X-VM-v5-Data: ([nil nil nil nil nil nil nil t nil] [nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil]) Date: Mon, 19 Mar 90 23:23:34 GMT Reply-To: LaTeX-L Mailing list From: CA_ROWLEY@VAX.ACS.OPEN.AC.UK Subject: Sections, generic headers, and other things To: Rainer Schoepf Status: R X-Status: X-Keywords: X-UID: 54 No reactions yet to my ideas of last month on how flexible/generic sectioning should be for the designer! Do I assume that you all agree with me, or think my suggestions are too dumb to be worth reacting to (apologies to Don H for my appalling style of English) or has everyone (very wisely) decided that they, like Rainer, have better things to do in Spring than tap away at their terminals. I have: I am off "on vacation" for a few weeks from Thursday. But before I go, here are a few thoughts on Frank's \genericheader message (which has also attracted zero comment) ---some specific and some general. 1. In addition to all the logical information which is not handled at all by it, here are some typographic features which it does not seem to cover: some would be easy to incorporate but others, particularly those concerned with special features of the first line(s) of the first paragraph of the section being started, may not be so easy to incorporate into this designer interface. Please note that I am NOT saying that 2.10, or even 3.0, must cater for all these possibilities, and I am certainly not claiming that they are good typographic practice. I do, however, think that we should be clear about what we are NOT providing as well as about what we are providing so that we can produce clear documentation. a: Special treatment of first line(s) of section: The following are amongst things I have seen or heard about, there may be others, and they may be combined (some apply whether the heading is run-in or not). Other indents (including negative), eg the first n lines may be set to a wider/narrower/displaced measure, possibly lined up with the headline itself. Dropped caps. Special font for first line(s) (see TUGBOAT Vol 8 no 2 july 87). Heading set in "a bite out of the text" either out of the first few lines on the right or the left or (ugh!) the center! Extra space after first line, especially in run-in style. b: Many column headings in multicol text (very common). c: Do we need to make it possible to override the autonumbering (other than by resetting the counter and/or \the). Some specific detailed comments on: > \generic@header{LEVEL}{NAME}{SPECIAL CMDS}{NUMBER LAYOUT} > {HEADING LAYOUT} > LEVEL and NAME as in LaTeX As I have already argued (in my last message to the list), more is needed here. > SPECIAL CMDS contains declarations and actions to be executed just > before the header and its vertical before space. > At the moment I suggest the following cmds (my names > are more or less arbitrary): It would be better to separate assignments/declarations from actions (ie typesetting actions), see below. > Switches: > > \@afterindenttrue to get indentation of the paragraph following the heading. > (default false) > \@runintrue to get a run-in heading. (default false). > Of course this implies \@afterindentfalse. There are a lot more possibilities than just indenting or not, see above. > I can think of \clearpage or something like this but there are perhaps other > uses aswell. This should be in the "actions" parameter. > Thinks like rules however should be placed (in my opinion) into > the argument HEADING LAYOUT. OK for rules that are part of the heading, but what about those which are part of the ending of the previous section, the ""trailing??"". > This would also be the place where \thispagestyle could be used, of > course only are something like \clearpage. > Does anybody think we also need an argument for things after the > heading? More than just an argument, see above. > NUMBER LAYOUT This argument should contain the layout for the object `number' > i.e. everything which is attached to it. > The kernel will pass it back in \headernumber to the style > if a number should be produced. Otherwise \headernumber > will be empty. Should it be passed back in an \hbox? Does the actual text to be typeset get resolved at this stage? Does the value of the counter get resolved at this stage? > HEADING LAYOUT This is the playground for the designer > where he has to specify the exact layout > of the heading. The heading will be a > vbox if it is not a run-in heading. And in an \hbox if it is "run-in"? Also need to allow for "run-in unless more than a given length", how does this fit with a single parameter \headpostskip?? > \@ifpresent{}{}{} Sounds generally useful: I am assuming it tests whether expands to "empty"? > \@iffitts{}{Test dim}{}{} > This test allows to make certain actions depending on > the length of a trial text (like header + number). Good: May want to test the vertical size as well as the "total length", but perhaps you meant that? 2. More general comments (ie not related to sections alone): A: Typographic specs (at least in this part of the galaxy) commonly use baseline-to-baseline measurements: a designer interface therefore needs to make these easy to use directly. B: It would be nice if the designer could be assured that the assignments made in an argument which is purely for this purpose, eg possibly the SPECIAL CMDS one in \genericheader and, possibly, the one in \list, would be executed in an "Assignment Environment" (ie with end-of-line and empty lines made no-ops, plus, possibly, space made a no-op so all typographic spaces must be explicit). C: Also, will 2.10 be \protect-free (or just the "redone" bits?) Happy Easter to one and all chris