Gunnar Ritter <[EMAIL PROTECTED]>:
> > See my long reply to Larry Kollar. It's not clear to me that anything
> > interesting can be deduced here, but I'm open to suggestions. What
> > kind of semantic-level tagging could we use in this situation? Would
> > <blockquote> be the right thing here?
>
> Perhaps .in or .ti should trigger similar actions to .nf, e.g. to
> examine the following line(s) whether they look like some kind of
> sample code or program output.
Actually, I already know that .in does not convey significant
information that neighboring tags don't, because I've audited for this
in my test corpus of > 13,000 pages. (Known exceptions to this rule
are emacs.1 and compress.1.)
Your thought about .ti is very much in accordance with the way the rest
of doclifter works. I'm now running a full test on all 13,000+ pages with
.ti moved from the ignore set to the complain set. I'm going to eyeball
all of those instances to see if I can spot usable patterns.
--
<a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>
_______________________________________________
Groff mailing list
Groff@gnu.org
http://lists.gnu.org/mailman/listinfo/groff
