| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 382.1 | I could try to flag it | IJSAPL::KLERK | A doctype in every pot (B.K.) | Tue Nov 06 1990 08:39 | 18 |
| Sounds like another thing I could add to the converter to make
"usual" SDML files processable for Online books. I'll keep it in mind.
Theo
P.S.
This relates to a topic in the DOCUMENT notesfile, where I mentioned
a possible converter to "correct" SDML files for proper book reader
processing:
- adding labels where omitted in sectioning tags and "hotspot" tags
for tables, examples and figures
- adding <figure_file>(book\ ... ) entries for <figure>s, based on
other <figure_file> specified destinations
- adding <reference> tags following "hotspot" tags where they were
omitted from the text (to avoid figures and such to become inaccessible
from the main text).
Seem like doing "hierarchy checking" is required too.
|
| 382.2 | what is 'usual' ? | OLD::UTT | Mary Utt | Tue Nov 06 1990 09:51 | 20 |
| Well...
1. IF skipping header levels is a reasonable thing to do, then the
Bookreader should handle it more gracefully -- more like .PS output
where the output is fairly ugly (e.g., 1.0.1) but at least there's
output and the writer can decide what to do about it.
2. If skipping header levels is NOT a reasonable thing to do, I think
that the tag trans should report it as an E-level error. For
VAX DOCUMENT.
For other authoring tools, the Bookwriter Routines (which all
authoring tools use to build .DECW$BOOK files) should also catch
the error. (The authoring tools may chose to catch it earlier,
too, but that's up to them.)
My question here is 'what is *usual*' ? Then we can start to figure out
how to deal with it.
Mary
|
| 382.3 | one 'nay' vote | OLD::UTT | Mary Utt | Tue Nov 06 1990 12:34 | 21 |
| Mail response to this query:
From: BOOKIE::RANDALL "Put me in coach, I'm ready to play" 6-NOV-1990 09:25:09.45
To: OLD::UTT
CC:
Subj: coding DOCUMENT header levels (ref. online_bookbuilding note 382)
Mary --
Please don't stop processing after tag translation! I can't think of
a good reason to skip header levels in a final version, but it's a
common error during draft stages, and there are already enough
problems getting a rough draft to print out. And it's sometimes hard
to find out where the missing header should go without looking at the
output text. Simply bumping the next header up a level is frequently
not the right thing to do.
A W-level error, perhaps, but not another killer. Please!
--bonnie randall schutzman
|
| 382.4 | Strict ordering only works for numbered heads | XLSIOR::OTTE | | Tue Nov 06 1990 16:53 | 26 |
| Hi Mary,
Aside from coding errors, I've never seen a good reason to
skip numbered heading levels--ie, from <head1> to <head3>; however,
if you are talking about making all heading-type tags follow
a hierarchical ordering, then you may have some trouble.
I've seen plenty of cases where a <head1> tag was followed by a
<subhead1>, <subhead2>, <head>, or <chead> tag, in cases where the
writer wanted a heading without a number.
If you want to enforce strict ordering of headings with the tag
translator, I'd suggest doing it only for numbered headings
(which I think are the only ones that the bookreader cares about anyway).
In the hardcopy documentation, when DOCUMENT goes from a <head1> to a
<head3>, the TOC would mark the missing header as 0 as follows:
Chapter 1 xxxx <-- <Chapter> tag
1.1 Introduction to xxx <-- <head1> tag
1.0.1 Details on xxx <-- <head3> tag
Could Bookreader do something similar and report the problem as
a warning: "DOC-W-MISSINGHEAD, Numbered Heading tag missing"
-Randy
|
| 382.5 | | VIA::EPPES | I'm not making this up, you know | Tue Nov 06 1990 18:24 | 8 |
| I like Randy's suggestion (.4). I can't imagine skipping a numbered header
level on purpose, but people have been known to do some, um, creative things
with DOCUMENT.
Like Bonnie, I would prefer that processing not stop (although I tend to compile
my SDML files in LSE first, anyway).
-- Nina
|
| 382.6 | another 'nay' vote | OLD::UTT | Mary Utt | Tue Nov 06 1990 19:43 | 26 |
| From: RTL::SHIRLEY 6-NOV-1990 14:14:55.13
To: OLD::UTT
CC: WORDY::WILLISON,SHIRLEY
Subj: Allow head2 without head1 in a document
Mary,
I got a copy of your note regarding a <head2> heading without a <head1>.
I would suggest you let the heading numbers be out of order and just make
it a warning when the file is processed with DOCUMENT.
I have been using DOCUMENT for a few months and compared to some other document
processors its great. However, I feel it is too restrictive when one is trying
to develop and plan documents. The warnings should be handled in
some phase of a book review cycle rather than making DOCUMENT crap out on a
user when they are just trying to use DOCUMENT as a tool.
DOCUMENT can be a great development tool but if it quits too soon the user is
unnecessarily frustrated. I have worked in great detail on an
outline of a document, and many times I just need to see what the organization
of the book is about. I may be talking from a programming viewpoint but I
prefer warnings.
I hope this perspective is useful.
Regards, John Shirley
|
| 382.7 | | MJFITZ::FITZELL | got those multi authoring cross platform blues | Fri Nov 30 1990 13:12 | 29 |
|
Thanks for the feedback but I'm not convinced that the error should be only
a warning.
What I'm hearing here is that it's considered an error in the documentation but
we should only flag it as a warning so the build can continue to facilitate the
write-review-edit cycle. The problem with that is there is no way for the
authoring tool to know if this is a preliminary version or the final version.
(and even if there was there's no way to enforce that the final version was done
in final version mode)
It was mentioned a couple of times how this is done in hard copy and even
though it's not pretty it can be acceptable. Unfortunately for online it's
quite noticeable and it's sure to generate QAR's. It looks something like the
following.
([] = box)
[]Chapter 1 xxx
[]1.1 Introduction to XXX
[] <--- Box with blank space after it
[]1.0.1Details on XXX
I don't mind letting the build get through to the Device conversion stage so
you can see all the error messages but I do have strong objections on letting
it actually produce a .DECW$BOOK file that has an error in it.
Mike
|