OR US26 point OR43 = I-5 point 299A?

[M3200]

OR: In the Portland area I recently drove south on I-5 and took the offramp for US26 East.  The connection involves ~ 2 blocks of city streets, and is well marked with US26 shields.  Since the offramp is not a direct connection between I-5 and US26, is that why the two 'points' are not shown as concurrent?

[yakra]

I don't maintain Oregon & can't speak for the person who does, but that looks to be the case. Sometimes connections are indirect, and a route just connects to what it does where it does. Trumpets are the most obvious.
Cases like this one with its big jumble of ramps are a bit more visually confusing; it's less obvious what does & does not connect.

That all said, I'd do things a little differently.
First & last point in each code block is unchanged:

I-5

Code: [Select]

297 http://www.openstreetmap.org/?lat=45.470319&lon=-122.685893
+x54 http://www.openstreetmap.org/?lat=45.469627&lon=-122.679992
298 http://www.openstreetmap.org/?lat=45.489171&lon=-122.676301
299A http://www.openstreetmap.org/?lat=45.498256&lon=-122.673329
+x59 http://www.openstreetmap.org/?lat=45.502301&lon=-122.673608
299B http://www.openstreetmap.org/?lat=45.505121&lon=-122.674584• Trims down some overdense shaping points near exit 297, while keeping the route trace from overlapping OR99W.
• Gets rid of unsuffixed 299 on 1PPI grounds. These are just the southern ramps of the 299A interchange. Label is not in use.
• Tweaks the coords of 299A to be centrally located within the diamond. A good opportunity for a graph connection with OR43.

US26

Code: [Select]

I-405(1C) http://www.openstreetmap.org/?lat=45.508098&lon=-122.684884
4thAve http://www.openstreetmap.org/?lat=45.505475&lon=-122.682028
OR10 http://www.openstreetmap.org/?lat=45.503670&lon=-122.677685
OR43 +RossIslBri http://www.openstreetmap.org/?lat=45.500467&lon=-122.675014
OR99E http://www.openstreetmap.org/?lat=45.501534&lon=-122.660722• Repositions 4thAve at the middle of the 1st couplet.
• OR10 stays as-is.
• From there, the route cuts straight across the diagonal from one end of the 2nd couplet to the other, not favoring one particular direction.
• OR43 is moved to the existing coords from the OR43 file. It's the end of the couplet, and also a good place for OR43 to tie in.
• RossIslBri becomes an AltLabel for OR43. The splits the difference for formulanone & jayhawkco's travels, a very good approximation under the circumstances.

OR43

Code: [Select]

HoodAve http://www.openstreetmap.org/?lat=45.492864&lon=-122.673028
I-5 +I-5(299A) http://www.openstreetmap.org/?lat=45.498256&lon=-122.673329
US26 http://www.openstreetmap.org/?lat=45.500467&lon=-122.675014• South of here, I'd consider trimming down some overdense points, but that's outside the scope of this post. (Yeah, I need to do that in ME & NB too...)
• +x5 is gone. It really didn't do anything.
• I-5(299A) -> I-5, as it's the only I-5 junction. Repositioned for a graph connection with I-5's repositioned point.
• Deleted +x6, which IMO kinda favors the northbound route. "OK, the route passes thru here, does it? How does it get to/from the adjacent points?"... just a confusing distraction. Let's cut right across the diagonal instead.
• The existing endpoint at US26 looks good. It's where the connections between the 2 routes all tie in, and also one end of US26's one-way couplet.

We gain 2 graph connections where before there were 0.
Still can't click directly between I-5 & US26, but I-5 <-> OR43 <-> US26 would work.

OR10

Code: [Select]

There are still a couple OR99W references, though OR99W no longer make it that far north.
OR99W -> BarBlvd
RossIslBri -> US26_E
US26/OR99W -> US26_W