Travel Mapping
Highway Data Discussion => In-progress Highway Systems & Work => Completed Highway Systems Threads => Topic started by: Duke87 on February 13, 2017, 10:13:21 pm
-
This system is now ready for peer review. Have at it, folks!
Depending on how soon you read this the last batch may be missing from the HB - I only just submitted the Git pull so it's waiting for the next update as of right now.
I have also uploaded a readme file which will be here (https://github.com/TravelMapping/HighwayData/tree/master/hwy_data/ISL) once the pull is merged. This file documents my assumptions and sources.
-
TH25: fine
TH26: fine
TH30: fine
TH31: fine
TH32: fine
TH33: fine
TH34: fine
TH35: fine
TH36: fine
TH37: fine
TH38: fine
TH39: fine
TH40: fine
TH41: KflFlu -> KefFlu, AegisGat -> AegGat
TH42: fine
TH43: fine
TH44: FerRod -> FerRoth (ð is not d)
TH45: fine
TH46: really? obviously it is TH46 and signed as such, but boy is that short
TH47: fine
TH48: fine
TH49: add point for Kringlan as grade-separated junction
-
TH46: really? obviously it is TH46 and signed as such, but boy is that short
0.10mi. Why not?
Is there any rule or do we wanna introduce a minimum limit?
-
TH46: really? obviously it is TH46 and signed as such, but boy is that short
0.10mi. Why not?
Is there any rule or do we wanna introduce a minimum limit?
No it's fine about inclusion. The comment was about the tiny route having such a decent number (I guess the Isle of Man's unsigned A8 really takes the biscuit).
-
TH1: add extra shaping points near TH51_S; ThverDal -> ThvDal; ThelVeg -> TheVeg; OldS1Thin -> OldTH1_Thi? Thi?; point for TH936; OldS1_N -> OldTH1_N; OldS1_S ->OldTH1_S; OldS1Mar -> OldTH1_Mar
-
Thanks for your input!
Couple general notes regarding characters:
- I consistently and deliberately substituted d for ð, on account of this being seemingly standard practice among sources within Iceland. See for example the official website of Vegagerðin, Iceland's DOT: http://www.vegagerdin.is/
- I've transcribed æ as ae but continued to treat it as one letter for the purpose of determining when to trim names. So, Ægisgata, per the standard rule of "keep all characters in the street name if there are four or fewer, trim to three if five or more" becomes ÆgisGat ("Ægis" is four characters), which is then transcribed into AegisGat
- Likewise for transcribing Þ into th.
-
Hm. I'd think that should go the other way round -- Transcribe before truncate, not vice versa.
-
There are some datacheck errors: http://tm.teresco.org/devel/datacheck.php#preview.
Many VISIBLE_DISTANCE and SHARP_ANGLE errors which must be fixed or marked FP before activation but there are also some more "critical" errors like:
isl.th001;x062;;;DUPLICATE_LABEL;
isl.th054;+X39;+X40;Byli;BAD_ANGLE;
isl.th054;+X40;Byli;Hva;BAD_ANGLE;
isl.th061;+X082;+X083;;DUPLICATE_COORDS;(65.563147,-21.804943)
isl.th061;+X081;+X082;+X083;BAD_ANGLE;
isl.th061;+X082;+X083;+X084;BAD_ANGLE;
isl.th211;TH211;;;LABEL_SELFREF;
isl.th248;TH248;;;LABEL_SELFREF;
isl.th325;x02;;;DUPLICATE_LABEL;
isl.th643;BasSta;TH645_W;;DUPLICATE_COORDS;(65.76426,-21.683664)
isl.th643;GeiVeg;BasSta;TH645_W;BAD_ANGLE;
isl.th643;BasSta;TH645_W;+X02;BAD_ANGLE;
...
TH54 errors: There's a visible and hidden wp at the same location.
TH61 errors: 2 hidden wps at the same location.
Most of the errors are already detected by the wp editor.
-
- I consistently and deliberately substituted d for ð, on account of this being seemingly standard practice among sources within Iceland. See for example the official website of Vegagerðin, Iceland's DOT: http://www.vegagerdin.is/
Yep, you're right. I anglicised the transliteration. Swedish turned ð into d and Welsh (OK, a Celtic language rather than North Germanic) uses dd for a similar sound - eg Caerdydd is written in English as Cardiff.
- I've transcribed æ as ae but continued to treat it as one letter for the purpose of determining when to trim names. So, Ægisgata, per the standard rule of "keep all characters in the street name if there are four or fewer, trim to three if five or more" becomes ÆgisGat ("Ægis" is four characters), which is then transcribed into AegisGat
- Likewise for transcribing Þ into th.
As yakra says, transcribe before truncate.
I did have more comments, but they got eaten by my computer crashing. What I can remember, beyond the truncate-before-transcribing issue that made labels too long was 'Gat' -> 'Gata' changes (with Ægisgata, I got bogged down in the Ægis and ignored the four-letter 'Gata' shouldn't be truncated to 'Gat').
