Travel Mapping

Highway Data Discussion => Updates to Highway Data => Solved Highway data updates => Topic started by: si404 on October 03, 2018, 07:32:08 am

Title: George Washington Parkway
Post by: si404 on October 03, 2018, 07:32:08 am
I thought I'd make a separate thread for this, with a poll.

Summary: The George Washington Parkway is currently active in usasf, but in preview in usanp. When the National Park highways system is activated, then the 'select freeways' version will be removed. The usasf version does not split at the VA/DC border, while the usanp version does. Which way should we have it?

Complicating factors: northbound has twice the length inside DC compared to southbound. The borders are totally unsigned.


Previous discussion:
I noticed this today when zooming in on our VA map around DC.  There seems to be two different George Washington Parkways.  They seem to have the same waypoints between I-495 and I-66.  But, then the first one goes on for about another 5 miles while the second ends at DC/VA.  I was wondering what the difference was.

(George Washington Parkway)
.list name:  VA GeoWasPkwy
14.03 miles
Green sign

(George Washington Memorial Parkway (Langley))
.list name:  VA GWMemPkwy
9.05 miles
Brown sign

Mike
The Green sign one was made a decade ago (and probably not changed since) and is in the active 'Select Named Freeways' system.

The brown sign one was made later and is in the preview 'National Park Highways' system.

The Green one ignores the border and continues on, while the brown one is chopped.

I'm not sure they are even concurrent - the Green one ought to be chopped and synced with the brown one - and when usanp becomes active, removed.
The reason for the disparity is because there is a section of the George Washington Parkway that technically is within the District of Columbia.  However, when the select named freeways system was created, it was decided to keep everything on that side of the Potomac River in Virginia for simplicity.

Personally, I think it should still stay that way, but whomever created the National Park Highways lists clearly disagrees with that.

I drove the stretch of George Washington Memorial Parkway between I-66 and I-495 yesterday, which I believe clinches that road for me. But I just noticed that the part of the road north of VA 400 is also listed in usasf (Select Freeways), albeit with no consideration for the district border (va.geowaspkwy is continuous even though the road runs in DC for a little over a mile). And they're separate files, with no overlap. Don't want to cross-post if I don't have to, but as a newbie I don't know the protocol.
The GW Parkway has been part of usasf for many years, long before usanp was begun. Since it belongs in usanp, the usasf file will be deleted when usanp is promoted to active, but until then, it will exist in both.
^ Also, at the time it was added to usasf, it was decided by TPTB that, for simplicity, everything west of the Potomac's main channel would be considered "Virginia", which enabled it to be added as a single file instead of three.
Title: Re: George Washington Parkway
Post by: si404 on October 03, 2018, 07:41:14 am
Personally I'm on the fence - I'm all for pragmatism here, having it in VA, but precident is that we chop such routes. We might fudge the small corner (though don't with the DC Beltway (http://travelmapping.net/hb/index.php?units=miles&u=si404&r=dc.i495), having a 0.07 miles of bridge in DC, despite no signs saying DC), but this is 1.41 miles (http://travelmapping.net/hb/index.php?units=miles&u=si404&r=dc.gwmempkwy) (the file follows northbound).
Title: Re: George Washington Parkway
Post by: oscar on October 03, 2018, 09:10:21 am
How about cloning the usanp files as is, and using them to immediately replace the GWP file in usasf? I think usanp handles GWP better, with appropriate splits where the parkway crosses into and out of DC. One could quibble with the usanp version, but I'm unconvinced it needs revisions.
Title: Re: George Washington Parkway
Post by: Jim on October 03, 2018, 10:16:32 am
I think we're pretty consistent, possibly to a fault, about breaking up routes across region boundaries, so I voted to split.
Title: Re: George Washington Parkway
Post by: michih on October 03, 2018, 10:41:01 am
http://travelmapping.net/hb/index.php?units=miles&r=dc.i495

The DC route is currently about 30% in MD according to OSM borders and should be moved. But are they really accurate?
Title: Re: George Washington Parkway
Post by: froggie on October 03, 2018, 12:53:15 pm
^ No, they're not.  Again, simplicity vs. accuracy.  If you want accuracy, OSM's borders would also need to shift west.
Title: Re: George Washington Parkway
Post by: michih on October 03, 2018, 01:05:22 pm
^ Yep, that's why I've voted to omit the DC segment. btw: OSM borders to too far east? That means our DC segment is more than just 30% in MD.
Title: Re: George Washington Parkway
Post by: oscar on October 03, 2018, 01:47:06 pm
^ Yep, that's why I've voted to omit the DC segment. btw: OSM borders to too far east? That means our DC segment is more than just 30% in MD.

