You mis-used the word "write". Not that it makes much sense to say that Markdown "writes" anything, and the best guess for what is meant when it is said that a markup language "writes" something is that it includes something as markup when one writes in that language, which in this case it definitely does not.
a2ps isn't writing ASCII and overstrikes, either. Obviously so, as its output is PostScript. It's reading ASCII and overstrikes, like pg, most, less, more, and even the humble ul command; and writing something else entirely. Reading stuff that was destined for old typewriter-like printers and turning it into something else is straying quite far from what strogonoff and I were actually talking about. That's a whole other discussion, and how bad less and more (and indeed a2ps) are at reading such stuff, even compared to what ul is capable of, is a lengthy discussion in itself.
Note that I did explicitly point to the part of groff that is used where it is the "tty" output post-processing, i.e. grotty, which handles what are called in the groff doco the "typewriter-like devices". PostScript and PDF and non-fixed-width fonts are the domain of grops and gropdf et al. and again not the typewriters that we were talking about.
In the context of GNU groff, it is what is written to those "typewriter-like devices" that is what is relevant to typewriters, in that groff (grotty) uses the same idea for the same effect, relies upon the same assumptions about glyph shapes not being crazily different from the norm (which turns out to be a fairly shaky assumption for the middle to late 20th century), and can (as typewriters with a few judicious glyph designs for things like single quotes and such could) actually construct a fairly decent subset of Latin-1 and some other bits and bobs using this technique from the typewriter days.
> in "ascii" mode it overstruck a plus symbol with a letter 'o', one of many such typewriting tricks, which no-one but those printing manual pages to old printers capable of the same typewriting trick would have ever seen
I was pointing out that you don't need an old printer to see them; `enscript` and `a2ps` are perfectly capable of showing them to you. So it is not the case that, as you said,
> Reading stuff that was destined for old typewriter-like printers (...) is straying (...) far from what strogonoff and I were (...) talking about.
The rest of your comment seems to be you getting mad that you found my comments hard to interpret sensibly.
a2ps isn't writing ASCII and overstrikes, either. Obviously so, as its output is PostScript. It's reading ASCII and overstrikes, like pg, most, less, more, and even the humble ul command; and writing something else entirely. Reading stuff that was destined for old typewriter-like printers and turning it into something else is straying quite far from what strogonoff and I were actually talking about. That's a whole other discussion, and how bad less and more (and indeed a2ps) are at reading such stuff, even compared to what ul is capable of, is a lengthy discussion in itself.
Note that I did explicitly point to the part of groff that is used where it is the "tty" output post-processing, i.e. grotty, which handles what are called in the groff doco the "typewriter-like devices". PostScript and PDF and non-fixed-width fonts are the domain of grops and gropdf et al. and again not the typewriters that we were talking about.
In the context of GNU groff, it is what is written to those "typewriter-like devices" that is what is relevant to typewriters, in that groff (grotty) uses the same idea for the same effect, relies upon the same assumptions about glyph shapes not being crazily different from the norm (which turns out to be a fairly shaky assumption for the middle to late 20th century), and can (as typewriters with a few judicious glyph designs for things like single quotes and such could) actually construct a fairly decent subset of Latin-1 and some other bits and bobs using this technique from the typewriter days.