Re: Reason behind semicolumns (i.e. "inline comments") in system output?

Jan J. via Sbcl-help <[email protected]> wrote:

> Is there any particular reason as to why some REPL output has
> semicolumns prefixing it, as though it is trying to comment it out?

I believe it's an informal custom. ANSI requires whatever messages
COMPILE-FILE and LOAD print when VERBOSE is true to be in the form of a
comment, and SLIME implements comment formatting. but I doubt the custom
is very consistently observed.

> I am curious about the source/history of this convention.

I don't know exactly, but here's some "just so" storytelling.

- Lisp development is canonically an interaction with a long-running
  session. A minimal interactive environment would have one channel for
  input from the user and one channel for output to the user.  In such
  an environment, if there's ever a need to transmit information other
  than a user program's "primary" output, prefixing any "non-primary"
  output (e.g., non-interactive error and warning reports, diagnostics,
  timings, tracing, progress reports, GC notices, etc.) with a semicolon
  makes it easier for a user to distinguish things.

- The custom might remain useful even where it's not absolutely
  necessary. For example, even if "primary" and "non-primary" outputs
  could technically go to distinct windows, users might like them
  co-mingled to make the sequential execution order clear.

- It's pretty ordinary to write Lisp programs that process programs. A
  "processor" and "processee" might each have "primary" output. For
  example, a compiler might have something to say about code in a file,
  and the code in the file might have its own things to say during
  compilation. If a processor shares its output stream with its
  processee, having the processor prefix its output with a semicolon (or
  just differently many semicolons than its processee) makes it easier
  for a user to discern origins.

Regards,
Richard