-
There's a little mistake in islth.csv file:
ERROR: Duplicate list name: ISL TH807
https://github.com/TravelMapping/DataProcessing/issues/45#issuecomment-304973676
-
What I can remember, beyond the truncate-before-transcribing issue that made labels too long was 'Gat' -> 'Gata' changes (with Ægisgata, I got bogged down in the Ægis and ignored the four-letter 'Gata' shouldn't be truncated to 'Gat').
What I can confirm from my travels around Iceland is that there are, seemingly, no official abbreviations. The entire name is always spelled out on street signs (e.g. "Ægisgata"). So, I used the first three letters as abbreviations for all Icelandic street name suffixes.
While one could certainly opine in favor of doing something different, this is a purely stylistic matter and one for which a decision on how to handle Iceland needn't impact any other country. I would therefore not think it worthwhile to go through every file and replace every instance of "Gat" with "Gata", "Myr" with "Myri", etc. unless for some reason it's really really necessary.
Also, thanks for the datacheck link. I had heard legends of the mystical datacheck page but never seen it prior to today. I've already fixed (or should have fixed - let's see!) all of the red errors and flagged some of the black ones as false positives.
How fussy do we want to be about the visible distance errors? I know there's a guideline that points should not be more than 10 miles apart, but a lot of Iceland is rather sparse and oftentimes there simply are not any intersections or points of any particular interest for longer than that. As it is I'm pretty sure I've co-opted a few people's private driveways to make waypoints out of.
-
How fussy do we want to be about the visible distance errors? I know there's a guideline that points should not be more than 10 miles apart, but a lot of Iceland is rather sparse and oftentimes there simply are not any intersections or points of any particular interest for longer than that. As it is I'm pretty sure I've co-opted a few people's private driveways to make waypoints out of.
These errors are normally not errors, but the check was added long ago in case a big chunk of route was missing in the middle.
-
How fussy do we want to be about the visible distance errors? I know there's a guideline that points should not be more than 10 miles apart, but a lot of Iceland is rather sparse and oftentimes there simply are not any intersections or points of any particular interest for longer than that. As it is I'm pretty sure I've co-opted a few people's private driveways to make waypoints out of.
These errors are normally not errors, but the check was added long ago in case a big chunk of route was missing in the middle.
In the cases where remote routes simply do not have any points worthy of a visible waypoint, these can be marked as FPs. I'd suggest waiting until just before activation to take care of those, in case some points are adjusted during review.
-
TH50 - fine
TH51 - fine
TH52 - fine
TH54 - fine
TH55 - fine
TH56 - fine
TH58 - fine
TH59 - fine
TH60 - OldTH60Abc_D -> OldTH60_AbcD ?
TH61 - off-route shaping points on other side of Fjord from GruSt
TH62 - fine
TH63 - fine
TH64 - fine
TH65 - fine
TH68 - fine
TH72 - fine
TH74 - fine
TH75 - fine
TH76 - fine
TH77 - fine
TH82 - fine
TH83 - fine
TH85 - OldTH85Abc_D -> OldTH85_AbcD ?
TH87 - fine
TH91 - fine
TH92 - EgiBra -> EgiBra_E (?) as concurrency diverge; OldTH92Abc_D -> OldTH92_AbcD ?
TH93 - fine
TH94 - fine
TH96 - fine
TH97 - fine
TH98 - fine
TH99 - check route in Hofn
-
TH99 - check route in Hofn
The route I've mapped is consistent with the path shown on Vegagerdin's official map (http://vegasja.vegagerdin.is/eng/), which I follow over OSM on account of it being a primary source. There is no signage in town along either routing.
Will fix the others in my next pull.
-
The route I've mapped is consistent with the path shown on Vegagerdin's official map (http://vegasja.vegagerdin.is/eng/), which I follow over OSM on account of it being a primary source. There is no signage in town along either routing.
That's fine. OSM was going "no" and signage was going "I have no idea" - it just needed a check.
-
TH2xx routes
All fine except:
TH275 - point at AsmVeg which is the extension of TH282.
-
9xx routes all fine.
-
TH783: Byli1 -> HolVeg; Byli2 -> ThraVeg (names from HERE map)
other 7xx routes: fine
8xx routes: fine
-
4xx routes - all fine
5xx routes - all fine except
TH590: Byli1 -> Byli (or replace with hidden point); Add point at OrrVeg (labeled as TH593 on Mapnik); Byli2 -> ManVeg (road name on HERE mapping)
6xx routes - all fine except
TH641: TH68 point coordinates don't match with TH68's corrisponding point (which needs to be relabelled from KroVeg)
TH643: OldTH643Fos_X -> OldTH643_X
TH645: TH643_S point wrongly uses coordinates for the BasSta point on TH643 file, rather than the TH645_W point
just the 3xx routes to go, which I'll do later today
-
I can't see anything wrong with the 3xx routes. Peer review complete.
-
Woo!
OK, so I've addressed all the items mentioned here and all of the datacheck errors (http://tm.teresco.org/devel/datacheck.php#preview).
What is the next step now? Is there anything else that needs to be done before islth can be moved from Preview to Active?
-
Next step is activating! :)