chlrn: Chile Rutas Nacionales

[neroute2]

Sources:
2018 route log: http://www.vialidad.cl/areasdevialidad/gestionvial/Documents/redvialnacional2018.pdf
Official maps/GIS: http://www.mapas.mop.cl/
Decrees defining routes: http://www.vialidad.cl/normativas/Paginas/caminosnacionales.aspx

[neroute2]

chlrn and chlsf are now in preview: https://travelmapping.net/user/region.php?rg=CHL
Notes are in the readme: https://github.com/TravelMapping/HighwayData/tree/master/hwy_data/CHL/chlrn

[michih]

When chlrn is ready for review, I would preview this instead of anything in Mexico. Let me know.

[neroute2]

One major issue to be taken care of before review. The routes are known as either Ruta X or R-X. But there are also (signed) secondary routes known as R-X (as well as other letters). There is at least one conflict: R-180 intersects both primary R-156 and secondary R-156. What should be done? Ignore the issue and disambiguate? Use Ruta156 instead of R-156?

[michih]

I had the same issue with primary Fylkesveier and secondary Fylkesveier in Norway and we never distinguished them. There were tons of Fv routes and identical numberings, not just one. It was recently changed, Fv routes are now unique all over Norway.

Just go with R156 and go with NEWS suffixes or 3-letter city suffixes. And omit hyphens.

[yakra]

Is this anything like the primary / secondary situation in Tennessee?

[neroute2]

Is this anything like the primary / secondary situation in Tennessee?

No. The primary routes follow a rough grid and are all referred to as R-X. The secondary routes are grouped by initial letter in horizontal bands from north to south. A-93, for example, goes to the northern tip of Chile, and Y-905 is on Isla Navarino near the southernmost point. The secondary R band includes Victoria (where primary R-5 meets primary R-181CH) and Angol (where primary R-180 meets primary R-182).

It would be like Michigan having county highways that begin with M-. In that case, we would use MI* and CRM*, but that's the U.S. standards.

[neroute2]

Another issue: there are a bunch of U-turns on R-5 that I label by the nearest km post. R-5 has a zero point in Santiago, with *N and *S posts north and south of the city. So that leads to points like R5_U328N and R5_U1059S, which give a LONG_UNDERSCORE error that "is always a true error and cannot be marked false positive".

[michih]

Interesting. The rule only says:

Use a _U suffix for interchanges that are nothing more than a U-turn ramp. If more than one is needed for the same highway, use _U1, _U2, etc.

No limit is indicated and we said that no clarification is needed:

Only problem is the manual sequential suffixes RN9_U1, _U2, etc., which would require a wholesale renumbering any time one is added.

It requires no such thing.

We will have the same issue in ARG and likely other SA regions.

[neroute2]

This should be ready for review now.

[michih]

@yakra, what do you think about this? https://travelmapping.net/devel/datacheck.php?rg=CHL&show=LONG_UNDERSCORE,LABEL_SELFREF

Interesting. The rule only says:

Use a _U suffix for interchanges that are nothing more than a U-turn ramp. If more than one is needed for the same highway, use _U1, _U2, etc.

No limit is indicated and we said that no clarification is needed:

Only problem is the manual sequential suffixes RN9_U1, _U2, etc., which would require a wholesale renumbering any time one is added.

It requires no such thing.

We will have the same issue in ARG and likely other SA regions.

@neroute2, just out of curiosity, what's the story behind the "CH" routes, e.g. R11CH? What does CH mean? It's not an abbreviation nor banner (which must be 3-letter abbreviations) but just part of the route, so it's technically according to the rules

[neroute2]

@neroute2, just out of curiosity, what's the story behind the "CH" routes, e.g. R11CH? What does CH mean? It's not an abbreviation nor banner (which must be 3-letter abbreviations) but just part of the route, so it's technically according to the rules

The routes with "carácter de internacional" (international character) (other than the single-digit longitudinal routes) get a CH at the end. My guess is that Chile was hoping other countries would renumber their routes to match, and the CH would indicate that it was the Chilean portion of the route.