You seem to be referring to DC's short segment of I-95/495, not the GW Parkway. The Parkway's DC's segment endpoints in usanp (dc.gwmempkwy) are very close to the DC/VA lines as shown in OSM. Perhaps the usanp GWPkwy route file was created/updated much more recently than the I-95/495 route file, which was part of CHM's very first route set.
Title: Re: George Washington Parkway
Post by: yakra on October 03, 2018, 01:56:53 pm
http://travelmapping.net/hb/index.php?units=miles&r=dc.i495

The DC route is currently about 30% in MD according to OSM borders and should be moved.
These border point coords are probably more than a decade old. In the early days, point coords were from Google, and their cartography was a bit different from what we have now. Also, less emphasis was put on having utmost precision at the highest zoom levels. Before that, I believe Terraserver was even used to get coords for the earliest versions of Tim's "Pennsylvania/Maryland Interchange Browser" on m-plex.com, from which the HB eventually evolved.

But are they really accurate?
They match up pretty close with the line as shown in ESRI layers. Absent any other evidence from known-good, definitive sources, I'd say they are.

Again, simplicity vs. accuracy.
I say this distinction doesn't come into play here the same way it did in in IN I-69Whe scenario. It's just a matter of considering accuracy, really; the only simplicity is that of simplifying the workload; taking the No-Build route & opting not to edit the files.

If you want accuracy, OSM's borders would also need to shift west.
^ Yep, that's why I've voted to omit the DC segment. btw: OSM borders to too far east? That means our DC segment is more than just 30% in MD.
Froggie, what's your source info for this claim? The USA Topo later on Acme Mapper (http://mapper.acme.com/?ll=38.79311,-77.03815&z=15&t=U) also matches OSM & ESRI.
The SHA_LINE_ROUTES_MD_2015 shapefiles are unhelpful, as they include 95/495 to a point obviously within VA (http://www.openstreetmap.org/?lat=38.793194&lon=-77.046328). (Does MDSHA maintain the entirety of the bridges or something?)
I have not yet downloaded any DC or VA shapefiles.

I believe the DC portion of 95/495 should be retained.



I'm also with Jim & Si about the consistent precedent WRT chopping routes. Chop it.
How about cloning the usanp files as is, and using them to immediately replace the GWP file in usasf?
I'll stay neutral on this, due to the issue of breaking .list files.

I think usanp handles GWP better, with appropriate splits where the parkway crosses into and out of DC.
Agreed. The border point locations look good, considering the odd layout of the area.

One could quibble with the usanp version, but I'm unconvinced it needs revisions.
Maybe move the DC GWMemPkwy US50 (now barely outside NMP tolerance vice DC/VA) point a bit more south.
The DC side seems the appropriate place the route thru the park, with the lion's share of the ramps being here, and the Memorial Drive circle visually tying it all together.
I find myself questioning the MemDr point itself, as the locations of its ramps meeting the mainline nearly coincide with adjacent points. Think like TX I-45 @ Vision Park Blvd (http://forum.travelmapping.net/index.php?topic=2702.msg11527#msg11527), or AL US90 I10/US98_W I10/US98_E (http://forum.travelmapping.net/index.php?topic=2601.msg11533#msg11533). OTOH, one can enter at WasBlvd and leave at MemDr. Distinct & separate ramps, albeit close together. I have no big problem with keeping it either.
Title: Re: George Washington Parkway
Post by: oscar on October 03, 2018, 02:08:28 pm
How about cloning the usanp files as is, and using them to immediately replace the GWP file in usasf?
I'll stay neutral on this, due to the issue of breaking .list files.

