Austin David

Curviest road we have is I-526

@JohnCh and I have been chatting a little about a "find along route" style feature. I've got a proposal I'd like to share, in case anyone has any thoughts. I've added followup below for why "find along route" is not easy. (note the following is AI-generated, but with a lot of guidance) Here's the experience from the rider's side — what they see and do, start to finish. The setup You're in the route editor, building a ride. The map fills most of the screen; your waypoints sit on it as numbered, draggable pins. In one corner of the map there's a single small icon — a pin+ button. That's the entire footprint when you're not using it. No panel, no sidebar row, nothing competing for space. Finding stuff You pan and zoom the map to frame a stretch you care about — say a section of road where you'll want a coffee stop. Nothing happens as you scroll — no searches fire, no spinner, no cost. The map is just a map until you ask it for something. You tap the pin+ icon. It expands in place into a short row of category chips: Fuel · Food · Coffee · View. Still nothing's been searched. You tap Coffee. Now one search fires, scoped to exactly what's framed on screen. A handful of results (~6) drop onto the map as a visually distinct pin layer — a different color/shape from your route waypoints and from the curated nearby-places pins, so it reads instantly as "candidates, not yet in your route." Alongside the pins, the control shows a compact list: each result's name, star rating, and rough distance from the center of your view. Importing one You tap a result — either its row in the list or its pin on the map. A small info card opens with the name, rating, and one primary button: "+ Add as waypoint." You tap it. The place becomes a new numbered waypoint in your route, instantly — the route re-routes through it, the new draggable pin appears, and that candidate pin disappears from the provisional layer (so you can't accidentally add it twice). The waypoint inherits the place's name as its label. Adding more, or backing out Want a fuel stop too? Tap another candidate → "+ Add." One tap each. Want different results? Tap another chip, or re-frame and search again — the old candidates clear and the new set drops in. Done exploring? Collapse the control (tap the icon). All the candidate pins vanish and the map returns to its clean baseline — just your route and that one small icon again. The feel of it The whole thing is invisible until you reach for it, and gone the moment you're done. It never searches on its own, never writes anything to your saved route until you explicitly add a waypoint, and never clutters the screen with results you didn't ask for. The motion is: frame → tap category → tap result → it's a stop. Three taps to the first one, one tap for each after. A couple of honest edge cases If the map hasn't settled (no bounds yet), tapping a chip just says "pan the map first" instead of firing a useless search. If your route's already at the waypoint cap, the "+ Add" button shows disabled with a "route full" note rather than silently failing. "View" maps to Google's tourist_attraction category, which is a bit coarse — it can surface a museum where you wanted a scenic pullout. That's a known imprecision, called out as an open decision (accept it, or curate the category list). ----- Why "find along route" is hard: - the various Google APIs are not free, but there is a $200/mo credit. "search for a place" is the single most expensive request - "Search for a place" can return only 60 things at at time. For a long route "all the places within a mile of the planned path" is a lot more than 60 - there is no API for "be smart about what you find". You can search for "gas stations" but only in a radius; finding "all the gas stations along a route" is a lot of searches - Google Maps has some magic and their own free-to-use stuff built in, and it works pretty well: "navigate A->B" + "find gas along the route" is a lot smarter than what is publicly available to a service like route 7. - I *have* implemented 'stuff close to a waypoint' -- it's a lot easier to use. Give it a shot, tell me what you think. I use it by dropping a pin near where I think there is stuff (ie a town), then search for fuel or whatever, review the results. It's not "anywhere along route", but it is straightforward to use. As editor you can trim results (hiding mcdonald's or sus gas stops, for instance). The remaining hits show up alongside your chosen waypoint. There's discussion about a "make this a waypoint" feature, not yet added.

whoa -- details?

as in, "outer banks is cool but mostly straight and salty" ?

