We'd be Better Off with 9-bit Bytes

The post argues that 9-bit bytes would have made the computing world easier in several places. With 36-bit IPv4 addresses, 36-bit UNIX timestamps, and 18-bit Unicode characters, many of the exhaustion problems we live with now would have arrived much later or not at all.

Reading notes

#

• 70s systems such as the PDP-10 used nine-bit bytes, and the post treats that as a plausible base for a different standard.
• A 9-bit byte world would give IPv4 36-bit addresses, enough for far more devices and users before exhaustion.
• 32-bit UNIX timestamps would become 36-bit timestamps and would not run out in 2038.
• Unicode would have more room, with 18-bit characters instead of 16-bit ones, making the available repertoire much larger.
• Pointer sizes and memory limits would shift too, with 32-bit operating systems allowing larger per-process address spaces.
• Other effects mentioned include roomier AS numbers, ports, process IDs, and user IDs, along with changes to instruction encodings and character sets.
• The main cost discussed is TCP sequence numbers, where 18-bit space might be too small for high-bandwidth connections and could force an earlier protocol upgrade.
• The post ends by suggesting that many early computing limits were shaped by numerological choices rather than hard necessity.