Travel Mapping
Highway Data Discussion => In-progress Highway Systems & Work => Completed Highway Systems Threads => Topic started by: michih on May 01, 2020, 03:00:18 am
-
System is ready for peer-review (2512mi): http://travelmapping.net/hb/index.php?sys=dnkmr
-
Viborg
- SR519_AstW -> SR519_Ast
- SR519_AstE -> SR519/535
- SR517_X - I don't like the choices of X here. _N has the SR517 heading south, _S has the SR517 heading north, there's a W-S concurrency and an E-N one. I think each concurrency should be E-W or N-S (doesn't matter which), and either use different, or use the same pair of cardinal directions but add village names.
- SR503_W -> SR503_Vej
- PR52_W -> PR52_S (MR heads west, but PR52 heads south-south-east. Concurrency paired with _N)
- PR52_S -> PR52_Vin
- PR26/SR186 is off, and would need to be PR26/186
- SR503_N -> SR503_W
- SR180_RanW -> SR180_RanN
Svendborg
- PR8_W -> PR8_Hor
- PR8_S -> PR8_Faa or PR43/44
- PR8_N -> PR8_W (I really don't get how you got _N at all here!)
- LehVej_W -> LehVej_E (concurrency)
- LehVej_E -> JuuVej
- SR305/PR9 -> PR9_E or PR9_Sim
- add point for SteVej (also SR305) as leads to correctly labelled SteVej_N
There's a systematic issue with your cardinal directions where they just make no sense half the time.
For the multiplex splits, add a suffix: an underscore followed by a direction letter. The direction letter should match the signed direction the concurrent route is splitting toward. US80_W in the US25 file means that US 80 heads west from US 25 at that point but is concurrent to the east.
If a non-exit-numbered highway is a cross road twice, add an underscored suffix. The direction letter refers to the relative position of the intersection along the route whose file is being made. US90_S is the southern of the two US 90 junctions along US 25, which runs S-N.
If a non-exit-numbered highway is a cross road a third time, a suffixless label is an option.
If more than two points for the same non-exit-numbered cross road are needed, there are two options which can be used in combination with or ignoring the previous options for pairs of identical labels.
1. Use alphabetical suffixes _A, _B, _C, etc.
2. Choose 3-letter suffixes for nearby towns if they are fairly close. The 3-letter suffix should be the first 3 letters of the town name. or a desired 3-letter abbreviation if the name consists of more than one word. Add a suffix with an underscore and those 3 letters.
And also, there's a systematic issue with PRxx/SRyyy points
Drop the prefix of the second highway if it is more than one character long. A5/A6 becomes A5/A6. I-5/I-6 becomes I-5/6. I-25/US50 becomes I-25/50.
-
There's a systematic issue with your cardinal directions where they just make no sense half the time.
I know. I got this totally wrong when I read the manual in 2015/16.
When the "other" route (concurrent or intersecting) was a W-E route, I used _W + _E. When it was N-S, I used _N + _S route. To be honest, I never changed my way although I read the rule again a few weeks ago - but I should!
And also, there's a systematic issue with PRxx/SRyyy points
I think dnkmr is my first system with 2-letter prefix (eurtr routes in DEU-BY* too) and I was not sure how to deal with PR + SR in a label. I've checked all MR routes for this now: https://github.com/TravelMapping/HighwayData/pull/4029
*Locally fixed.
-
For the multiplex splits, add a suffix: an underscore followed by a direction letter. The direction letter should match the signed direction the concurrent route is splitting toward. US80_W in the US25 file means that US 80 heads west from US 25 at that point but is concurrent to the east.
I've never seen a "signed direction" anywhere except of in NA. Is the whole paragraph meaningless when there are no signed directions?
Just a random example:
D124_??? http://www.openstreetmap.org/?lat=45.639890&lon=5.109251
D313 http://www.openstreetmap.org/?lat=45.640102&lon=5.122056
D124/D125 http://www.openstreetmap.org/?lat=45.644014&lon=5.132868
Should it be D123_W since it is first heading west or D124_S since the main direction is south and the other main direction is north? Or - if the rule does not apply lack of direction signs - should I do it as desired?
