So what about a new architecture for a personal computer that is built around an FPGA or a couple FPGAs (but probably just one large FPGA, or maybe two so that it's possible for it to give itself a "brain transplant") which fan-out into some I/O...
Maybe multiple FPGA's make sense, in a "network" of sorts, but more like a tree so that "CPU" type processors can be realized as well as I/O processors purely through software, all the way down to utility processors like battery charge controllers? Hmm... I don't exactly know the topology of this sort of network or how it would be physically realized without being overly complicated.
Really, a physical architecture which is logically one large piece of FPGA fabric connected to physical I/O connections with perhaps only some buffering or other physical safety measures would be ideal, so that the "physical" (metaphysical) architecture of the machine could also be altered through software (think of microcontrollers that can remap GPIO pins via configuration instead of moving wires). The ideal form of this would limit the need to change the actual physical machine to essentially mechanical adapters that connect GPIO pins to various electrical connectors like header pins or USB ports or sub-D connectors or more fun things like those aviation-style connectors I used on the proton packs, etc. This puts the realm of creating such connectors close if not entirely within the realm of home or small-scale fabrication, making the machine expandable by the people who use it.
So for example a machine like this could be used to create the portable version of my Apple //e + SwyftCard, though nothing but software and hardware that I can fab at home. It could then be re-used for other recreations, or to create something entirely new.
Something like this, while not something I can achieve in the short-term is enticing because it's not limited to one type of machine, or a single idea; I don't need to design and fabricate completely different computers to realize different stages of development. I can use this to simply "clone" an existing design or realize a new one, and every step in-between. I don't need to "settle" on a set of I/O interfaces, devices, etc. because they too can be "written" (and modified) later. A "motherboard" designed this way could be incorporated into many different form-factors with radically different I/O capabilities (a traditional keyboard-monitor-mouse, or a fully-immersive cybernetic VR interface, or a robot, etc.). Virtually everything about the machine could be designed and modified by the people who use it. In many ways, it's the ideal heart for the "Parametric PC" I've been working on for years.
I wonder if I can restrain myself to focusing on designing and realizing this? Settle on some relatively stable off-the-shelf machine for the time being and focus on creating this new machine? It's so easy for me to get distracted and frustrated working with what we have today that I'm always looking for a way to quickly stop that pain; can I endure it in order to realize something like this? Do I have the discipline, the fortitude?
I'm going to have to learn a lot to complete it, and I know that the smart way to do that is to work with others, but what I don't know is if I can convey why this is good, why it's important and induce the same level of excitement in it as I have. It could simply be that I am crazy, and the powerful draw this idea has on me something unreal. There is no way for me to know.
I think the best I can do is see if I can even convince myself that this is worth setting-aside other things to pursue, making sacrifices and seeing the work through. Sharing the progress and the setbacks and seeing if others are drawn to the path. See if I can abandon or suspend my scatterbrained nature long enough to make this real enough that I can realize something I need with it, at which point it will be ready for other to do the same.
Jason J. Gullickson, 2026