gbrrr: United Kingdom Ring Roads

[michih]

The preview system has 57 routes for about 357 miles. 47 TM users have travels so far.
Please report issues here!

[yakra]

St Helier Ring Road clockwise & anticlockwise should be split into 2 separate ConnectedRoutes.
Consider using CW & ACW abbrevs to be consistent with irlorb.

[Duke87]

St Helier Ring Road clockwise & anticlockwise should be split into 2 separate ConnectedRoutes.
Consider using CW & ACW abbrevs to be consistent with irlorb.

Wait, why? Why is this handled differently from how any other route with one-way pairs is handled?

[si404]

Wait, why? Why is this handled differently from how any other route with one-way pairs is handled?

Concurrencies - several places where its concurrent with different roads in different directions.

[neroute2]

I suppose it's better than this BS...

It looks like Jersey is explicit about not having one-way pairs on numbered routes, with the possible exception of A14 at A7 and A8. B86 even changes direction in the middle!

Edit: A7 is explicitly signed (no parens) on the piece of SHRRAnt that OSM shows as X3. But then the B74 portion has parens. Here's A9 sans parens on B87.

For consistency, if we're keeping it mostly as-is, A7/A14 should be redone with B82 ending at A7 - or does it use Janvrin Road to A14?

Finally, should gbrrr and jeyc be different colors?

[yakra]

Wait, why? Why is this handled differently from how any other route with one-way pairs is handled?

I darn near clawed my eyes out when I saw that done on irlorb. "WTF", etc.
Taking a closer look at the wildly differing routings and even termini (and junction numbers?) I threw up my hands and said whatever.

That said, what's arguably appropriate for irlorb probably shouldn't necessarily be extended everywhere.
Much more straightforward standard couplets here.

Concurrencies - several places where its concurrent with different roads in different directions.

Not buying this argument. This is just gonna happen with couplets. Just gotta plot the route where it goes and roll with it. MB Rte42 may have some counter-intuitive point labels, but MB Rte37 just cuts across the couplet even if EB & WB widely diverge; it doesn't have separate EB & WB files.

I suppose it's better than this BS...

Did you open a topic on the Updates to Highway Data board about this, rather than just commenting in an unrelated GBR thread where froggie's much less likely to see it?

Edit: A7 is explicitly signed (no parens) on the piece of SHRRAnt that OSM shows as X3. But then the B74 portion has parens. Here's A9 sans parens on B87.

What do parens mean?

[michih]

What do parens mean?

This. "(A1)" means "to A1".

I suppose it's better than this BS...

OMG! Please report it as yakra wrote.

Wait, why? Why is this handled differently from how any other route with one-way pairs is handled?

Concurrencies - several places where its concurrent with different roads in different directions.

I fully agree!

[Duke87]

Concurrencies - several places where its concurrent with different roads in different directions.

Not buying this argument. This is just gonna happen with couplets. Just gotta plot the route where it goes and roll with it.

Even better example.

But yeah mapping the separate directions separately makes no sense since this implies you need to drive it both ways for a full clinch and we don't do that for any other route. If you have half-concurrencies, meh, so be it.

[michih]

Concurrencies - several places where its concurrent with different roads in different directions.

Not buying this argument. This is just gonna happen with couplets. Just gotta plot the route where it goes and roll with it.

Even better example.

But yeah mapping the separate directions separately makes no sense since this implies you need to drive it both ways for a full clinch and we don't do that for any other route.

Seconded. Both examples.

If you have half-concurrencies, meh, so be it.

Disagree.

[neroute2]

I suppose it's better than this BS...

Did you open a topic on the Updates to Highway Data board about this, rather than just commenting in an unrelated GBR thread where froggie's much less likely to see it?

The connections between US 29, US 80, and AL 81 is a little uneven at the town square in Tuskegee. Usually, those just get straightened out, right?

There's points AL 81_N and AL 81_S on US80, but AL 81 just pierces the town square without the additional points. (US 29 only has the AL81_N point because of its concurrency with US 80.)

It is explained upthread why I did things the way I did and why I do not intend to change it.

As for split one-way pairs, I do still need to fix MEX 40 and 85 thru Monterrey; 40 is split by the river, with a different local number for each side, and 85 has wildly different routes for each direction. Edit: finally fixed in an inelegant way

[yakra]

If you have half-concurrencies, meh, so be it.

Disagree.

I'd say I agree with Duke87, but what exactly am I agreeing with? What do "you have" and "meh, so be it" mean?
If they exist in the field, yeah -- so be it. Chin up & accept it, and take reasonable steps to accommodate it if needed. The assumption here is, plotting a 2nd route is not "reasonable steps".
IMO it's better to keep half-concurrencies out of the data -- more on this below.

Even better example.

Ooh, yes. Very good example.
60, 87 & 287 are all centered between each half of their couplets. As they should be.
Dead center, 87 & 287's route traces overlap, but the routes themselves don't. One's on the inner roads, the other on the outer.
Did you drive 87 or 287? You may or may not have driven on 60, depending on which direction.
Did you drive on 60? You drove on something else. Depends which direction.

On a smaller scale, there's where US6 & MA138 cross the bridge into Fall River & link up with MA79.
It was even more confusing back before the MA79 freeway was demolished -- US6 on the frontage rd, MA79 on the mainlines, MA138 on the frontage rd NB & the mainline SB.
Another "maybe you drove something else / maybe not / you drove something else but what was it" scenario.

In both cases I opted to break any concurrencies with hidden points, and let the user decide what they have & have not traveled & enter it manually.

[Duke87]

In both cases I opted to break any concurrencies with hidden points, and let the user decide what they have & have not traveled & enter it manually.

I do think this is the best way to handle half-concurrencies like this.

[si404]

If they exist in the field, yeah -- so be it. Chin up & accept it, and take reasonable steps to accommodate it if needed. The assumption here is, plotting a 2nd route is not "reasonable steps".
IMO it's better to keep half-concurrencies out of the data -- more on this below.

I view breaking a load of concurrencies and having the route mapped differently to the underlying routes to be far more unreasonable than two routes (fully understanding how unreasonable that is - it was not a decision I took lightly).

Even better example.

Ooh, yes. Very good example.

Yes, it is a very good example of a real world thing messing with our ideal mapping that took an age to wrangle out a solution, that solution is best of a bad situation, but still sub-optimal and not something that would necessarily translate to similar situations.

I don't see how this situation isn't likewise in need of a similarly bespoke sub-optimal solution that wouldn't necessarily translate to other situations.