-
It's understable the systematic issue - and thanks for cleaning them up.
When it comes to a "signed direction" when no direction is signed, I've tended to go instinctively with whether its a rough E-W or N-S alignment and then applied that to the concurrency. Circular routes are more absolute directions (eg you had an O_S and O_E and that seemed fine to me), and if there's two concurrencies and it doesn't look silly then you can go with E-W for one, N-S for the other.
-
yes, I got the hint with your message about your trip to Denmark
Doing this roughly in order of length:
Ærø, Glyngøre, Hurup Thy - all fine
Svendborg
- HegVej -> HagVej (also PR8Faa)
- SkeVej is off (also PR44)
- PR9/44 is slightly off (also PR9, PR44, SR163, SR167, SveMot)
- SR305_W -> SR305_S ?
Nykøbing
- does it not start at the ferry?
- PorVej -> PomVej
- GroVej_E -> GroVej_S (paired with GroVej_N)
- GroVej_W -> GroVej_E
- StovStr_N -> SkovStr_W
- VanVej_E -> VanVej_S (paired with VanVej_N)
- VanVej_W -> VanVej_E (concurrency split with the road heading east)
- SkovStr_W -> SkovStr_Nys or _S
- SR283_N -> SR283_S
- SR297 is off (also SR283)
- SR283_W -> SR283_N
- KarVan_E -> KarVae_E (also KarVan_W)
- VarVej is slightly off (also SR289) and check name
- SR289_E -> SR289_W (concurrency split with road diverging W, not eastern junction)
- PR9_StoS -> PR9_StoE (paired with _StoW)
- SR289_W -> SR289_Kar (for Karleby)
-
yes, I got the hint with your message about your trip to Denmark
Thanks, I really appreciate that you work on this :)
However,... I was not sure whether I could be no. 1 in Denmark. I checked the preview+active column after my last day and saw that I'm still no. 2. When I wrote the report about my trip, I wanted to write that I missed the goal by xx miles but then saw that I'm no. 1 - for active systems. I was quite surprised and just wanted to write this - and that it's only temporary till dnkmr will be active.
Again, thanks for continuing the review! :)
-
Aalborg
- SR190/180 -> SR180 as SR190 doesn't intersect, merely ends.
- StrVej -> O1_N (also SR180)
- PriGade -> O1_S (also SR180)
- OstAlle -> O2 (also SR180)
- SR507/595 -> SR595
- MarMolMot -> SR507
- SR567_E -> SR567_W (concurrency split)
- SR567_W -> SR567_E (concurrency split)
- SR567_LogE -> SR533_LogE (SR567 ends here concurrent)
- SR33_LogN -> SR533_LogW or SR533_LogN
- SR187_W -> SR187/535 and extend SR535 to this point? If not then SR187_N (paired with SR187_S)
- SR561_N -> SR561_W
- SR561_S -> SR561_E
- SR561_W -> SR561_Ged
-
Just note to self: 7 out of 11 routes are reviewed now (Svendborg twice).
-
- KarVan_E -> KarVae_E (also KarVan_W)
æ is one character, not two!
https://travelmapping.net/devel/manual/wayptlabels.php#latin
Waypoint labels may never use international characters. Instead, use the closest latin character without a diacritical mark.
https://en.wikipedia.org/wiki/%C3%86
Æ (minuscule: æ) is a character
-
Viborg - fine
Esbjerg
- SR401_KruW -> SR401_Kru
- LeijVej_N -> LeijVeg_W
- LeijVej_S -> LeijVeg_N (not a concurrency split)
- SR401_Far -> SR401_E ? (as concurrency split)
- PR8_Bur -> PR8_E (as concurrency split and only two intersections with route)
- PR8_Byl -> PR8_W (as concurrency split and only two intersections with route)
- SR443 -> SR401/443 (concurrency switch)
- SR401_Log -> SR401/429 (junction with both routes)
- add point to match OstEnde at other end of road?
