Bottom margin vs. descenders question

[Peter Flynn]

On 25/06/2020 05:22, Doug McKenna wrote:
> Dear typesetting cogniscenti,
> 
> In "official typesetting rules" (TeX or otherwise) are descenders
> supposed to be completely above the bottom margin bounds of a page?
> Or is just the baseline of the bottom line of a full page supposed to
> conform to the margin bounds?

I am not aware of any official typesetting rules for this unless you 
mean Hart's Rules (OUP) or Turabian (US) or similar. It's been many 
decades since I looked, but if you can get a copy you could check. Or 
call someone at OUP (I am asking). Fyffe on Copyfitting is silent on the 
question.

In my experience, when a client checks the type-area of a job nowadays, 
it's from the top of the cap-height on the first line to the baseline of 
the bottom line. It's not exact, but assuming the area contains the 
right number of lines as specified, it is a very close approximation.

Which is slightly different from how it was when type was metal.

The top of the cap-height is indeed *usually* level with the top of the 
metal body (although a few typefaces with swash characters or 
exaggerated parentheses may hang above this). But at the bottom, the 
same needs to be taken into account, so the measurement would be made to 
the bottom of a descender (subject to the same rare exceptions as for 
the top).

If the type is set solid, then baseline-to-baseline will give the 
correct measure. If it is leaded (as in TeX, by default 10pt on a 12pt 
baseline, which is 2pt leaded) then for N lines of 10pt you need to 
allow N-1 occurrences of 2pt leading. In practice, however, the extra 
leading under the last line is retained.

In practice too, on a printed proof, you can just align a typescale with 
the top of a cap on the top line, and see how many lines of type have 
been used for whatever measure you were supplied with in the spec. If 
you have an accurate on-screen typescale, you can usually do it with a 
suitably constructed PDF.

As most block of type are followed by vertical white-space, the kern of 
underhanging swash characters in decorative fonts is probably the only 
circumstance in which you need to be concerned about this. Or, as 
happens occasionally, you have very picky clients, always anxious to 
prove they are smarter than you are, which is almost certainly not true.

Peter