[Q] TFM files headers

[barbara beeton]

On Fri, 13 Sep 2019, Taylor, P wrote:

> barbara beeton wrote:
>       On Fri, 13 Sep 2019, Taylor, P wrote:
>
>             Tomas Rokicki wrote:
>
>                   A google search immediately returns
>                https://en.wikipedia.org/wiki/TeX_font_metric
>
>             which has the appropriate reference to tftopl as
>             well as being a
>             pretty easy read itself.
>
>             Do you have any idea of which iteration of TeX David
>             was writing, since TeX82 would require "\font
>             \A=CMR10" whilst TeX78 would have required (as David
>             himself notes)  "\:A=CMR10". 
>             Philip Taylor
>
>       The article by David Fuchs is from TUGboat volume 2 (1981),
>       so TeX was in flux.  In fact, in the first paragraph, there
>       is this sentence:
>
>         For instance, when you say \font A=CMR10 to TeX (\:A=CMR10 in
>       the old lingo), ...
>
>       so that should settle that question.
> 
> Well, it didn't settle it for me, which is why I asked !  It was clear that
> \:A=CMR10 was from TeX78. because the first thing I did on reading that was
> to take my copy of TeX and MetaFont and look up the syntax of the
> \font-equiv command, which was exactly as reported by DF as "the old
> lingo".  But it was (and still is) unclear which version of TeX
> required/allowed \font A=CMR10 rather than \font \A=CMR10.  Now you say that
> "TeX was in a state of flux", but this is the first suggestion that I have
> ever read that TeX did not go direct from TeX78 to TeX82 — are you saying
> that there were intermediate (but unnumbered/unnamed) versions, and if so,
> were they found "in the wild" or only within the safe haven of Leyland
> Stanford University ? And having declared \font A = CMR10, how did one then
> use A, and why did words containing "A" not cause trouble ?

I'm not aware of a version of TeX with that syntax ever got loose.
But if \font were defined to take a single token, then "A" wouldn't
necessarily have to be \catcoded to something other than "letter",
so I don't see any problem with ordinary words, or, indeed, \cs-es
containing that letter.  But I'll think about it some more.
 						-- bb