- PR11/24 -> PR11_N (match concurrent either-side PR24's label)
- JerVej -> SR191 (also E20)
- OHavVej -> SR447_E (also E20 without suffix)
- SR447_E -> SR447_S (paired with SR447_N at other end of concurrency)
- PR25_W -> PR15_E
- PR15_E -> PR15/16 or PR15_Rin
- PR28_S is off
- StdVej_E -> StdVej_N
- StdVej_W -> StdVej_S
- is BaekByVej SR537? (it's signed as such on GSV). If so, add route and relabel point (also SR181, and the PR16/28 point at the other end)
- BejVej -> BjeVej (also SR181)
- add point for GjeVej where it meets SR513 (also SR513)
- ferry terminal looks to have moved in Sprogovej (also SR181)
-
- OHavVej -> SR447_E (also E20 without suffix)
- SR447_E -> SR447_S (paired with SR447_N at other end of concurrency)
SR447 is a point in use. Keep as alt label. OHavVej -> SR447_Esb
- BejVej -> BjeVej (also SR181)
Already in use for SR181. BejVej -> BjhVej
-
Odense
- SR547_N -> SR547_All
- SR547_W -> SR547_N
- PR16_W -> PR16/547
- SR547_S -> SR547_W
- PR21_N -> PR21_S
- RooVej -> RodVej
- SR445_Tor -> SR445_E (paired with _W)
- SR170/445 -> SR445_W (not a junction where SR170 diverges)
- SR170_Hod -> SR170_S (paired with _N)
- SR409_N -> SR409_W (paired with _E)
- StoVreVej_W -> StoBreVej_W
- SR453_W -> SR453_N (paired with _S)
- SR453_N -> SR453_Vin
- SR445_E -> SR445_Gam
- SR453_E -> SR453_W
- SR170/409 -> SR409/461
- add point on MR and SR442 to corrispond with StaVej_S?
- YdrVej -> YdreRingVej (also PR28)
- PR28_FreW -> PR28_FreN
- remove NorGade_S
- PR28_FedS -> PR28_FreS
- SR313_Lun -> SR307/313
- SR311/329 -> SR311/317 (not a junction with SR329)
- SR327_N -> SR327_W (paired with SR327_E)
- SR335_N -> SR335_S
- SR303/O2 is off (also SR303 and O2Ode)
- SR167_N -> SR167_E
- SR335_S -> SR335_W (paired with _E)
- SR163_W -> SR163_BroN
- SR163_S -> SR163_BroS
- SR163_E -> SR163_LanS
- SR163_N -> SR163_LanN
- SR301_S -> SR301_E (paired with _W)
- PR8_N -> PR8/323 ?
- SR323_N -> SR323_Lin
- SR301_BirN -> SR301_BirW
- SR165_N -> SR165_E (paired with _W)
- SR315_N -> SR315_W (paired with _E)
- SR315_S -> SR315_N
- SR165_E -> SR165_N (paired with _S)
- SR315_W -> SR315_S
- SR160_E -> SR160_W
-
Wow! What a mess I've done....
- remove NorGade_S
Why? It corresponds to O_N
-
Why? It corresponds to O_N
Then it should be called O_S and intersect with OFre - just like the point right next to it (https://travelmapping.net/hb/showroute.php?units=miles&u=si404&r=dnk.mrode&lat=55.562315&lon=9.753161&zoom=17) that makes this point superfluous.
-
Why? It corresponds to O_N
Then it should be called O_S and intersect with OFre - just like the point right next to it (https://travelmapping.net/hb/showroute.php?units=miles&u=si404&r=dnk.mrode&lat=55.562315&lon=9.753161&zoom=17) that makes this point superfluous.
Ok, two wps at the same location. Sure -> DEL.
-
æ is one character, not two!
As are the Greek characters: θ, χ and ψ (universally transliterated (https://en.wikipedia.org/wiki/Romanization_of_Greek#Modern_Greek) as th, ch and ps respectively). Similarly the ß (https://en.wikipedia.org/wiki/%C3%9F) ligature is rendered in latin basic as 'ss' (or 'sz') rather than just 's'.
