I don't think we can't demand from the community anything other than naive implementations, thus the standard which we've decided to continue of limiting it to ASCII.

During the 1978 session of Harvard's Summer School. The undergraduate college had a Harvard-Radcliffe Student Timesharing System (HRSTS) that was (barely?) able to satisfy demand for general computing power during the normal school year with a monster PDP-11/70 configuration that included a fixed head disk to speed swapping. All in the typical glass cage, with the Unibus extending to a room open to the public room with self-service DECtapes and a line printer. And there was a Xerox Daisywheel in the main terminal room, this is where my focus on backups and word processing started. A rich system with a lot of nice stuff on it, including a default shell that provided help in getting some command arguments correct, probably TENEX/TOPS-20 inspired, the CS department had a PDP-10 as so many others did in that era.

And for a variety of reasons, including the first edition of K&R's The C Programming Language having an awful explanation of arrays and pointers, later the Lions' Commentary on UNIX 6th Edition, with Source Code. Didn't run that code, but the study was invaluable in a community with very little UNIX experience.

There wasn't much of a jump from from Version 6 to Version 7 to BSD2.x, since pretty much everyone was running UNIX on affordable split-I&D PDP-11s with a maximum address space for a program of 64KiB code, 48KiB data, and 8KiB stack (the latter two due to the PDP-11/45's MMU granularity), it was only with the 32 bit macroarchitecture VAX and 68000 based systems that things started to get really wild and diverging.