We're all doing a terrible job representing the southeast. The following is a proposal, all of these are editable -- click to sign in (free for life), make copies to change stuff. Overview: Blowing Rock -> Cherokee -> Asheville https://route7.austindavid.com/route/QlYEaMKAeTmv 1 day to gather, 2 days in the mountains, 1 day home. Day 1: go to Blowing Rock. Resorts are cheap during the week... or at least, they were in the "before times" when I used to travel a lot. Wilmington route for SENC: https://route7.austindavid.com/route/fro05kDTO8i6 Charleston route for me: https://route7.austindavid.com/route/DaxxB9cfN1zG Greenville route for uglyfast et al: https://route7.austindavid.com/route/DgsB6zLJjAh8 Day 2: ... is a lot of BRP still closed? https://www.nps.gov/blri/planyourvisit/roadclosures.htm -- seems to be some significant closure between blowing rock -> asheville. Blowing Rock -> Cherokee: https://route7.austindavid.com/route/9vPhadSdsw_U Day 3: Cherokee -> Asheville: https://route7.austindavid.com/route/MmF50JNmMRLf There's room for Deal's Gap in here. I've never driven it, I hear mixed things. Day 4: Asheville -> home Wilmington for SENC: https://route7.austindavid.com/route/-O6uPcRau8GP Charleston for me: https://route7.austindavid.com/route/rJ-0jxW25_N0 Again, just a proposal. I'm tired of working and tired of looking at my garage.

Try just making a route with route7 and (I assume on your phone) click "Navigate full route". That sends the first 10 waypoints directly to Maps, just a big 10-point URL. Maps will automatically route turn-by-turn to the first waypoint, and when you get there it starts to the next. I did it yesterday with Android Auto, it was surprisingly easy. Exporting GPX & importing to Beeline is like 13 clicks total. Also very easy, but not quite as easy as the one-click Maps kickoff.

Can users see other people's public routes? Not implemented. I haven't spent much time on this feature, other than the distinction between "public" and "unlisted". Right now there's no exposure or search for the "public" ones. The primary use-case is "past a link in the forum", which works for both public and unlisted routes. How would you expect "discovery" to work? search, browse, pins-on-map ? (any of those should be ~easy) Can users import routes from other sources? Not yet... how would it work? GPX-style import should be super easy, at least for coarse GPX without a lot of waypoints. If you have an idea or application in mind, happy to discuss. re: vibecoded: it's currently private, I'm tryin to decide what to do about it. Are you thinking to make a Beeline-like WatchOS app with turn-by-turn? Gemini says: Because it's a Google platform, Google Maps is deeply integrated into Wear OS right out of the box and gives you full turn-by-turn navigation, including a live, interactive map view on your wrist, standard text directions, and haptic turn alerts. It also works fully standalone if you have a LTE/cellular model. I assume you want to simplify the experience somehow, with more like the Beeline arrow and a distance to next turn?

https://route7.austindavid.com/ is open and free to use. Cost-modeling says "always free" but usage will help tell if we approach any API usage above the free tier.

I was gonna ask about a "better" app for planning rides, but it got away from me and I just made something. Vibecoded, anyway. The app itself is just planning; actual turn-by-turn routing is Maps or Beeline (or anything that can use GPX export). Who wants to help me test and share feedback? The concept is a fairly straightforward maps (Google maps) experience to add a series of waypoints to build a long route. Once the route is set up, it can be shared, edited, or navigated. Smaller routes (<10 waypoints) can go straight into Google Maps. Each leg from here -> Waypoint can be navigated (click -> send to maps). The whole route can be exported, then imported into Beeline or anything that reads GPX. For Beeline it's a few clicks to create a new ride from the export. Mobile experience is the same, works surprisingly well on a handset. The sample route below should be viewable here: https://route7-prod.web.app/route/Bj6Gm4GZKw3U Message me a gmail address and I can whitelist for editing.

Observatory? The backside is great. Y'all have such great roads...