On the other side of things, 'll' was a single letter (coming after 'l') rendered with 2 latin characters in Spanish until 2010. It remains a separate letter in Welsh (as are 'ch', 'dd', 'ff', 'ng', 'ph', 'rh' and 'th', though they don't have 'k', 'q' or 'z'). We've treated these as two characters for counting (and so a point 'Ffordd Pwllheli' would be 'FfoPwl'). There's also the Dutch 'IJ' (which unlike the Welsh/Spanish double-character letters sees both capitalised. And like æ they often appeared as a single character, though typing has removed that a lot more), though that's not now considered a distinct letter.
https://travelmapping.net/devel/manual/wayptlabels.php#latin
Waypoint labels may never use international characters. Instead, use the closest latin character without a diacritical mark.
It has no diacritical marks - the ligature turning a and e into one single character is not a diacritical mark. The closest single latin basic character to æ is probably 'e'*, but it always gets rendered 'ae' in latin basic, just as θ is rendered 'th'.
*Pretty much every æ (which re-entered English via Latin loanwords. In Latin, while the ligature was not a distinct letter as it is in Danish, it was a single character) is reformed to be 'e' in American English, whereas British English tends to render it 'ae' (though demon and ether are exceptions in En-UK): encyclopedia, pediatrician, medieval (archaeology is the En-US exception).
-
Sjæ (https://en.wikipedia.org/wiki/%C3%86)lland
- SkovVej_S -> SkovVej_Ska ? (doesn't pair with SkovVej_E, which is concurrency split, unlike this one)
- EngVej_W -> EngVej ? (doesn't pair with EngVej_S, which is concurrency split, unlike this one)
- FveVej -> FjeVej
- PR57_E -> PR57_N (paired with PR57_S)
- add KulVej_W to go with KulVej_E as it's a road to a the large settlement of Høng
- SavyVej_S -> SaeVej_S (is Sæbyvej, so even if you render æ as 'a', it's wrong)
- SR255_W -> SR255_S
- OrdVej_N -> OrdVej_W (paired with OrdVej_E)
- HojHovGade_N -> HojHovGade_W ?
- HojHovGade_S -> HojHovGade_E ?
- OddVej_N -> OddVej_W ?
- PR21_N is off overbridge
- SR155_S -> SR155_E (paired with SR155_W)
- SR231_E -> SR231_S (paired with SR231_N)
- BakRag_W -> BakRag_S (paired with BakRag_N)
- SR231_S -> SR231_Mor
- PR23_Kva -> PR23(2) ?
- PR21_S -> PR21(19)
- SR155 -> SR155_Hol ?
- HorVej_W -> HorVeg_N (paired with _S)
- SabyVej -> SabyVej_W (though really SaeVej_W)
- HorVej_N -> HorVej_Sae
- SkovVej_N -> SkovVej_Son
- PR16_Kre -> PR16/211 like PR16/205 at the other end of the PR16 concurrency.
- PR16_S -> PR16_E (paired with PR16_W)
- SR152 -> E47/55_N or E47/152 or something
- SkoVej_E -> SkoVej_Sko
- SR152_N -> SR152_S (concurrency split)
- SkoVej_W -> SkoVej_E (concurrency split)
- SR152_S -> SR152_N (concurrency split)
- O2_E -> O2_S (paired with O2_N)
- BorsGade -> BorsGade_W
- O2_S -> O2_E (paired with O2_W)
- PR21_W -> PR21_Kob
- SR156_E -> SR156_HedE
- SR156_S -> SR156_HedW
- SR156_W -> SR156_TreW
- SR156_N -> SR156_TreE
- PR23 -> PR21(13) (also change PR14 lavel to PR21)
- SR155_E -> SR155_W or SR155_Ros
- SR255_S -> SR255_W (paired with _E)
- E47/55 -> E47/55(33)
- SR154_S -> SR154_W
- JunVej_E -> JunVej_Bon
- JunVej_N -> JunVej_E
- JunVej_S -> JunVej_W
- JunVej_W -> JunVej_N
- PR59_N -> PR59_SteE or _E
- PR59_W -> PR59_SteW or leave
- PR59_S -> PR59_Lan
One thing I don't get is that you happily town suffix some street names which you have multiple of (eg HovGade), but then don't want different 'LynVej_Abc' points so do the weird take a random other letter for one of them (so Lyngerupvej becomes LyrVej as Lyngbyvej already took LynVej. It's not wrong, but I'd have preferred LynVej_Ger and LynVej_Van)
-
æ is one character, not two!
