| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 683.1 | DOCUMENT can do better than RUNOFF | CASEE::CLARK | Ward Clark | Fri Jul 17 1987 08:28 | 20 |
| I've always felt that RUNOFF's .TEST PAGE directive was a kludge
because it forces the user to be aware of how many lines will be needed
to format a chunk of text.
I much prefer the technique used by FORMAT (an antique RUNOFF-equivalent
for IBM systems) that allow the user to put "keep" and "release"
commands around any arbitrary portion of text, forcing the text to be
formatted as a contiguous block. There was also the concept of
"floating keep", i.e., keep this body of text together but allow to be
formatted out of sequence.
This FORMAT capability is very similar to DOCUMENT's current support
for figures.
Unfortunately, the <KEEP> tag is already begin used for "keep a small
number of words together on the same line". It would have been nice
for this tag to have been called something like <ONE_WORD>, leaving
<KEEP> for the more general case.
-- Ward
|
| 683.2 | | CUPOLA::HAKKARAINEN | This too shall pass | Fri Jul 17 1987 09:50 | 8 |
| How about <group> and <endgroup> to identify an artibrary chunk
of text?
It's often not lines that we are trying to handle, but a relationship
between two entities in text, a relationship that isn't defined
by other types. (Perhaps <friends> and <endfriends>.)
An optional argument might allow the group to float as needed.
|
| 683.3 | another entry, and generalization of the idea | VAXUUM::DEVRIES | M.D. -- your Device Doctor | Fri Jul 17 1987 10:19 | 18 |
| Or perhaps <unit> ... <endunit>.
Maybe this concept could be generalized to express in what way you
wanted the unit treated specially:
<unit>(noword_break)
<unit>(noline_break)
<unit>(nopage_break)
<unit>(float)
<unit>(nofloat)
.
.
.
The concept applies regardless of the word chosen -- in fact, maybe
<keep>() is as good as any.
--Mark
|
| 683.4 | | AUTHOR::WELLCOME | Steve | Fri Jul 17 1987 11:37 | 6 |
| I'd agree that .TESTPAGE is far down on the list of desirable ways
to accomplish the task, but I totally agree that we need SOMETHING
in DOCUMENT to perform that equivalent function. The <unit>...
<endunit> idea looks good to me. I think it would be handy to have
a parenthesis version as well for small units (such as a word one wants
to keep from being hyphenated).
|
| 683.5 | | MARTY::FRIEDMAN | | Mon Jul 20 1987 14:45 | 3 |
| Do any doctype gurus have a suggestion for a .DESIGN parameter to
force headers to stay with the subsequent paragraph?
|
| 683.6 | <BLOCK> Tag | DICKNS::DENHAM | Jeff Denham, DTN 223-6770 | Wed Jul 22 1987 15:05 | 1 |
| I'm partial to <BLOCK> ... <ENDBLOCK> myself.
|
| 683.7 | | BUNSUP::LITTLE | Todd Little, NYA SWS, 323-4475 | Sun Jul 26 1987 15:33 | 8 |
| re: .5
Are you getting widow headers? The BL8 definitions for headers
called for 80pt of free vertical space on the page at the time the
header was used, and if that much space wasn't available, a page
eject was performed.
-tl
|