My notes so far on your comments so far.
I'll be hitting the road in a few days. Before then, I'll do a pull request for the changes indicated below. While I'm out (at least two weeks), I won't be doing much work on cansk, but can respond to questions. My notes below illustrate my general approach, especially on adding waypoints.
SK 2
...
-WP54 (SK202) is in the wrong location (SK 202 needs to be fixed as well - it should routed south to Twp Rd 184)
2013 GMSV shows an SK 202 marker on the alignment shown in OSM/Mapnik, and that the alternate alignment shown in Google Maps is
only a detour route. The official highway map shows both routings as paved, without indicating which one is SK 202.
OTOH, the 2016 provincial traffic count report (which I didn't have when I drafted the SK 2 and SK 202 route files) seems to agree with you. The traffic count report item on SK 202 is unclear about how it connects to route SK 2. But the SK 2 list indicates that its junction with SK 202 is about 4.91km SE of the SK 42 junction, which would put the SK 2/202 junction right around township road 184. So I'll make that adjustment.
This seems to be just one example of conflicts among the online mapping sources. I usually go with OSM/Mapnik over Google Maps, but sometimes GM is right, and sometimes both are wrong. And official sources can be unclear as well, though here one of them is clear enough. Grrrr.
SK 2 (other)
-add WP @ 3 St (Rockglen)
-add WPs @ Dominion St, 6 St (Assiniboia)
Not needed for shaping. One waypoint in Rockglen, and three waypoints in Assiniboia should be enough for communities of their size, unless someone specifically needs another for their list files. Especially with all the new systems being added in the Eastern Hemisphere, I'm being conservative about adding points (labeled or shaping) to TM's server.
-the SP between WPs 29 and 30 is off
By about 0.01 mile. I'm tweaking it since SK 2 needs other changes anyway, but otherwise I'd let it go. I don't view shaping points (unlike labeled waypoints) as a high-precision exercise, so long as the route as shown in OSM/Mapnik stays within lateral tolerances.
-add WPs @ Hodlfast, Dana access roads
SK732_E covers Holdfast's south access.
As for Dana, seems to be a tiny ghost town near a closed military site (along with Sage Hill, closer to SK 2). Not crazy about adding otherwise-unneeded points for every little town near a highway, especially the ones not right on the highway. (Turns out that West Bend, for which I just added a point on SK 310, also fits that description, so that'll go too.)
-add WP @ Pulp Haul Rd north of Prince Albert
Not needed on SK 2 for shaping or otherwise, and it's less than two miles from existing SK 2 waypoint SK355. There is a PulpHaulRd point on SK 55, but that one is needed for shaping and/or distance, not because of the importance of the road.
SK 3
-WP 26 (SK697) should be moved 2 miles east
Google Maps says SK 697 goes south from SK 3 there. OSM/Mapnik says it only goes north, from where we have it. GMSV confirms the former, but not the latter. But the traffic count report suggests that Mapnik is right.
So I'll split the difference, create a new SK697_S point at the location you suggest, and rename the existing SK697 point to SK697_N.
-match WPs with SK 4 through Glaslyn and with SK2 through Prince Albert[/b]
Thanks. These probably won't be the only broken concurrences needing a fix -- I'll have to look for others.
-add WP @ Twp Rd 464 (Birch Hills)
-add WP @ Marcotte St (Hudson Bay)
TR 464 is only about 0.3 mile from waypoint SK25. SK 25 and WilSt seem good enough in Birch Hills.
As for Marcotte St. in Hudson Bay (on SK 9 as well as SK 3), even if a point there were needed for shaping (not really, the curve in Hudson Bay between the two SK 3/9 junctions is just barely within lateral tolerance), it's also not shown in OSM/Mapnik, only in Google Maps. So we don't have any open-source/non-commercial coordinates for that intersection, unless perhaps we go to the shapefiles which I'd rather not do here. As a matter of policy, TM avoids grabbing coordinates from commercial mapping like Google Maps, though it's OK to consult those sources on other matters such as waypoint names.
SK5
-change WP49 to SK746
I think you meant waypoint 50, SK745 (please refer to labels rather than or in addition to waypoint numbers, which can change as I edit the file). Anyway, SK745 => SK746. Also, RR51 => TinyRd (identified in the traffic count report).
-add WP @Spadina Cr, 2nd Ave (Saskatoon)
-add WP @Anaheim Access Rd
-add WP @ 1st St (Canora)
All of these are very close to existing waypoints, and I think don't help. The Annaheim Access Road is interesting for being paved, which is unusual for such a long unnumbered rural road in Saskatchewan. But it does go a long way off SK 5 to a rather small community, plus its intersection with SK 5 is only about a mile away from existing waypoint SK667.
SK6
-add WP @ SK729 (Twp Rd 202)
-add WP @ SK743 (Twp Rd 310)
Added. SK 743 was
really hard to find, since online mapping doesn't show it anywhere near SK 6, nor does it usually show township and range road numbers that could otherwise be used as reference points. But 2013 GMSV confirmed my guess about where it would be.
-delete WP 50,51
I'm guessing you meant 51 (EdgRd) and 52 (FooRd). Adding SK743 between them makes them less necessary, but I'm inclined to leave them in.
SK7
-change WP 37 to SK672/762
Done.
SK8
-add WP @ Wright Rd (Moosomin)
-add WP @ Ellise St (Rocanville)
WriRd is needed for shaping. EllSt is not.
-remove "600" from WP23 (only SK308)
Will do that and other changes. OSM/Mapnik has SK 600 concurrent with SK 8 and SK 308 from Rocanville to Welwyn, which would be a weird unsigned "implied concurrence". However, other maps have SK 600 avoiding Rocanville altogether, winding from Spy Hill to Welwyn well east of Rocanville, or including only segments from Welwyn south, so I question OSM/Mapnik's showing three SK 8/SK 600 intersections. I'll rename the three points on SK 8, and the two on SK 308, with SK600 references.
-rename WP 62 (not "SK662"?)
2009 GMSV shows an SK 662 marker at that location. The latest official road map is unclear about SK 662's routing.
I was going to label the point "SwanPla" until I found the above GMSV imagery (looking in vain for a township road sign, since wasn't sure from the 2008 MapArt atlas or otherwise what the TR number was). If it becomes clearer that SK 662 doesn't go to Swan Plains, I might change the point label back to SwanPla.