One of my rules of personal conduct is to never say "I told you so".
*Boy* is it hard to do it sometimes.
So long (again), LJ, and thanks for all the fish! It's unfortunate, but things are what they are...
Not that I was running out of reasons to like Mullvad but this just happened: https://mullvad.net/en/blog/2019/8/7/open-source-firmware-future/ .
I'm sure the end result is going to be utterly horrific, but with any luck, it's going to be the last one we really need to put up with for a while. Realistically, there hasn't been any real innovation in this space since SIMICS. If no innovation is gonna happen, how about we at least have just one stagnating, bloated, awkward tool, instead of a dozen? (5/5)
With any luck, the next step in all this game is going to be a "consolidated" IDE, the way basically ARM devices you care about (can) use the ARM GNU toolchain. Silicon vendors go to incredible lengths to save money on software (some of them horrific, but let's leave the depressing stuff for another time). (4/n)
The same goes for the move to gcc & friends.. Yeah yeah it's GNU but do you know what less fun than reaing the GPL? Maintaining a list of compiler bugs and workarounds for your firmware, which we actually did at one point because back in 2015 $highprofilevendor had bugs that hadn't been fixed since 2003. That's, what, 84 years in software years? That bug was so old Computer Eisenhower was still in office when it was born. (3/n)
Things used to be so bad that I was actually glad when everyone left their proprietary, inscrutable IDEs behind and moved over to Eclipse-based (CrossCore, SystemWorkbench, whatever). I sometimes grumble about Eclipse but hey, at least it doesn't crash every day, there's no 16-bit code left in it and oh my god are those anti-aliased fonts? (2/n)
Laugh at silly wannabe hardware hackers all you want, but it takes less time to write a modest program in CircuitPython that it takes to download the nRF SDK over a slightly crowded WiFi connection.
Right, I wouldn't use CircuitPython for production development, either. But this isn't (just) about how much high-level tools for embedded development have progressed -- it's also about how much "serious" tools have stagnated. (1/n)
It just struck me that a (partial!) installation of Xilinx' Vivado is about 15 times bigger than the hard drive of my first computer, and about 2,300 times bigger than the hard drive of the first computer I used.
I bet half of that is just to tell me it can't synthesize my design FOR THE LOVE OF GOD WHAT DID I DO NOW
I first read about Christopher Kraft in Lovell and Kruger's book about Apollo 13, when I was 12 years old or so. That's when it dawned on me that space exploration wasn't just about astronauts. Rest in peace, Mr. Kraft.
The Amiga Workbench had a bunch of different UI styles over the years. Some of them are well known. This page has some that arent: https://www.gregdonner.org/workbench/ .
I like how the biggest and most important thing that IBM needs to do now that the Red Hat acquisition is closed is to ensure everyone that Red Hat is still gonna be just like Red Hat and nothing like IBM because that's just how much people have come to trust IBM. Yeah.
It's hard to imagine this is the same company that gave us the IBM 360...
"Docker protects a programming paradigm that we should get rid of."
I mean, if we go cloud, maybe we should go cloud all the way :). I like this guy.
A fascinating piece on console cartridges with built-in communication systems.
No flash, only DRAM!
When developing open firmware like coreboot and LinuxBoot, or doing research into early boot security, you end up waiting all the time on the SPI flash chips erase and write cycles. spispy replaces the slow flash with an open source FPGA and DRAM controller for instant updates.
*This* is how you hack: an adventure in Palm emulation. This is some of the most interesting stuff I've read in years. http://dmitry.gr/?r=05.Projects&proj=27.%20rePalm
I just found out that Joe Armstrong died and I can't find the right words to say how sad I am right now. I never had the chance to write Erlang professionally, but learning it many years ago had a profound influence over the way I think about my programs, and over the way I think about my profession. I had a great deal of admiration for his work and his approach to engineering.
Rest in peace, Joe Armstrong!
A classic piece on the perils of miscommunicating complex technical matters -- and on the unfortunate pervasiveness of this practice.
I toggle bits. Embedded/systems developer with healthy curiosity towards other fields, computer history enthusiast, occasionally studies drumming, humanities.
Server run by the main developers of the project It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!