Lane

What if the core insight of Duff's rc shell (Plan 9, Bell Labs), taking the sequence instead of strings as the primitive object, can be interpreted with virtual double categories as the semantic frame and sequent calculus as the type theory of its morphisms? I'm not entirely sure yet, but I am optimistic about what may lie in this interpretation for the shell, an often maligned setting of computation.

@mralancooper i imagine that we are confronted again and again with the principle that an interface is defined by what it enables just as much as by what it constrains

what if the system shell was a tool that always suffered from not knowing its semantics were best described by sequent calculus?

If we lived in a society that used automation to free everyone from material needs, we might be able to appreciate that having a reproducible example of a god awful/irresponsible coder/user would help us foresee the worst possible mistakes to make in systems engineering. Such could only before be discovered before at considerable expense, or when code was actually being used in live production where mistakes come at a much higher price. Claude Code ironically demonstrates this last anti-pattern

When designing a program that is a dependency for other programs, one must take a lot of care in how much of the overall plan for implementation is realized before it is recommended for public use. Even if further revisions are additive, the kinds of programs that people will make utilizing a system existing at one stage of implementation will vary from what they might make using the more realized system. If your system has two phases P1 and P2: Prog(P1 + P2) != Prog(P1) + Prog(P2) in general