Tuesday, March 11, 2003 Hans Hagen wrote: HH> btw, here are shorter defs: HH> \def\startdisplay HH> {\relax\ifvmode\else\endgraf\noindent\fi\ignorespaces} HH> \def\stopdisplay HH> {\par\doifnextcharelse\par\donothing\noindent} HH> watch the \relax (prevents lookahead) and \doifnextcharelse Good to know, thanks. BTW, have you had a look at the patches I submitted to the list around 10pm yesterday? What do you think of them? HH> anyway, HH> isn't a \stopdisplay followed by an empty line still supposed to be a HH> display? Yes it is, and it conceptually belongs to the previous paragraph, but the following text is not part of the same paragraph and should therefore start with indentation. HH> , so the full model is: HH> \startparagraph HH> text or not HH> \startdisplay HH> \stopdisplay HH> text or not HH> \stopparagraph HH> the startparagraph and stopparagraph can be kind of automatic, but that's HH> just an added bonus for those who want to use empty lines as signals HH> (something which is not common, for instance in xml where proper tagging HH> takes place). Anything that gives a "paragraph end" should work. This is why in the patches I submitted I made a small change: it suffices to give a \noindent, regardless of what comes after, so that: (1) if paragraph doesn't end (by any mean: no \par, no empty line, no </para>), the next text will typographically belong to the same paragraph (i.e. will not be indented) (2) if paragraph ends (by any mean: \par, empty line, </para>), the next text will be indented, since the paragraph end will "absorb" the \noindent. -- Giuseppe "Oblomov" Bilotta