(There are a lot of tracking improvements in the development queue, but even one or two feature requests from free tier users would be enough to bump something away up the priority list.)

This basically describes a product my $dayjob sells: dojofive.com/embedops/

RAM/Flash tracking over time works well, but it doesn't quite have the interaction with the CI graph to get PR-level resolution yet. Similar metrics for unit tests, HIL tests, SBOM monitoring, etc.

We're gonna be the ones bringing up that next generation!

I'm worried about the amount of institutional knowledge we'll have to struggle to hold on to, but that's also an opportunity to let go of some ideas that deserve retirement!

Those of us who happened to end up in that sparse middle are going to have some work to do, though

I'm pretty sure I still have a Galileo around here somewhere, imagine if that monstrosity had kept going

At least I also got to discover a bunch of bugs in my own company's software, as well!

And that's on top of a bunch of items between mcuboot, the Nordic qspi driver, the qspi xip driver, and nrf70 xip patches having various conflict points and poor documentation

Zephyr spdx doesn't work with Nordic, and Nordic's response is that one should use their ncs-sbom plug-in. Except that plugin doesn't work if you use a nonstandard C library, or link time optimization, or look at it funny...

Imagine a town with only burger joints and one or two salad places 20 miles away. It would be hard to say the evidence is overwhelming that the majority of the town’s residents prefer burgers to salads. To me, that’s the suburban/urban housing preference debated distilled to its essence.

The overwhelming evidence of the past 100 years also includes redlining, white flight, government subsidies that enabled the growth of the suburbs, the construction of the interstate highways system… that most people now live in suburbs is not a factor of preference alone.