Bob Diertens <[EMAIL PROTECTED]> wrote:
> This was already there in the "PWB/MM Programmer's Workbench Memorandom
> Macros", D.W. Smith and J.R. Mashey, October 1977.
>
> So there only a bug in the groff_mm manual page.
>
> The same document also mentions the names used by PWB/MM in the chapter
> "Extending and Modifying the Macros":
>
> registers: Aa (most common, accessible to users)
> An (common, accessible to user)
> A (accessible, set on command line)
> :x (mostly internal, rarely accessiblem usually dedicated)
> ;x (internal, dynamic, temporaries)
>
> Seen this, I suggest not to provide aliases for registers :x.
> This only clutters up the name-space for registers.
I consider this an inconsistent argument. If the text
by Smith and Mashey (in whatever version) is considered
to be some sort of authoritative specification, then it
is also part of that specification that these registers
are reserved, and any document that uses -mm must not
access them in any way. Otherwise, the document would
be non-conforming.
As I wrote, I do not regard the Smith and Mashey paper
as authoritative in a strict sense, especially since it
even encourages to use undocumented features. Thus I do
not actually believe that it makes sense to argue with
terms like "conformance".
However, if you consider the text authoritative, then
a document that uses :p with the intent of setting a
footnote number is just as non-conforming as a document
that expects that it can use :p for its own purposes.
Consequently, it does not matter whether groff_mm uses
:p or not.
Gunnar