Yeah, that's what I'm thinking; I've got an NC getting suspension now, if I can confirm it's "about right" then it'll save me $150 and half a day. I was expecting my dedeon rear end to be pretty consistent. Maybe I should check my neighbor's live axle charger and see if it has any "toe".

Mine came in over the weekend. It's the "G2" which has some sort of internal hardware and runs an app on my phone. I've reached out for some support, hopefully will hear back during the week. I did a check of my wife's car, which has unknown alignment but I assume it's OK. The numbers weren't crazy, but I didn't save them. IIRC it was something like -0.25 toe front + rear, but very precise-looking values like -0.18* and -0.37* The app has a few modes including a 4-point and 6-point check, plus a thrust angle which uses a body reference. The body reference I chose was the sill where the scuttle hits the "door". I'll assume my car is still square, it's never touched anything. The rear is a dedeon with machined plates, so I further assume the toe is controlled very precisely, tho thrust angle could be off slightly. note that in the photos below the "+1.00*" is a configuration I added, it doesn't actually mean anything for the measurements. here's #1 and #2. The pictures look alarming but the actual values are small. Note that the car didn't move at all during these two measurements, I just repeated the same test twice. The procedure involves holding the phone+device flat (the diamond in the center is your level reference), then hold it against the reference point on the body or the wheel. One of those will blink saying which one it wants next, you align it then tap to confirm. It sends you back and forth a few times to get the measurements before showing the result. I do expect some toe in here, and I think it's underestimating twice. As this is a dedeon I don't feel like toe has much room to wear -- it's steel or aluminium all the way around. I also did the "6 point" check for toe front+back, twice: again, the car wasn't moved between tests. This time the rear toe is more consistent but the front varied a lot. I have not tried to independently measure in 15k miles but I expect the front to be toed in a little. I'd already reached out to the team to see if they could help with the square-check. I don't know if a good set of thrust values would really help here... maybe? My dedeon doesn't have any thrust angle adjustment, I think the two swingarms will constrain that angle. At best I could try and shim the assembly left or right but I don't think the gyraline (as currently provided) can account for that. And It was shimmed + measured when I built it... The above 6-point rear toe check is encouraging, but the fronts and the thrust angle seem ... variable. I'm not sure what the expected accuracy is. I assume something like 0.5 or 0.25 degrees, just based on the swing in the numbers below. But if that's the case I have a front that may or may not be neutral. Based on the very obvious inside wear on both tires after 17k miles I assume they're toed in just fine. I have not yet attempted camber measurement. Again with the dedeon I expect the rears to be very precisely constrained. The fronts were aligned when new and with no accidents and mostly grocery and road trips, I expect the fronts to be about -1.5*.

And sun, and keeps the heat in for the winter. Zip up and you get weather protection too

If you click around a lil you'll find some pics and writeups about the other setups -- @JohnCh and I have traded a few, for instance. Your setup is different than our Duratec-based systems. For the sake of future readers, could you post a couple more photos while you have it open, and of course share your experience when it's all sorted?

Any idea how fuel is getting from the filler all the way forward to your can? From the photo I assume the evap line is coming from the very tip top of your filler neck, then plumbed forward to the evap can. Then usually a purge valve between the evap can and your intake manifold. "fuel in the manifold" is clearly not ideal. In typical operation I understand that the purge valve would open to pull vapors out of the canister, which USUALLY is not a big deal. But if you're getting enough fuel in there to fill that canister and overflow into a catch can (then overflow THAT), something seems not right? The fuel tank needs to have some ventilation -- fuel expands and contracts, and as you empty the tank something needs to let some air back in. I'm pretty sure I have a vent in my filler neck up near the top, but I don't have a photo handy. Either way, I'd want to be sure I understood why fuel was coming out of the tank and down that line; if you block or otherwise vent the same tube that's delivering overflow to your evap can, it will just deliver overflow to your floor or something.