[ntp:questions] Re: Latest changes toTWikiGettingStarted

Peter Boettcher boettcher at ll.mit.edu
Mon Oct 6 17:42:02 UTC 2003

hmurray at suespammers.org (Hal Murray) writes:

>>Given that Dave will be using GoLive for his HTML pages and nobody has
>>found a good way to convert these pages into other formats (I'm not sure
>>if the existing html2man script will actually convert the .html files into
>>useful .mak mages) this means we will be supporting a variety of formats
>>for a while.
> One glitch to add the the checklist...  If you use something like
> html2man, what environment does it need?  perl?  Is it reasonable
> to assume the system you are installing NTP on has perl?  Will it be
> the right version?  (Or will html2man run on all versions?)  ...

It's a beast.  It needs:

# This file require the Perl HTML::Parser module, which in turn
# requires HTML::Tagset, which has no dependencies.  [...]

That's why I'd rather have the man pages generated at package time
rather than build time, with copious disclaimers and warnings included
to assuage Dave's concerns about newbies reading (potentially) poorly
converted docs instead of the real stuff.  But I'm gonna try to stay
out of that fight.

I consider the whole thing a hack, which has required some significant
modifications after the switch to Adobe Super-HTML-Author 5000, which
generates it's own brand of crappy HTML.  (Though I probably shouldn't
complain... it's much better than lots of other machine-written HTML).

I'm not interested in pursuing this, but I think the "real" answer
would be to maintain the docs using some sort of XML tools, from which
"nice" HTML and MAN could be easily produced.  (Not to mention PDF for
printed docs, and LaTeX just because).

One of these days I'll do the whole bitkeeper thing and make a few
updates to html2man.

Peter Boettcher
MIT Lincoln Laboratory
boettcher at ll.mit.edu
(781) 981-5275

More information about the questions mailing list