Xref: utzoo comp.text:8050 alt.hypertext:750
Path: utzoo!utgpu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!cbnewsl!floyd!saal
From: saal@floyd.ATT.COM (Sam Saal,1A-110,6956,Cap Gemini)
Newsgroups: comp.text,alt.hypertext
Subject: Hypertext and CASE for on-line docs (long)
Keywords: CASE hypertext existing documents
Message-ID: <1991Feb22.154333.10409@cbnewsl.att.com>
Date: 22 Feb 91 15:43:33 GMT
Sender: @cbnewsl.att.com
Reply-To: saal@floyd.ATT.COM
Organization: AT&T Bell Laboratories
Lines: 63
Followups-To: comp.text

Sorry this is so long; please respond via email as I don't read all the
newgroups.  I'll post a summary after a week or so.

My project is considering implementing a Hypertext interface to our
documentation.  We will have to convert several documents to Hypertext
and I have some questions about it.

From my understanding about Hypertext, what works best is information
"chunks" no larger than what would fit on (at most) a couple of
screens.  Further, each of these chunks should maintain P/PI
(Page/Paragraph Independance).  That is, because the Hypertext
developer may not know how the user got to a screen (which link,
sequential or otherwise), each screen must give enough of a context to
keep the user from having to start back at an earlier point in the
Hypertext document.  This is especially important for supporting
information (as opposed to the root document).

IN addition to the significant effort level necessary to convert
existing traditional documents to Hypertext, this presents some
interesting problems For example:

Some of the documents are deliverables, "legal" definitions of
requirements.  Converting them implies rewording some things and that
means the Hypertext version could not be used for arbitrating issues
we'd normally be interested in.  That is, by having two different
versions, and someone needing to decide whether a requirement was being
met, we could not use the on-line, convenient, Hypertext version and be
sure we had the official requirement.

With that huge introduction, here are my questions:

Is the problem I mentioned really something to worry about?

How important is it to maintain the small chunk-size?

What experience do people have converting existing documents to
Hypertext?

Is it worthwhile to convert documents one at a time?  If you do it that
way, each time you add a document to the Hypertext you'll have to go
back to each document already in the system to add links to the
additions.

If existing traditionally published documents are converted to
Hypertext and then reissued, are there fast/easy ways to redo the links
in the new issue or must youstart all over?

One possible solution to a few of these probelms is to look to a CASE
tool as hybrid solution.  CASE has some of the links between source
documents, although they are far more structured than what you'd expect
in Hypertext - hierarchical instead of "random."  The problem with this
solution is that we're looking for a documentation system that would be
useful to us as well as to our customers.  CASE's main use seems to be
for developing a system.  Can anybody recommend a system that allows an
exportable read-only subset that can be delivered as an on-line
documentation system?  Does such a beast exist?  Does it make sense?

I'm sorry this has gotten so long.  Please respond via email as I don't
read all these groups on a regular basis.  I'll post a summary of
responses to the groups listed after a week or so.
--

Sam Saal        saal@floyd.att.com


Brought to you by Super Global Mega Corp .com