NH: I-93 Exit 3 / NH111APel relocation

[yakra]

https://www.facebook.com/NHDOT/posts/1010270692363206
"the new NH Route 111A... is expected to open sometime this summer/fall."
I read some plans somewhere once that the old northbound ramp and part of the mainline will be repurposed as a relocation of NH111A.

Needs action now:
A short relocation of NH111 on the west side of the I-93 interchange, as yet poorly represented in OSM.
A new northbound I-93 carriageway built in the WIDE existing median means the Exit 3 point will have to be repositioned.
For those who've just driven the old northbound alignment, an *OldI-93 point may be worthwhile, pretty close to where the existing shaping point between exits 2 & 3 is.

[froggie]

Unless there were separate points on the 111 list for northbound and southbound 93, I don't think an Old93 point is necessary.  A new point certainly isn't necessary on the I-93 list.

[yakra]

This point would be for the I-93 list. It would be useful for those who've only driven the old northbound alignment. Similar things have been done for interstate realignments in other places, though not necessarily for just one carriageway AFAIK. There's already a point at just about that exact location; it just needs to be unhidden and have its coords tweaked slightly.

[froggie]

What's the point (pun intended) of having such a point on the I-93 list?  The distances are effectively the same (less than 0.2mi total) and the 111 interchange hasn't really gone anywhere.  Just seems like a lot more picky detail than is necessary.

It'd be one thing if this was a major reroute/realignment, but the new alignment is entirely within the old median.  I just don't see the point.

[mapcat]

Looking at the satellite imagery, it appears that the current waypoint (3) has been off for a while and ought to be moved. But given the scope of the construction (at least as Google Earth shows it), it seems unnecessary to identify this as a noteworthy highway relocation. I'm fairly conservative when it comes to re-clinching segments that have been moved, but if I had travelled this segment of I-93 prior to the construction, I know I would ignore the *OldI-93 point and continue to consider the segments clinched.


I'm with froggie on this.

[yakra]

The distances are effectively the same (less than 0.2mi total)

Per WPTEdit, *OldI-93 is 1.20 mi north of Exit 2, and 1.69 mi south of Exit 3.

and the 111 interchange hasn't really gone anywhere.

On the N-S axis, the interchange hasn't really gone anywhere.
On the E-W axis, the change is more significant, as mapcat notes. (I've moved the point 0.11 mi per WPTEdit.)

From the comments, I'm not 100% sure my proposal is understood. So here's a diff of my proposed changes to look over.
https://github.com/yakra/HighwayData/commit/6d444b42be9a94e0a3f99d2bf226068b9c757ec2

It should be noted that a point is already required in the general area of *OldI-93 (formerly +x2a) due to shaping.
I've just tweaked its coords a bit and unhidden it. If we have to have a point there in the DB anyway, and it might be useful to travelers, why not have a visible one. Travelers have the option to choose to reclinch, mark as declinched, or not, to their personal taste.

[mapcat]

I guess my thought was, if there's a *oldI-93 waypoint, then that implies that a user should consider the new alignment declinched.


With a controlled-access road, however, is a special *old point really needed in any case? Any user who wants to call it declinched and go back to get it again would have to drive the entire segment between exits, so wouldn't it make just as much sense for the user to remove the entire segment from his .list?

[yakra]

They help display stats & maps more accurately for those who've used them. I may never get back to New Brunswick again, but until then, this map tells a story.

[froggie]

I guess I just don't see where the changes in the northbound alignment are significant enough to warrant putting in such a point, especially given that the new northbound alignment is wholly within the old median.

[froggie]

Regarding 111A, I have a design approval graphic that (assuming the plan hasn't changed since then) shows a realigned 111A meeting 111 just a hair (about 50ft) west of where the old northbound on-ramp loop departed from 111.  It will utilize space formerly occupied by that onramp and the old northbound mainline, but the old alignments themselves will not be directly repurposed.  Part of the old northbound alignment is slated to become a park-and-ride lot.

[yakra]

I guess I just don't see where the changes in the northbound alignment are significant enough to warrant putting in such a point,

To nitpick: I'm not so much putting in a (new) point as repurposing an existing shaping point.

especially given that the new northbound alignment is wholly within the old median.

My philosophy: We don't drive on the medians (or the HB traces). We drive on the carriageways; the roads themselves. If I had driven only the old northbound carriageway, and it closed, I would consider marking it as unclinched appropriate: I've not driven either the new NB carriageway or the SB carriageway that are approximated by the HB trace, after all.
I avoid the 1.69 mi false positive clinch of marking NH I-93 MA/NH 3, and I avoid the 1.20 false negative clinch of marking NH I-93 MA/NH 2.
See, I think of it less in terms of "the new northbound alignment is wholly within the old median" and more in terms of "the old northbound alignment is significantly outside the new footprint".

