On 6/29/20, Ingo Schwarze <schwa...@usta.de> wrote: > I like the idea of converting it to -ms somewhat less. I mean, we > have groff documentation split out into info(1), man(1), and HTML > (for mom) - do we really need a fourth format?
It's more than three currently. The pic(1) manual doc/pic.ms is already in -ms format (and requires pic for obvious reasons). The -me macro package is documented in -me format. Granted, doc/meintro.me and doc/meref.me are historical documents for which there is probably little demand in the 2000s (though enough that someone in 2014 contributed a French translation of one of them). They're certainly not important enough to justify the work of converting them to man or info format just for parity with the core of the documentation. > "What is the difference between groff and traditional roff with > respect to this feature?" is not a question users often need to > ask. True, and its infrequency argues in favor of the information being siloed in groff_diff. But "infrequent" is not "never." I'd say there was actually less need for this file when groff was first developed, when it was the only free *roff implementation available. Now there are several, making interoperability a concern to some users. Directing those users to a document specifically addressing this topic seems like the best approach. Arguably groff_diff could be made even more useful to those users by pointing out where Heirloom has implemented the same functionality but with different syntax.