The closest single latin basic character to æ is probably 'e'*, but it always gets rendered 'ae' in latin basic, just as θ is rendered 'th'.
Sjælland is a perfect example how we could complicate it. Should it be Sja, Sje or Sjae?
I'd like to keep Sja as-is (æ -> a).
-
One thing I don't get is that you happily town suffix some street names which you have multiple of (eg HovGade), but then don't want different 'LynVej_Abc' points so do the weird take a random other letter for one of them (so Lyngerupvej becomes LyrVej as Lyngbyvej already took LynVej. It's not wrong, but I'd have preferred LynVej_Ger and LynVej_Van)
When there are two Lyngerupvej, I use LynVej_Abc and LynVej_Def.
But Lyngerupvej and Lyngbyvej are different names and a different rule applies: https://travelmapping.net/devel/manual/wayptlabels.php#differentnames
If two named cross roads have different names but would have identical labels (Milford Parkway & Millville Parkway both would be MilPkwy), either add a 4th letter to a 3-letter part of one of the labels, or choose 3 different letters for that label.
-
(doesn't pair with SkovVej_E, which is concurrency split, unlike this one)
(doesn't pair with EngVej_S, which is concurrency split, unlike this one)
(paired with PR57_S)
(paired with OrdVej_E)
(paired with SR155_W)
(paired with SR231_N)
(paired with BakRag_N)
...
Is there a rule for how to handle "pairs"?
-
About the direction suffix thing.... I usually applied the suggested changes but one or two of the last post were points in use I didn't change as suggested.
It's understable the systematic issue - and thanks for cleaning them up.
When it comes to a "signed direction" when no direction is signed, I've tended to go instinctively with whether its a rough E-W or N-S alignment and then applied that to the concurrency. Circular routes are more absolute directions (eg you had an O_S and O_E and that seemed fine to me), and if there's two concurrencies and it doesn't look silly then you can go with E-W for one, N-S for the other.
https://travelmapping.net/devel/manual/wayptlabels.php#split
For the multiplex splits, add a suffix: an underscore followed by a direction letter. The direction letter should match the signed direction the concurrent route is splitting toward. US80_W in the US25 file means that US 80 heads west from US 25 at that point but is concurrent to the east.
The grey segment is just a "should" and only refers to countries like US where directions are signed.
That means, we don't have a rule for most of the review comments. Should we add the "instinctive way"? Any suggestion?
-
æ is one character, not two!
The closest single latin basic character to æ is probably 'e'*, but it always gets rendered 'ae' in latin basic, just as θ is rendered 'th'.
Sjælland is a perfect example how we could complicate it. Should it be Sja, Sje or Sjae?
I'd like to keep Sja as-is (æ -> a).
Sja is obviously right whether it's 'a' or 'ae', because a third letter theta gets left as t rather than making a four-letter abbreviation. You'll note that I've not corrected labels where æ is second letter - it's a bit of a can of worms. Should note that the older systems in Denmark takes Sæletters as Sae, and Sjæletters as Sja. I can go through and do a check if you want...
For the multiplex splits, add a suffix: an underscore followed by a direction letter. The direction letter should match the signed direction the concurrent route is splitting toward. US80_W in the US25 file means that US 80 heads west from US 25 at that point but is concurrent to the east.
The grey segment is just a "should" and only refers to countries like US where directions are signed.
That means, we don't have a rule for most of the review comments. Should we add the "instinctive way"? Any suggestion?
Pretend that the road is signed with directions, and go with that. So if a roughly N-S route splits off a concurrency to the south, _S. Obviously complicated a little when there's two concurrencies with the same road...
Is there a rule for how to handle "pairs"?
