AmpBenzScientist @AmpBenzScientist@qoto.org

@lore In my country they say that opposites attract. I love RISC-V and FPGSs.

It makes a lot of sense with the ISA but I've been told before that I wasn't looking at the correct ISA and that it's not written in stone. Hopefully things have changed because it wasn't easy to work with and making absolute garbage is bad for sales. Performance was surprisingly good but my two experiences were on the more cursed chips.

Just do your research before you buy one. I was hoping for an open ecosystem and I found out that it was actually a portal to hell. If you thought that setting up a development environment was bad, RISCY 5ish was a fresh layer of hell thanks to all those companies involved.

I really despise the manufacturers of my chips and I know they could have been more transparent or followed a standard. RISC-V is a fantastic ISA but the companies were horrible. It could have changed but until I see domestic production and entirely open designs, I wouldn't buy one and expect anything but disappointment.

@lore Performance is going to be lackluster unless one pays to get the highest end hardware. There are a few newcomers like whatever is in the Lichee Pi 4. It's just the first proper T-Head C906 implementation I've heard of. A quad core RISC-V SoC that can out perform the Pi 4 it seems.

SiFive makes their own RISC-V SoCs but they are their own IP. The good news is that they seem to follow the standard. The T-Head cores were opened to the public. They might be a little different than the standard but they were designed for and used in servers and data centers.

It's not impossible to get a good RISC-V CPU or MCU but it definitely takes some research. I think StarFive has a few suitable products for tablets and laptops. The MCUs will be the best and worst category. I will admit that I haven't been disappointed by the performance with the few MCUs I've bought. The disappointment and frustration is getting them functional but they just work and lean towards being overpowered for their intended application. The first RISC-V MCU that I bought can emulate a NES and run Linux. It's a K210.

@lore I wish I could tell you that it is that simple. So RISC-V allows customization, modification and building one's own IP.

It's mostly that simple honestly. The problems arise when Reverse Engineering. They are getting much better. A RISC-V CPU destroyed an Apple Silicon M1 in a benchmark while using less power. RISC-V is getting DSP, Vector and numerous other extensions in the future. The assembly language is fantastic.

@lore He actually met RMS after a friend convinced him to go to an event on campus. That's why he would release his Kernel as GPLv2. I believe it was released in 91 and GNU was messing around with Hurd. GNU saw it and the possibility of making a complete OS was now possible.

Of all the strange things, he met his wife after teaching a course and getting the students to email him. A woman asked him on a date in one email. I'm pretty sure that is not acceptable even in Perkele land.

He scuba dives too. I wonder if Torvalds and Freemo will ever meet.

@lore Android has changed quite a bit over time. There is a list of companies and products that haven't properly released their source code. I presume that Android has gone with more "permissive licenses" to mitigate the damages by the companies who think they don't have to abide by laws. Android has also cracked down on bad practices and support in the past.

Apple does release code but it's due to licensing if I recall correctly.

@lore Microsoft saw that they were losing developers to Linux and had to come up with something. Essentially they had to look friendly because they were losing ground. This is where something called WSL allowed people to "use Linux" so they wouldn't actually use Linux. They are still the company that made computer manufacturers implement UEFI and only allow Windows to boot. It was claimed to be for security but the only secure device they have made is the Xbox One. It has still not been broken. It would be a nice little Linux box because of the 8 cores and fast ram but they are content with E-Waste.

@lore Arduino has redeeming values. I'm glad this isn't about the RaspberryPi. It's been a disappointment for it's entire life. There's a large community surrounding it but nearly every computer before it was a single board. Prices are high because of scalping or poor planning. It requires the same amount of money in equipment as its retail price to work. For something so targeted at developers, it has horrible documentation and a tendency to not have any safeguards to prevent accidental frying. Open Source? Nope, they just rode off the hype to sell products.

The Pico is well designed. It's the most impressive thing that RaspberryPi has sold. It even looks like a real Engineer designed it.

