On 3/31/20, Ralph Corderoy <ra...@inputplus.co.uk> wrote: > Doug's idea of \s314 being 314, not 31, is better, but the effect can > already be achieved with groff's existing extension of (xx to [yyy...]
True, the functionality is already there. But the principle of least astonishment argues that the bare numbers 39 and 40 not be treated differently. Syntax that surprises someone who's been as involved with a piece of software as Doug should be looked upon with skepticism. It would surely not occur to anyone writing a document in 2020 to say something like "\s65 golden rings" to get a point size of 6. The proposed change quoted above would break historical documents that use such constructions -- but they're poor style anyway: anyone writing with an eye to clarity would already have said "\s[6]5 golden rings" or "\s6\&5 golden rings". (Even the groff info manual, in its sole example using \s with an absolute point size, uses unnecessary brackets to aid clarity: "\s[20]x\s[0]".)