List files will break at some point, if not now then when usanp is activated and the usasf George Washington Parkway file is removed.
Title: Re: George Washington Parkway
Post by: yakra on October 03, 2018, 02:09:07 pm
True.
So to that end, it doesn't matter when it's done.
Title: Re: George Washington Parkway
Post by: froggie on October 04, 2018, 12:49:43 pm
Quote from: yakra
Froggie, what's your source info for this claim?

- Extensive personal experience (I lived less than a mile from the bridge when I was stationed in the area and bicycled across it regularly)
- The boundaries are highlighted on the bike/ped path in brass trim.  They don't show up in most online imagery but are discernable (barely) in ultra-high-res DDOT imagery if you know where to look.  I also have photographs on VHP (http://www.vahighways.com/photos/wwb/bike-ped/virginia.htm).

Since the other day, I've done some more research into this.  While the Virginia/DC boundary is normally the mean low water mark on the Virginia side, it was adjusted in 1945 to account for pier development on the Alexandria Waterfront...so it juts out a little bit from the river edge there.  OSM's boundaries appear to be based on DC's boundary shapefiles, which are pretty close to the actual (as defined by what's on the bike/ped path.  Difference is approximately 20 feet.  You could call that a rounding error on the scale we typically use for TM WPT files, so one could argue that OSM's boundaries are acceptable for our purposes.

That said, if we are going to keep the DC I-95 and I-495 segments, they should definitely be recentered.  And recentered on the bridge median instead of being located in the outer loop local lanes where they currently exist.

To answer another of your questions, the Woodrow Wilson Bridges are jointly owned by MD SHA and VDOT.  DC has zero ownership in the new bridges.  One argument I'd make in favor of eliminating the DC segment of I-95 and I-495.
Title: Re: George Washington Parkway
Post by: yakra on October 04, 2018, 03:19:47 pm
Quote
Difference is approximately 20 feet.  You could call that a rounding error on the scale we typically use for TM WPT files, so one could argue that OSM's boundaries are acceptable for our purposes.
I'll argue that. Though, a difference of 20' isn't so terribly bad either.

Quote
That said, if we are going to keep the DC I-95 and I-495 segments, they should definitely be recentered.  And recentered on the bridge median instead of being located in the outer loop local lanes where they currently exist.
Agreed.

Quote
the Woodrow Wilson Bridges are jointly owned by MD SHA and VDOT.  DC has zero ownership in the new bridges.  One argument I'd make in favor of eliminating the DC segment of I-95 and I-495.
Disagree. If a road physically exists within a region, it exists within a region, regardless of who owns or maintains it, etc.
Title: Re: George Washington Parkway
Post by: froggie on October 05, 2018, 11:24:13 am
^ Again, simplicity vs. accuracy. 😌
Title: Re: George Washington Parkway
Post by: yakra on October 06, 2018, 12:32:49 am
I hope you're not using that aphorism to argue in favor of removing DC I-95/495... :(
Title: Re: George Washington Parkway
Post by: dave1693 on October 06, 2018, 01:03:00 am
Quote
That said, if we are going to keep the DC I-95 and I-495 segments, they should definitely be recentered.  And recentered on the bridge median instead of being located in the outer loop local lanes where they currently exist.
Agreed.

Quote
the Woodrow Wilson Bridges are jointly owned by MD SHA and VDOT.  DC has zero ownership in the new bridges.  One argument I'd make in favor of eliminating the DC segment of I-95 and I-495.
Disagree. If a road physically exists within a region, it exists within a region, regardless of who owns or maintains it, etc.

I also disagree. For supporting evidence, I note that the mile markers *on* the WW Bridge reset TWICE -- the DC mileage (all 0.8 miles of it) is marked separately from the VA and MD mileage, though it's not obvious. You have to look for it, and there's no marker at the DC/MD line.
Title: Re: George Washington Parkway
Post by: froggie on October 06, 2018, 12:41:11 pm
^ That must be new since I left the area, because there were NO DC milemarkers on the bridge at all when I lived in Huntington.

Quote from: yakra
I hope you're not using that aphorism to argue in favor of removing DC I-95/495...

