Disk Drive Documents - test-drb@ccmp.vtda.org

7 Jun 1999

please see (non-combative) comments embedded below.

Dick

-----Original Message-----
From: Tony Duell &lt;ard(a)p850ug1.demon.co.uk&gt;
To: Discussion re-collecting of classic computers
&lt;classiccmp(a)u.washington.edu&gt;
Date: Monday, June 07, 1999 6:45 PM
Subject: Re: Disk Drive Documents

...
   All of this
suggests that Word documents are a complete non-starter, and
 however convenient PDF is in some ways, it shouldn't be the *only* format
 provided.  I think we're all agreed on that?  I'd also submit that
 PostScript is less useful than either.  There are few systems that can
 handle PostScript but not PDF; all can handle ASCII+GIF; it can be a real 
Hmm... Well, if it's level 1 postscript you can print it on an old Apple
LW2NT printer, and those are pretty cheap and common now. They're also
dead easy to fix :-).

Maybe I was being a little biased, as that's the printer(s) I have here,
so I have few problems with postscript. On the other hand, postscript is
hardly suitable as a medium for distributing documents (IMHO).

Are there really more machines with modern PDF views than ones that can
run ghostscript? I thought that was available for virtually all
modern-ish OSes, and anyway, source is available so porting it should be
possible. 

The problem with PS is that it's not trivial (or achievable with freeware)
to display, scale, search, or selectively print portions of it, is it?  PDF
allows all these.  True, that's no help if you can't run Acroread.

...
   pain
extracting individual pages from PostScript, especially the 
True. I think I was suggesting (but not seriously suggesting) a separate
postscript file for each page.

 PDF is nice because it can preserve the original layout, with diagrams in 
This is great if you're the sort of person who believes layout is more
important than content. It can be a right pain if the user wishes to
alter the layout for whatever reason. 

Perhaps not the art of the layout, but the order in which things are
presented certainly can make the different between a very informative and
easy-to-use document and one which is impenetrable.  That doesn't mean it is
MORE important than the content, but it's important enough.

...
   the right
places etc, but for many purposes having the diagrams separate
 (and viewed in a separate window) is actually a nicer way of working -- 
Agreed...

Actually, I've just thought of something.

I have the shop manauls for some old (1960s) Citroen cars. Like all Citroen
products, these manuals are unconventional.

You open them in landscape format. The top part is then a 'book' of
diagrams (only). The bottom part is the text, which refers to the
diagrams by number.

Actually, this is _extremely_ convenient when you're working. You can
keep a particular diagram open while you read the text that goes with it.
Or turn up a different, but related, diagram.

The same thing could apply to computer manuals. How many times have you
found yourself flipping between a circuit description and the schematic.
I suspect that's one reason why DEC used to publish the printset
(schematics) and maintenance manual (description) as separate books.
Many times I've wanted to keep both open at once.

One other thing to consider. Given separate text and diagrams, it's easy
to (automatically) combine them into one document. It's much harder to
separate them again.

[...]

  There are plenty of document formats that
don't keep everything in one
 file. Why should that be a requirement?  There are probably more systems 
In fact whenever I've written a large document, I've always split it up
into separate files. Makes it a lot easier to manage (like splitting up
source code into separate files, I guess).

[PDF]

  [1] unless I discover that the "text"
within it isn't OCRed or typed, but
 scanned bitmaps, in which case I'll likely throw it away again. 
YES!. A PDF file shouldn't be used as a way of grouping a number of
single-page bitmaps together.

-tony
 I don't like those minimal-effort-PDF's which look like faxed documents,
badly aligned with the page boundaries, and looking like poorly rendered
dot-matrix images either, but I think it's highly important, for the
confidence of the user if not out of respect for the originator, to put
forth a creditable presentation of the original document.  Making it barley
readable means nearly illegible which is nearly worthless.  OCR'ing,
however, invites editing, reformatting, and other sins, which , out of
respect for the document's originator, and probably still owner, should be
avoided.  Besides, having what is, for all intent and purpose a complete
de-facto facsimile (not fax) of the document in question as opposed to an
edited and possibly corrupted copy is a great confidence builder the third
day you're trying to effect a repair in which you NEED the document.  That's
another reason to make the file monolithic in my view.