At 2022-01-12T09:53:28-0800, Jon Steinhart wrote:
> Ralph Corderoy writes:
> > > I never understood "interpolating a register" to have any relation
> > > to the definition of interpolate that I learned in math.
> >
> > The first definition makes sense of it:
> >
> >     1. (transitive, intransitive) To introduce (something) between
> >     other things; especially to insert (possibly spurious) words
> >     into a text.
> >
> >         in verse 74, the second line is clearly interpolated
[...]
> Yeah, I looked it up too.  Yes, the argument can be made that one of
> the definitions can be forced to sort of fit; I'm guessing that #3
> didn't exist when troff was written.  So I'm gonna stick to my point
> that using that word is awkward and makes the document a bit harder to
> understand.  Especially in the context of programming languages, of
> which troff is one.  To the best of my knowledge, nobody talks about
> "a = b;" as interpolating b.

I've found the term highly useful and have greatly increased its usage
in groff documentation.  (I have been a stickler for a disciplined
lexicon in every software project I've been involved in.)

I prefer it to a popular alternative, "expansion", which is
misleading--especially to novices, who then make the reasonable
assumption, given the everyday meaning of that word, that whatever
results from the process will be larger in some sense than what was
there before.

An argument could be made for the word "replacement", but I've found it
useful to reserve that plain-spoken term for discussion of things a
human might do (perhaps in the course of editing a document or
developing a macro).  To my ear, "interpolation" sounds fancy enough to
refer to something you let a machine do, without being _excessively_
technical in tone.

Regards,
Branden