@lore I didn't think he was unpleasant. I was a member of the FSF. I will say it would be unpleasant to be stuck in an elevator with RMS and Torvalds.

Torvalds was also one of the first people that I know of that was important to Computer Science and wasn't an Engineer, Mathematician or Physicist. Bill Gates, Steve Balmer and RMS were Mathematicians. Wozniak was an Electrical Engineer and Computer Scientist.

Torvalds would be unpleasant if he was incorrect. His stubbornness about the license of his Kernel was probably the right call.

@lore Microchip PIC was used in The Stamp I think. The Stamp was used to teach Embedded Systems Programming before the Arduino came along. The lower cost of the UNO and realistic language choice of C made for a more realistic experience if one would go on to work for a company that specialized in Embedded Systems.

The reality, especially for open source developers, is that they have next to no documentation and have to essentially wait for Reverse Engineering of the hardware, firmware and software to be completed so they can use these tools to do what they wanted or learn that they can't do it.

FLOSS developers might seem like pricks in many ways from how they tend to explain things to using a license like the GPLv3. The reality is that every step of the journey was a struggle and companies love to sue. So when something is reverse engineered and released, the GPLv3 is the choice of someone who endured the mountain of red tape, threats and hardship. No one will be able to steal their code because of the GPLv3 Superior Clause which means that the company has to stop use of the code or they will be required to release all of it under the same license.

@lore I think you would be pleased to know that MIPS now produces RISC-V MCUs. Maybe they realized that no matter how impressive it seemed on paper, they had a dumpster fire for people that had to work on them. I will admit that they were impressive and I have seen some of the most professional Low C code used on MIPS.

@lore I try not to complain as much as I used to. Do I have gripes about working with embedded systems work? I don't right now but I haven't been trying to resolve dependencies for a few hours while trying to figure out what device is better to communicate with the device in question.

Part of the reason why Embedded Systems are not always well documented is because someone had to bang their head against the wall until they could get it to work. It usually takes a few people to refine the documentation and get it to where people can use it. This is because they probably despise the hardware and they get shit on for their hard work.

@lore Yes I can flash a ATMEGA328 bootloader but for everything else it was completely worthless despite supported hardware lists indicating that it should support what it did not.

I just use Bash and GNU Emacs. I would rather just use Emacs with documentation handy to write a program for a board and use a Bash utility to flash it. It seems like it would be a problem but they rarely ever fail me. Bash is incredible with the right utilities. Having a Black Magic Probe just adds to the experience.

In all honesty, I believe Arduino Studio is counterproductive especially if one is learning. Realistically everything else will require another IDE and using the terminal. It's better to get used to this early on instead of getting a false sense of competency just for it to be crushed by something that wasn't designed for children.

@thor I don't think Ska has a museum and it's not cool. Fedora tipping intensifies.

@thor That would be oddly intimate but when in Rome...

@lore With some areas it's incredibly niche and stepping off the path results in not having much help. The amount of digging for information, research and experimenting is considerably large. That's just for a prerequisite to a prerequisite to finding the solution.

In my experience it's best to think of it as many small projects. I'm not good with documenting my progress. It looks almost like I have no clue what I did because I will fail to include parts of a month or so of effort.

Start> time passes with unintelligible comments being made >the work is released without any context. I have to go back and explain what I did and why I did it after I recover from the sleep deprivation and collective frustration of the work.

Perhaps what I'm trying to say is don't be like me. 72 continuous hours is not an acceptable amount of time to spend on a project especially when it is a regular occurrence. I have learned my lesson, I at least take notes now.

It's good to allocate a limited amount of time to a project per day to keep morale up. The time spent not doing that thing is when really good realizations are made. Never give in and never give up but remember that it's a project. Projects require patience and pacing to be successful.

Furthermore, you can always find friends along the way. There are many people who enjoy exploring similar things and they could easily be tempted to explore something new. Just be aware that motivations vary so the friends can almost be trusted to go on tangents but it doesn't mean that they no longer care.