We'd be better off with 9-bit bytes

Brillant, that 36 bits would be three bytes.

"DEC's 36-bit computers were primarily the PDP-6 and PDP-10 families, including the DECSYSTEM-10 and DECSYSTEM-20. These machines were known for their use in university settings and for pioneering work in time-sharing operating systems. The PDP-10, in particular, was a popular choice for research and development, especially in the field of artificial intelligence. "

"Computers with 36-bit words included the MIT Lincoln Laboratory TX-2, the IBM 701/704/709/7090/7094, the UNIVAC 1103/1103A/1105 and 1100/2200 series, the General Electric GE-600/Honeywell 6000, the Digital Equipment Corporation PDP-6/PDP-10 (as used in the DECsystem-10/DECSYSTEM-20), and the Symbolics 3600 series.

Smaller machines like the PDP-1/PDP-9/PDP-15 used 18-bit words, so a double word was 36 bits.

Oh wait. Its already been done.

If we had 9-bit bytes and 36-bit words, then for the same hardware budget, we'd have 12.5% fewer bytes/words of memory. It seems likely that despite the examples in the article, in most cases we'd very likely not make use of the extra range as 8/32 is enough for most common cases. And so in all those cases where 8/32 is enough, the tradeoff isn't actually an advantage but instead is a disadvantage - 9/36 gives less addressable memory, with the upper bits generally unused.

Actually, the PDP-10 didn't have any byte size at all, it was a word-addressed machine. (An early attempt to implement C on this machine came a cropper because of this.) It did have a Load Byte and a Store Byte instruction, which allowed you to select the byte size. Common formats were Sixbit (self-explanatory), ASCII (5 7-bit bytes and an unused bit), and (more rarely, I think), 9-bit bytes.

My first machines were the IBM 7044 (36-bit word) and the PDP-8 (12-bit word), and I must admit to a certain nostalgia for that style of machine (as well as the fact that a 36-bit word gives you some extra floating-point precision), but as others have pointed out, there are good reasons for power-of-2 byte and word sizes.

But with 5 valued electronics, Up, down, left, right and charm...

You could have the equivalent of 45-bit numbers ( 44 + parity ). And you could have the operands of two 15 bit numbers and their result encoded in 9 quint-bits or quits. Go pro or go home.

Side note: I'm not sure why folks downvoted you. Even if they disagree it seems like an honest question.

> For instance, it's quite difficult to use Linux with IPv6 in a static configuration without any form of autodiscovery of addresses or routes; I've yet to achieve such a configuration. With IPv4, I can bring up the network in a tiny fraction of a second and have it work; with IPv6, the only successful configuration I've found takes many seconds to decide it has a working network, and sometimes flakes out entirely.

On IPv4 I assume you're doing something which boils down to (from whatever network configuration tool you use):

  ip addr add 192.168.1.100/24 dev eth0
  ip route add default via 192.168.1.1 dev eth0

Which maps directly to:

  ip -6 addr add 2001:db8:abcd:0012::1/64 dev eth0
  ip -6 route add default via 2001:db8:abcd:0012::1 dev eth0

If you're also doing a static ARP to be "fully" static then you'll also have an additional config which boils down to something akin to:

  ip neigh add 192.168.1.50 lladdr aa:bb:cc:dd:ee:ff dev eth0 nud permanent

Which maps to this config to statically set the MAC instead of using ND:

  ip -6 neigh add 2001:db8:abcd:0012::2 lladdr aa:bb:cc:dd:ee:ff dev eth0 nud permanent

In both cases you either need to still locally respond to dynamic ARP/ND request or also statically configure the rest of the devices in the subnet (including the router) in a similar fashion, but there's not really much difference beyond the extra bits in the address.

> Challenge: boot up an AWS instance, configure networking using your preferred IP version, successfully make a connection to an external server using that version, and get a packet back, in under 500ms from the time your instance gets control, succeeding 50 times out of 50. Very doable with IPv4; I have yet to achieve that with IPv6.

I have a strong aversion to AWS... but if there is anything more difficult about this for IPv6 than IPv4 then that's entirely on what AWS likes to do rather than what IPv6 requires. E.g. if they only give you a dynamic link local gateway it's because they just don't want you to use a public address as the static gateway, not because IPv6 said it had to be so by not supporting unicast gateways or something.

There's also nothing about IPv6 ND that would make it take longer to discover the gateway from a statically configured unicast address than IPv4 ARP would take, but AWS may be doing a lot of optional stuff beyond just being a dumb gateway in their IPv6 implementation - again, not because IPv6 itself said it should be so but because they want to do whatever they are doing.