I'm not...but it certainly fit your response to me.  The DC I-95/495 points have already been well-established for awhile now and are technically accurate.  The confusion removing them would cause exceeds any benefits that would be gained.
Title: Re: George Washington Parkway
Post by: dharwood on October 06, 2018, 09:40:48 pm
I believe that DC segments should be shown for both the GW Parkway and I-95/495.
Title: Re: George Washington Parkway
Post by: yakra on October 07, 2018, 12:09:33 am
Phew!
*wipes brow*
Title: Re: George Washington Parkway
Post by: dave1693 on October 10, 2018, 07:18:51 pm
^ That must be new since I left the area, because there were NO DC milemarkers on the bridge at all when I lived in Huntington.

The WW Bridge milemarkers appears a short time after the bridge redo was completed. The catch is that the DC milemarkers look EXACTLY like the MD milemarkers... what gives it away is:
1) heading from VA to MD, you get mile markers up through 0.8 (IIRC) and then the next marker is either 0.2 or 0.4;
2) heading from MD to VA, the milemarkers descend to either 0.4 or 0.2 and it's obvious that 0.0, which is not posted, is in the middle of the bridge, which is followed (IIRC) by another 0.8 milemarker, then 0.6, 0.4, possibly 0.2, and next you see a VA milemarker with 177 point something.

That is, however, all irrelevant to the GW Parkway. I haven't voted in the poll because I don't have a strong opinion either way.
Title: Re: George Washington Parkway
Post by: Duke87 on October 10, 2018, 07:47:08 pm
One thing I will assert with certainty: for sake of accuracy there definitely needs to be a DC segment, and the route definitely needs to be split.

What's more indigestion-inducing is the question of where to have the DC/VA point. There is a full half-mile where southbound is in VA while northbound is in DC, and splitting the difference isn't an option since this half-mile section is a tangle of ramps with access points to and from multiple directions.

There is no perfect solution here, but I am generally on board with the way usanp maps it, following the "median"(ish) while keeping the route in DC for the portion where it's only actually in DC northbound. In addition to the previously provided argument of most of the ramps being in DC, I will also add in support of this that by standard convention northbound is the direction of index.

The US 50 point needs to move south a bit though.

Title: Re: George Washington Parkway
Post by: froggie on October 10, 2018, 11:59:41 pm
^ That must be new since I left the area, because there were NO DC milemarkers on the bridge at all when I lived in Huntington.

The WW Bridge milemarkers appears a short time after the bridge redo was completed. The catch is that the DC milemarkers look EXACTLY like the MD milemarkers... what gives it away is:
1) heading from VA to MD, you get mile markers up through 0.8 (IIRC) and then the next marker is either 0.2 or 0.4;
2) heading from MD to VA, the milemarkers descend to either 0.4 or 0.2 and it's obvious that 0.0, which is not posted, is in the middle of the bridge, which is followed (IIRC) by another 0.8 milemarker, then 0.6, 0.4, possibly 0.2, and next you see a VA milemarker with 177 point something.

That is, however, all irrelevant to the GW Parkway. I haven't voted in the poll because I don't have a strong opinion either way.

I have to question those milemarkers considering the 0.8 marker on the Outer Loop Local Lanes is both before Milepost 178 and clearly within Virginia.  I don't think any of the mileposts you're referring to are physically within the DC portion of the bridge.  Nevermind that even the bridge bike/ped path, which would technically have more "mileage" in DC because of the angle of the DC/MD border, has less than 0.1 mile within DC.
Title: Re: George Washington Parkway
Post by: michih on January 10, 2019, 04:02:05 pm
Summary: The George Washington Parkway is currently active in usasf, but in preview in usanp. When the National Park highways system is activated, then the 'select freeways' version will be removed. The usasf version does not split at the VA/DC border, while the usanp version does. Which way should we have it?

George Washington Parkway was deleted and replaced by usanp route last Sunday. Can this thread be closed now?
Title: Re: George Washington Parkway
Post by: si404 on January 10, 2019, 04:19:05 pm
I thought it had already been categorized as 'solved'. Done now.