[yakra]

Yikes, sorry I'm late to the party, guys. -_-
I rarely ever check the unread replies page here on this forum, unlike on AARoads.
Everything from Reply #7 onward is flagged as . On that note:

Another issue: there are a bunch of U-turns on R-5 that I label by the nearest km post. R-5 has a zero point in Santiago, with *N and *S posts north and south of the city. So that leads to points like R5_U328N and R5_U1059S,

That explains those unusual looking labels. If I had seen this post I would have allowed more time for discussion and not been so quick to slap DataProcessing#473 on top of #472.
Now I understand why these points were labeled as they were, though I don't agree with the intent behind it.

which give a LONG_UNDERSCORE error that "is always a true error and cannot be marked false positive".

Sidestepping the issue of whether _U559 or _U1059S are "good" for the moment:
Even just going by what's in the manual now, it's possible -- however unlikely -- that a route could have 100 U-turn interchanges. If a _U100 suffix flags an error that can't be marked as FP, that's plusungood.
So yes, this datacheck should be changed. And even if I don't like those letter sub-suffixes, I can avoid making the datacheck overambitious & flagging them.

So that's LONG_UNDERSCORE.

@yakra, what do you think about this? https://travelmapping.net/devel/datacheck.php?rg=CHL&show=LONG_UNDERSCORE,LABEL_SELFREF

As for LABEL_SELFREF?

DataProcessing#472 avoided flagging errors at _U# labels. So far so good.
(Mostly. There was still a bug with INVALID_FINAL_CHAR cases with nothing after an underscore.)

Then came #473.
• What I thought I was doing: fixing something still broken that #472 failed to fix.
• What I arguably actually did: re-broke something that #472 fixed.

Reverting #473 while avoiding the bug in #472 is easy enough to do. Since for the moment it's already in there for better or worse, I'd still like to make the case that it stay.

First, I should clarify my comment:

Only problem is the manual sequential suffixes RN9_U1, _U2, etc., which would require a wholesale renumbering any time one is added.

It requires no such thing.

Then it needs to be clarified that it doesn't.

The potential need for renumbering is a maintenance issue, and the manual already says that Care must be taken to ensure that changes do not "break" a user's list file.
From a maintenance standpoint, _U1, _U2, etc. suffixes are similar to _A _B _C etc. suffixes. Or fudged letter suffixes for same-numbered exits. Sure, when first added to a .wpt, they'll be neat & orderly. But sometimes necessary highway data changes + the need to avoid breaking .lists will result in out-of-order labels or breaks in the sequence.
That's what I meant by that.

We will have the same issue in ARG and likely other SA regions.

Will we though? IMO the issue of how to number U-turn ramps is secondary -- these cases aren't U-turn ramps at all! Even if ARG wants to call them "retorno", that's a misnomer. They're just regular old interchanges.

Labeling them with _U### misapplies the rule, taking it outside the scope of interchanges that are nothing more than a U-turn ramp in an attempt to fit a square peg into a round hole.
If it does come to the point of adding km numbers tohave unique labels, why add the extra R5_U clutter?
For the issue of How to label unsigned junctions when there's no clear name available, I think a solution working within the existing rules is preferable.

My thoughts from the Rule Change thread:

How would we ever get to option 9? Are there situations where 8 can't be applied? IMO 9 is more limiting than 8, which could be used for roads without km-posts.

Seconded. I've only quickly skimmed the threads for the South American systems, but I don't see any reason we can't use truncated place names, or a need to introduce a new rule.

Have we exhausted all other possibilites for names, and there's not even a truncated distant placename to go with?

This, primarily. Do ARG or CHL have, say, county equivalents that could be used in a placename-type label?

If it fits the RN9_U111 criteria, which are already specified in the manual, let's go with that.

Emphasis added. I don't think it does, as noted above.

Then maybe a km number, (alone, no need for "km") if there are truly no other possibilities.

I don't see any reason "Mi", "Km", etc should be added. It's just extra cruft.

Thinking about this some more, I'm willing to walk back on it a bit -- "km" might be useful to indicate the label is not a standard exit number, but a placeholder for inadequate highway data or other rare situations.

[neroute2]

These interchanges are literally signed RETORNO. Spanish (and Portuguese) for U-turn. Yes, there's often farm access (which may include through roads), but if there's no sign giving the farm a name, we're not going to trawl property records for the owner's name.