Deri James <d...@chuzzlewit.myzen.co.uk>:
> This seems to be the difference between Ingo and Eric's approach. Ingo is
> correct in saying we should be trying to win hearts and minds of man page
> authors to use macros which include semantic information, but Eric says
> we must stop any man pages which include presentation markup which
> Doclifter specifically can't handle, from being displayable by groff.
This is mistaken; doclifter can already handle almost the full range of
requests found in man pages - though not the full range found in all
troff documents, which is quite a bit larger.
In truth, the request set actually used in more than 98% of all manual
pages is already quite small. It has been shrinking for decades.
One early form of selective pressure began when people started
routinely viewing the pages in terminal emulators. Many troff requests
render badly or are no-ops in that context, so there was a tendency
for them to get removed.
More recently, since 2002, I've been pushing fix patches to simplify markup
and make it less presentational. As a result, there may no longer be any
instances of (for example) .ce and .ul left in manpages in the wild.
I have been very methodical about this. Last year I cleaned up the
manual pages for X - every one of them. The man pages for groff itself
are among a very, very small set that still contain significant
troff-oriented markup. In a full Ubuntu install, just shy of 12000
man pages, there are fewer than a dozen such.
> If I have misunderstood Eric's intentions with regard to the purpose of
> introducing the .hygiene command, then it would be very helpful if he could
> elucidate further.
The reason to write .hygiene isn't doclifter, it's to allow other people
to write specialized man page renderers that could be two orders of
magnitude *simpler and faster* than doclifter. That can't happen now
because every viewer has to be nearly a full troff emulator.
--
<a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>