As for the NH111 list, I'm thinking:
• EntDr @ Enterprise Drive, old NH111 location
• I-93, @ new I-93 median. This is the only I-93 point.
• Point for new NH111A location
• Existing NH111A_Win -> RanRd
• Existing RanRd, rather than being renamed, is probably better off deleted.

NH111APel list

Regarding 111A, I have a design approval graphic that (assuming the plan hasn't changed since then) shows a realigned 111A meeting 111 just a hair (about 50ft) west of where the old northbound on-ramp loop departed from 111.  It will utilize space formerly occupied by that onramp and the old northbound mainline, but the old alignments themselves will not be directly repurposed.  Part of the old northbound alignment is slated to become a park-and-ride lot.

This is probably the document I remember seeing.

[froggie]

My philosophy: We don't drive on the medians (or the HB traces). We drive on the carriageways; the roads themselves. If I had driven only the old northbound carriageway, and it closed, I would consider marking it as unclinched appropriate: I've not driven either the new NB carriageway or the SB carriageway that are approximated by the HB trace, after all.

A) You're taking it to a level of detail deeper than most users are likely to go.

B) If we follow your line of reasoning to its logical extreme, we should have separate lists for opposing directions on a divided route.  I don't really see that as feasible, especially given past concerns about processing time with an increasing number of route lists and points.  Tim would have gone pale at such a suggestion.

[yakra]

A) You're taking it to a level of detail deeper than most users are likely to go.

That's what I do.

B) If we follow your line of reasoning to its logical extreme, we should have separate lists for opposing directions on a divided route.  I don't really see that as feasible, especially given past concerns about processing time with an increasing number of route lists and points.  Tim would have gone pale at such a suggestion.

I don't see that as a logical extreme following from my post, or my line of thinking behind it, at all. And of course I'm in no way anywhere near proposing that that be done. That would be preposterous.
I wrote what I wrote to help demonstrate how a traveler can end up deciding, "I once drove on something that is not now part of I-93 at all," and mark something as unclinched accordingly.
And while this is a bit different from your more extreme example (of roughly doubling the # of points in the DB) it should still be noted that my proposal does not affect # of points at all, in the slightest; the number of points on I-93 stays exactly the same.

Really, I'm a bit baffled that my proposal has received the amount of pushback that it has. I can't help but think that if we had some more voices participating in this thread, others would agree with me in finding my proposal reasonable. (But so far, there's only the three of us, so we just don't know.)
So that's probably why I'm doubling down here; forgive me if I continue to do so for a bit.

For one, I'm still not 100% sure that my proposal is 100% understood, based on some previous comments; I just want to make sure that it is. I'm not proposing a second point at or near the Exit 3 interchange.
Again, Here's the diff between my proposal and the I-93 file currently in TM/master. And Here's a link directly to the I-93 file I'm proposing. Please load it into WPTEdit and have a look if you've not done so.

Fact: If I trace directly from Exit 2 @ (42.777385°, -71.244793°) to Exit 3 @ (42.809149°, -71.272135°), I-93 is not within tolerance.
Fact: A shaping point is required on I-93 between Exits 2 & 3.
Fact: If I locate a point at (42.785865°, -71.264878°), then that puts the I-93 trace within lateral tolerance. Indeed, there is already a shaping point close by.
Fact: This point also closely approximates the divergence of the old NB alignment from the current alignment.
Fact: Whether this point is hidden or visible will have zero effect on DB complexity & near-zero effect on processing times. (There's just the small overhead of rendering one more waypoint marker overlay in hb.php or mapview.php and that's it.)
Opinion: IMO, if a point already has to be at a given location, and it can serve a purpose for travelers, then it should be made visible.
Fact: It is possible for a traveler to decide "There is a segment of what is currently I-93 that I have not traveled on." It is then possible for them to choose to mark the end of their travels at an *OldI-93 point, if available.
Opinion/Conclusion: Travelers should be given that option; this point should be unhidden as *OldI-93. It does no harm in terms of # of points, DB/algorithm complexity, processing times, etc. It does some good in allowing travelers greater precision & detail in tracking/mapping where they have left the roadway.

[mapcat]

Will you be identifying the change as Newsworthy?


To me, seeing something in Updates about a change to the alignment implies that a user ought to consider the segment declinched. Something this minor seems unnecessary, especially when I consider posting Updates.


Cleveland slightly relocated I-90 a couple of years ago (and thanks to this conversation, I am remembering to finally make this adjustment). As you can see from the current trace (look at Exit 171), it's been moved enough to justify realigning the waypoints. My instinct would be to make the adjustment without fanfare. Same with a small shift of I-75 in Dayton (Exit 54) and, in a few months, the I-71 bridge over the Little Miami gorge (between Exits 32 and 36). But now I don't know.


When I-40 was relocated several blocks south of its former route through Oklahoma City, I declinched the affected segments. IMO this example is closer to where the bar ought to be set.

[froggie]

Eric:  I'm not disputing the facts behind your decision.  What I'm disputing is that this is a significant enough realignment to warrant the inclusion of such points.