No, but like the above 'instinctive way', you don't see roads signed North one way and East the other, but N-S and E-W. And so a route running concurrent should leave the concurrency in the opposite direction at either end of the concurrency.
About the direction suffix thing.... I usually applied the suggested changes but one or two of the last post were points in use I didn't change as suggested.
That's fine.
-
Should note that the older systems in Denmark takes Sæletters as Sae, and Sjæletters as Sja. I can go through and do a check if you want...
I think it is not necessary.
Pretend that the road is signed with directions, and go with that. So if a roughly N-S route splits off a concurrency to the south, _S. Obviously complicated a little when there's two concurrencies with the same road...
That's how I drafted all systems from the beginning*. I always estimate the whole route but my reviewers (mostly you) did not always agree with that.
*When there is only one concurrency and no additional intersection - when there are only two "A123" labels.
Is there a rule for how to handle "pairs"?
No, but like the above 'instinctive way', you don't see roads signed North one way and East the other, but N-S and E-W. And so a route running concurrent should leave the concurrency in the opposite direction at either end of the concurrency.
I still prefer a rule in the manual.
Just an example where only _W and _E make sense but it is not a "concurrency pair":
D540_W http://www.openstreetmap.org/?lat=47.247912&lon=3.330708
D540/D977 http://www.openstreetmap.org/?lat=47.254343&lon=3.362101
D540_E http://www.openstreetmap.org/?lat=47.256700&lon=3.369560
-
Should note that the older systems in Denmark takes Sæletters as Sae, and Sjæletters as Sja. I can go through and do a check if you want...
Done: https://github.com/TravelMapping/HighwayData/pull/4212
-
Similarly the ß (https://en.wikipedia.org/wiki/%C3%9F) ligature is rendered in latin basic as 'ss' (or 'sz') rather than just 's'.
There are also ä, ö and ü. I've dealt with umlauts as one character only. ä -> a, not ae.
-
Similarly the ß (https://en.wikipedia.org/wiki/%C3%9F) ligature is rendered in latin basic as 'ss' (or 'sz') rather than just 's'.
There are also ä, ö and ü. I've dealt with umlauts as one character only. ä -> a, not ae.
Very true.
I'm leet now with post 1337
-
Just seen we have 99 systems in Preview :o Time to try and keep that number down by reviewing the last route of this system.
Kolding
- SR401 -> SR401_Dyb
- SR481_E -> SR481_Gra
- SR401_KruE -> SR401_Kru
- SR401_KruW -> ForVej (presumably the route was changed, but this label wasn't)
- HarVej -> HarVej_N
- SR481_S -> SR481_E
- SR170/481 -> SR170_Sog
- SR170_Sko is off
- SkaVej_S -> SteVej (both sides of point, MR runs on SkaVej)
- O_KolS -> O_KolW (O diverges west)
- O_KolW -> O_KolS (O diverges south) and is off
- SR191 labels for O concurrency need changing as they refer to road names.
- SR184_S -> SR184_E (or something, so the other two are a N-S pair)
- SR184_W -> SR184_S
- RydeMolVej_N -> RydeMolVej_E (or _S to _W - make pair match)
- Lyn -> EjeVej_S (MR is on Lyngen s of point)
- SR581 is off (also PR26), and perhaps wants a suffix on MR?
- Dar_E -> Dam_E
- is KliVej SR557 (https://www.google.co.uk/maps/@57.042679,8.514962,3a,27.1y,9.56h,86.47t/data=!3m6!1e1!3m4!1sJJieOkSMiPwP78mi18ws6g!2e0!7i16384!8i8192?hl=en)? Need to add that route to browser if so.
- PPR11/29_Ols -> PR11/29 (doesn't need disambiguating)
- perhaps add a clarifying point (hidden or visible) between HotVej_W and SR596_Bje
- JylGade_W -> NorSka or something (as route is on, rather than intersects JylGade here)
- PR55_HalE -> PR55_HirE (not sure where Hal came from, but is obviously meant to be paired with _HirW at the other end of the Hirtshals Sondre Ringvej)
Done!
-
Thanks! :)
https://github.com/TravelMapping/HighwayData/pull/4256
Activation with next site update.