
-  move shade1 and leader to webindex.azm; remove style.css dependency.
   Improve css classes etc.
   instead of <style="text-align:justify"> use <p class="j"> etcetc.

o  faq.zmm:
   there is a <br> at the end of \ans. Doing this with <div
   style="margin-bottom:1em"> inside the <td> element failed miserably in NS
   -- at an unexpected place, namely too much ws at *beginning* of the
   faqsec-table.

o  zoemtut:
   there is a double <br> in html itemize env. single sufficient?
   instead of the <br><br> to create paragraph skips between dt
   entries, wrap each dt entry in a div, give it a class, and set
   its margin-bottom attribute. Well, try at least.

o  can I use .ns and .rs in troff for verbatim environment?

o  after vbt env, no need to enter \par for html.
   How is that for troff?

*  TP not good voor toc in nroff -> paragraphs skip.
   need sth else. using tab positions now: it will not wrap.
   can copy the solution made for faq.zmm.

o  A space at EOL could be unpleasant in troff. Confirm/refute.

o  netscape sfcks because it does not correctly remember the align setting
   if you do X, where X is probably setting the aligning locally somewhere
   between <div></div> tags.

o  should try to modularize the html macro's as much as possible.
   did this for <a name=..> and <a href=...> (\_htmlanch and \_htmlhref).
   Same for roff, but less urgent as roff support is not as complex
   as html support.


*  [  when using \fB in the \bf{} definition, no \n should occur in
   [  the argument.
   [  ? is that true ?
   [  how can I prevent this from happening?
   [  how about a directive: forbid char X until revoke? (mmm. heavy)

   It is only partially true. It may lead to unwanted paragraph
   skips in troff if the '\fB' or '\fP' part ends up on a line by itself.
   This would be cured by escaping the newline.

*  Related:
   How to prevent context sensitive stuff like "\n." in troff.
   Right now I map '.' always to '\&.' -- which is ok I guess.

o  consistent typesetting of symbols and arguments in docs.
   clmdist uses bf. Should be sth else?
   italic for generic arguments 'k'.
   typeface for actual argument, '5'.
   ???? for mathematical symbol.

