C + 3,5 C + 1,5 V).

(it begins at the CS142 at Coll de la Gallina, and ends at the Spanish border. It's 10m wide arranged as two 3.5m lanes with 1.5m verges either side)

Quote
CS131:
Add shaping point at http://www.openstreetmap.org/?lat=42.443562&lon=1.535076 ?


CS360:
Add shaping point at http://www.openstreetmap.org/?lat=42.581664&lon=1.517323 ?


CS310:
Remove +X850068
Remove +X342717
Remove all shaping points between CG3 and CamiPad

CS370:
Remove ALL shaping points!

CS510:
Remove ALL shaping points!
So on one hand you suggest shaping points that are unnecessary to keep within the tolerance on certain routes, but on the other hand you strongly insist on removing lots of shaping points as they are unnecessary to keep within the tolerance.

Now there's a reason why I have these shaping points. The roads are switchback mountain climbs and are much longer than the straight line distance. I believe WikiSara's lengths come from the traces they draw over Google Maps (saves me doing it).

Let's look at some lengths:
  • CS360 0.32km without shaping point, with suggested shaping point 0.51km, on wikisara 0.5km. As it is, without that shaping point it's ~64% of the stated route (wpteditor more precise than WikiSara, which creates discrepancies). Good shout on adding it.
  • CS310 with shaping points 7.23km, without the ones you said 5.75km, on WikiSara 8.7km. 83% of the length or 66% with your changes - probably not a good idea to implement it.
  • CS370 with shaping points 3.01km, without 2.22km, on WikiSara 3.2km. So TM's route goes from 94% of the length to 69% - seems worth keeping the shaping points.
  • CS510 with shaping points 1.19km, without 0.55km, on WikiSara 1.1km. TM's route goes from 108% to half the actual length if I remove shaping points. I know what's more accurate!

Now there's scope for possibly removing a few shaping points on the switchback routes, but all of them is not something I'm prepared to do. The numbers are clear, especially on CS510, that some sort of displaying the hairpin bends and the like is a useful thing to have.

Offline michih

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 3231
  • Last Login:September 16, 2021, 12:11:48 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #3 on: June 29, 2021, 12:54:03 pm »
So on one hand you suggest shaping points that are unnecessary to keep within the tolerance on certain routes, but on the other hand you strongly insist on removing lots of shaping points as they are unnecessary to keep within the tolerance.

The "unnecessary" shaping points are because the "alignment" sticks out at the end of the route. It's fine to me if you don't like my approach.

Now there's a reason why I have these shaping points. The roads are switchback mountain climbs and are much longer than the straight line distance. I believe WikiSara's lengths come from the traces they draw over Google Maps (saves me doing it).

We care about the trace* but do we really care about the actual length? Sorry, I didn't know that and never saw it so extreme. I remember albsh where sometimes every curve had a wp but you removed a lot of shaping points back then.

* https://travelmapping.net/devel/manual/includepts.php#sharpturns
Quote
If the route has sharp turns or switchbacks and adding a few more shaping points there would significantly improve the trace, consider adding a few more, but be conservative. Not every curve needs a shaping point. Few curves ever need more than one shaping point.

Offline si404

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 1536
  • Last Login:September 02, 2021, 01:19:56 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #4 on: June 30, 2021, 06:02:39 am »
It's fine to me if you don't like my approach.
My main issue is the conflicting approaches coming from you. One is the manual's conservative, but still not essentialist, shaping point policy to add some to improve the trace. The other is purge shaping points with a passion: "Remove ALL shaping points!"

The barking order to go far beyond (to the point of going against) what the manual suggests would have been better received (though still not that well received due to the attitude of capital letters and exclamation marks) had it not been coupled with a sensible approach to shaping on other routes adding inconsistency that made me go "huh?"

Certainly I was a bit lazy on the routes where you suggested adding shaping points (and they are added). And certainly I went a little overboard on other routes where you suggest purging all sps.
I remember albsh where sometimes every curve had a wp but you removed a lot of shaping points back then.
I removed a lot mostly as every curve had two shaping points to avoid sharp angle errors and I was pretty much doing every curve. These routes are already much more like the end product - a point at every hairpin bend and maybe one or two others (some because the route would otherwise cross itself) is perhaps a little loose, but it's far more conservative than Albania was. I certainly don't believe I've done every curve here (see CS510 example where I found 67% more curves on a short route!).

Certainly there's scope to remove some shaping points on these routes, and I will reshape them now. But "ALL!" is extreme. As I've established with the length differentials it significantly improves the trace to have at least some of the switchbacking traced.

Looking at CS510:
'every curve' (5sp, 1.31km (119% of stated length))
Code: [Select]
CG5 http://www.openstreetmap.org/?lat=42.570955&lon=1.486309
+X616645 http://www.openstreetmap.org/?lat=42.569448&lon=1.490756
+X395059 http://www.openstreetmap.org/?lat=42.568000&lon=1.491625
+X676164 http://www.openstreetmap.org/?lat=42.570187&lon=1.492006
+X384310 http://www.openstreetmap.org/?lat=42.570461&lon=1.490804
+X440300 http://www.openstreetmap.org/?lat=42.570722&lon=1.492499
MasRib http://www.openstreetmap.org/?lat=42.568842&lon=1.492209
'current file' (3sp, 1.19km (108%))
Code: [Select]
CG5 http://www.openstreetmap.org/?lat=42.570955&lon=1.486309
+X395059 http://www.openstreetmap.org/?lat=42.568032&lon=1.491523
+X384310 http://www.openstreetmap.org/?lat=42.570461&lon=1.490804
+X440300 http://www.openstreetmap.org/?lat=42.570722&lon=1.492499
MasRib http://www.openstreetmap.org/?lat=42.568842&lon=1.492209
'final product' (2sp, 1.04km (95%))
Code: [Select]
CG5 http://www.openstreetmap.org/?lat=42.570955&lon=1.486309
+X395059 http://www.openstreetmap.org/?lat=42.568000&lon=1.491625
+X440300 http://www.openstreetmap.org/?lat=42.570544&lon=1.491759
MasRib http://www.openstreetmap.org/?lat=42.568842&lon=1.492209

I've removed the specific 2 sps from CS310, and two of the four in the section you wanted them all to go: now 7sps instead of 11, and 6.36km (73%) long. I've also removed one pair of hairpins from CS370: now 3sp instead of 5 and 2.81km (88%) long. Github
« Last Edit: June 30, 2021, 06:09:06 am by si404 »

Offline michih

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 3231
  • Last Login:September 16, 2021, 12:11:48 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #5 on: June 30, 2021, 11:59:30 am »
Fine to me. I'll have a final look on the system when your changes are live.

Offline michih

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 3231
  • Last Login:September 16, 2021, 12:11:48 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #6 on: July 01, 2021, 11:55:44 am »
The NMPs can be marked FP.
The data check errors should also be FPs (but please check yourself again).

I have no further comments.

Offline si404

  • TM Collaborator
  • Hero Member
  • *****
  • Posts: 1536
  • Last Login:September 02, 2021, 01:19:56 pm
Re: andcs: Andorra Carreteras Secundarias
« Reply #7 on: July 02, 2021, 04:30:25 pm »
Activating with the next update.