On 2014-04-01 13:43, Theo de Raadt wrote:
Then the semantic markup would all be in one place -- in the source
files. There would be no need for seperate manual pages, and the
pipeline could be simply:
ssh-keygen -? -> usage2pod -> pod2mdoc -> doclifter -> docbook2mdoc
-> man
More food for thought.
If whole man pages were embedded in every executable, though, disk
consumption would balloon. Perhaps it would be better to delete all the
current (crufty) man pages, store basic usage() information in the
binary, and encourage users to read the source for any further
information.
--
Matthew Weigel
hacker
unique & idempot . ent