Neal Wilson ✅ @neal@social.typica.us

Okay, I'll finish refining this later. Now it's time to go home and make samosas.

Typica supports hardware that can run faster than that, but the AD converter on that particular piece of faster hardware is awful and mis-matched on other hardware so I'm using the extra samples to calculate a linear regression (since the value being measured is expected to be changing) and eventually throwing away the raw measurements used to compute that just to get something that's surprisingly usable out of the 3-4 significant bits per measurement.

Setting the table view to show all measurements and the hardware to collect data at its fastest rate is fine. The data scrolls by too fast to use, but it's only 50 measurements per second and if that's problematic on even an old CPU clocked at over a few hundred MHz that would be a sign that something is seriously wrong. I wouldn't mind a little less RAM usage, though, not that this is a recommended configuration.

Current channels are now almost working in Typica. There are a few things where the behavior isn't quite exactly what I want, but it's basically usable.

@paulhara Welcome to the fediverse. If you have any questions, let me know.

My sister will get the laptop which might end up being a nice usable for her needs machine once it's wiped. My father had so much junk installed on that (which all tries to launch itself at startup) I don't know how he continued to find it usable, but my sister has decent digital hygiene and I think the hardware should be an upgrade from what she's using now.

Once the phone is wiped, the shop is going to use that as a secondary credit card terminal so I can take my personal iPad back.

New hard drive arrived. My father had a lot of photos and such on his devices that we don't want to lose, but we would like to start re-purposing the tech. The best of the photos we have hard copies on as he had a really nice photo printer (one of those that can do big prints with something like 10 separate ink carts), but we want to keep the digital versions as well. That stuff will probably get copied to multiple drives so any family member who wants that can maintain their own copies.

snax named after scooby tied for first, but there was an episode where the gang took a tour of the scooby snax factory and I don't think anybody acknowledged that possibility. Was this to preserve plausible deniability and avoid needing to pay the dog for use of his name? If so, why did nobody dress up as a monster to keep THAT under wraps?

(I'm legitimately surprised by just how large the page count there has gotten. It used to be a much smaller program. I blame having users.)

Yesterday I did a cweave/pdftex on the code I was working on because I wanted to double check the typesetting on a table I added to the documentation. I'd just like to apologize to anybody else who tried to do that for my failure to notice an old typo that caused the last several hundred pages of source documentation to be set in typewriter font. I'm not going to comb through and fix all the typography, but that one bad error is fixed and ready for my next commit.

printer: jam in tray 2

There was no jam.

Was expecting an $8.70 payment. Got $100.00 instead. Unexpected, but welcome.

Which is it?

If MLB was going to pipe old crowd recordings through the stadiums and CGI in the fans for broadcast, why not just find a good baseball sim video game with current players (or get EA to throw something together for them) and just fake the whole season? Pay the players for an NDA about the scam and let them not get sick.

What ever happened to Red Herring? Did Fred ever manage to frame him for something?

Longer term (probably not for the next release) I'd like to do output channels as well. There's some limited support for outputting to Modbus RTU devices, but support for a broader range of output types might be useful if I ever have a good test platform for experimenting with control automation strategies. I have a theory that manipulating heat settings in a drum roaster is mathematically very similar to the problem of aiming WW2 era anti-aircraft guns. (if true, we can do better than PID)

Getting lots of code written today. I'm adding more hardware support for Typica, specifically Phidgets devices with channels that can use the CurrentInput API. I'm probably not going to finish that today, but I'd like to get that and some other stuff done before the upgraded optical head comes back so I can start logging data from that automatically. Before the next release I'll probably go back and add VoltageInput support now that I have an easy way to test that and maybe DigitalInput?