But I doubt their transistors are much smaller than the original 6502?

Wiki says the original 6502 used an 8um process. The current CMOS W65C02 and W65C816parts are 0.6um according to their datasheets.

For the stand alone chips, there's no real motivator to use the smaller die processes

Welcome rupy! You might ask, why are smaller transistors desirable? The usual reasons are: reduction in cost, and increase in clock speed.

For the stand alone chips, there's no real motivator to use the smaller die processes.

There are certainly 6502 done with modern processes, just not stand alone chips, they're 6502 cores bundled with something else.

For me there was a big light bulb that went on when I realized that a TSC instruction followed by a TCD puts your Direct Page on stack. Suddenly Direct Page ceases to be a crowded place! And you get some amazing new address-mode flexibility -- for example being able to use a three-byte indirect pointer that's on-stack.

That was yesterday, today my brain is racing on the fact that understanding the complete package from hardware to software is not that far from my brain to do, specially if it's limited to 8-bit. I now even have a basic understanding of the Gerber file format.

And there is so much that I would like to improve in the computers of today. Here is a coarse list:

- HDMI/DVI are bad because they encrypt data and you really want G-sync but without licences. Display drivers have scanline legacy that makes no sense with a IPS LCD. The screen should have hardware v-sync buffer built in and draw immediately when the GPU serves a complete frame.
- USB is too expensive and generic (USB3 is just broken), you want the latency on the keyboard/mouse input to be even lower. Only keyboard/mouse input is interesting for now, everything else is a waste of time. I2C or SPI are better standards.
- You want P2P LoRa instead of Bluetooth and WiFi (again licenses). If you need speed you should use a cable.
- File systems are really bad, they have magnetic disk legacy drivers. They do not allow better indexing and searching natively (even if EXT4 made improvements). We need indexable linear compression. A small SLC NAND disc should be in the SoC for the tiny operative by default, you do NOT need a GUI for the base OS. That tiny SLC NAND should be extendable without the SoC needing to "mount" it or anything. Disc over the network should be a native thing.
- OpenGL is really bad, it assumes you want separate memory between the GPU and CPU. Shaders require very verbose communication between the CPU/GPU for now established standard things like skin mesh animation.
- Sound should be "vectorized", or a compilation of wave forms instead of discrete recordings. Think SVG for music.
- Memory should be level one memory all around (Steve Barlow of RPi fame was kind enough to explain the reason for the different levels of memory to me but I am convinced there must be some solution by integrating memory into the SoC if it is expensive and large so be it but we need to avoid cache misses at all costs, the programming languages are all very bad because either they are performance closed (Java/Swift) or they have poor fragmented compilers (C/C++).
- Everything should be integrated on the SoC with some better bus between everything, for example; the GPU should be able to do networking. So each GPU core could optionally handle multiple sockets to offload the kernel for servers where the GPUs are mostly a 3rd wheel today.
- Ethernet is bad and expensive. There should be a Internet 2 that is open and works both over Gb fiber and P2P LoRa. HTTP, TCP and IP are the final network abstractions, UDP is not needed because you can add a ackfreq property to TCP.
- Most new standards are over engineered: HTTPS, WebSockets, HTTP/2.0, IPv6, Vulcan, the list goes on and on. IPv4 should be extended with the internal IP just like the IPv6 workarounds do today, it will happen that way anyway, just take longer when you don't spell it out!
- Patents and licences are a waste of time. Ideas should be open and free.
- You need new hardware and people are afraid of that because of the arbitrary invention of money. Money and laws do not exist! There is no spoon.

The real reason for all this commotion is that we have peaked Moore's law, now it's the time to drop the legacy dependencies and build a first attempt at a "final" computer with everything in the SoC so that we have a licence free, open foundation for humanity to maintain truth and trust when stored solar energy has peaked. I think that peak Moore's law and peak-oil are actually the same thing, at least they happen at the same time and I don't think that is a coincidence.

I know this is a mouth full and laws and money make you instinctively pull back from it like if it's something dangerous. But we are all going to die, the only thing that matters is what you do before you die, and I'm not going to sit here and wait for us to destroy any hope of civilization we have left. We "the nerds" have to take control.

In practical terms, I will continue to learn, and one step in that process is probably to build my own breadboard computer with your W65C02S.

So no I don't have a "commercial" product in mind, I'm trying to build something that transcends the fiat we use to wage war. Something that outlives all bureaucracy. Society is as over engineered as our computers are becoming. Vulcan is like filling out your tax return.

Eventually I will need to build this thing with peak Moore's law transistors. But I don't know when, I'm just putting thoughts out there.

Maybe Moores law will continue and maybe we'll have fusion, but I'm not holding my breath.

And just so you know solar panels, wind turbines and electric cars are not going to save us, because electricity is not an energy source. Only photosynthesis has a free lunch, only that lunch takes millions of years to be served.

Thanks for replying with such openness.

Kind Regards,

/marc