|
|
|
dogs on the upstairs dock
Saturday, September 20 2025
location: 940 feet west of Woodworth Lake, Fulton County, NY
After I done my usual weekend morning rituals, I took Charlotte on a walk via the Mossy Rock Trail down to the lake, where I only spent a minute or two, and then continued along the lakeshore to the outflow, crossed the creek and went into the woods to the north, eventually crossing East Bifurcation Creek and gradually climbing the hill back southward. After crossing West Bifurcation Creek, the terrain quickly becomes very steep as one climbs the hill that our cabin sits near the summit of. For most of this walk, Charlotte was racing around on her own adventures, though periodically she'd check in. As for Neville, he wasn't at the cabin when I returned, so perhaps he had come on the walk too and fallen far behind.
I went back to working on the issue with my I2C Arduino slave, which also functions as a hardware watchdog. ChatGPT thought my spurious slave-initiated master reboots was being caused by glitches on either the reset line or in the I2C connection. Since I haven't really seen this sort of thing, I was skeptical. So I added a way to retrieve information from the slave about the last master reboots it has triggered. I was already having it store such information; all I needed to do was add commands to retrieve them and some code in the master to grab this data interactively. (The "instant command" system is ideal for this.) With this running, I saw that the reboots were being triggered logically somehow and were not the result of electrical noise. I couldn't figure out why, though, as the master was petting the slave enough that there should've been no such reboots. So I added a way to retrieve how many seconds had elapsed since the last pet, according to the logic, when each reboot was triggered. This produced some illuminating information, as the logic somehow was seeing billions of seconds having elapsed since the last pet, something that was clearly not the case. I asked ChatGPT about this, and it suggested that the problem was my mixing of signed integer values with unsigned longs. Evidently a C++ doesn't do the work to make sure when doing the math between these types that a negative number is not actually a very large number (which, because of the way two's complement works, a negative can be mistaken for).
While I was letting the latest I2C slave code run to look for additional issues in the logs, I turned my attention to the beams I have out on the upstairs deck. I'd gotten a fourth beam recently, and Gretchen had helped me get it up onto the deck last weekend (which wasn't easy, since it is 12 feet long and weighs about 100 pounds). Using a power saw, a chisel, and a hammer, I notched the fourth beam the same way I'd notched the other three. The day was beautiful at this point in the afternoon, and the dogs, who had been on the balcony beanbag nearby, came out onto the upstairs deck to sun themselves (it's not a place they have really hung out on before). But then my tape measure made a scary noise and Charlotte decided to leave. Neville, though, stayed out there with me even when I was making loud sounds with the power handsaw.
Late this afternoon, I grabbed a beer and walked down to the dock and decided to take a kayak out for a counterclockwise padded around the main body of the lake, completely circling Throckmorton the Loon in the process. I noticed that our Oru kayaks (which are about 12 years old at this point) have started to split on the outside surface on some of the creases adjacent to the bottom of the kayak. They don't yet leak, but the splits are definitely a cause of concern. I will have to see what sort of plastic they are made out of to determine how best to repair the material.
Back at the cabin, I fried up a pan of mushrooms, tofu, and onions and boiled up a pot of medium shells pasta, which provided the basis for several subsequent meals as well as light snacking.
A view from the kayak this afternoon. You can see our dock on the shoreline. Click to enlarge.
This loon does not look like Throckmorton. Note the grey near the corners of the beak-mouth. This looks like an older loon. Click to enlarge.
For linking purposes this article's URL is:
http://asecular.com/blog.php?250920 feedback
previous | next |
