alexfru wrote:These days you have plenty of stuff to play with, but the feel is different. It's no longer something novel or cutting-edge, just a simplified and less powerful version of the PC you've already been using (I mean the Arduino and SBCs with lots of I/O ports, some capable of running a full Linux distro with graphics and all).
I agree. I think there's more fun to be had in software.
Different programming languages offer radically different solutions. Lisp blew my mind despite my only learning a sort of cheap bad Lisp. Forth amazed me with the power of delving into the language, then seduced me with its extremely flexible syntax which is in part a consequence of that power. Even Tcl offers mind-blowing possibilities in playing with variable scope.
You can come up with operating systems completely unlike either Linux or Windows. Plan 9 blew my mind despite its various faults. It's extremely expressive with very little code. You can export *anything* on the network so long as the latency isn't too high. (That's one of the faults.) You can do
almost anything with shell script. (So close!) Some OSs inspired me by providing identical user environments whether on real hardware or hosted under another operating system. Those were Inferno and Active Oberon. Oberon had previously inspired Rob Pike to create the Acme text editor, which in turn inspires me. There's all this network of inspiration, at least half of which is alien to even the progressive projects in the Linux world.
None of these things are new, but they're not part of the mainstream. There's room for experiment. There's room to create new things, whether with them or inspired by them. There's room to try not-so-new things which you just don't normally see, and have your mind blown by them! Then experiment with the concepts that blew your mind!
Personally, I'm taking inspiration and experience from Plan 9, improving it, adding other ideas of my own, and implementing in Forth. I'm also tempted to take on a challenge: Using tricks from Plan 9, how powerful and flexible an operating system can I make for an 8-bit computer?
You know, there is room for experiment in hardware. It's impossible to get near the miniaturization and performance of manufactured hardware, but see above challenge. I think it's possible to hand-build a serial computer faster than 100MHz. The limit there is the memory, making even a kilobit of serial memory out of transistors would be an excruciatingly repetitive task, and the result would be large. Perhaps data could be streamed to and from it with an Ethernet interface. Haha! I forgot all about FPGAs when I wrote that. They would help, to say the least!
