Introducing Myself

Before joining the fray and offering myself up for sacrifice to this august
group, let me say a few words about myself and what I'm involved in by way
of introduction.  I had hoped to join in several weeks earlier than this,
but I've been sidetracked by several events out of my control that have
taken my time and attention.

I'm president of a small company called Neville SGML Associates, Inc. 
We're based in northern Virginia.  My partner, Carlton Neville, and I
attended the XML workshop at WWW6 (Monday, April 7) and met many of you
there.  I've been working with computers and documents for more than thirty
years in various capacities, mostly in what we might call system analysis
and engineering, although the new term process reengineering seems to
describe my activities more clearly.  Carlton introduced me to SGML in 1990
and got me hooked.  We are currently designing a repository system for
portions of the federal government based on SGML, using HTML to transmit
documents between sites.  We first heard about XML at the HyTime conference
in Seattle last August and were much cheered to hear what we heard about
it.  From what we've learned since, we're hopeful that XML will better meet
our needs (and more efficiently) than either SMGL or HTML.  What we see so
far says that it will, provided a few key ideas are kept or incorporated,
such as obligatory end-tagging and the optional use of DTDs (for our
purposes we suspect from our analysis so far that we will have to require
their use, but our requirements aren't necessarily everyone's).

Long-term requirements that have been articulated to us on this contract
make it clear that neither SGML nor HTML individually can answer the mail
(no pun intended).  The need to publish documents in multi-media (and we
include printing on paper as a media, believe it or not) points to SGML for
some tasks and, where web-like intranet implementations are necessary,
HTML, but not precluding SGML.  Since like most large organizations,
plagiarism of organizational documents by other members of the organization
for different purposes is a key part of doing business (like not
reinventing the wheel), an SGML environment, or one like it is also
suggested.  Basically, in this system, it will be necessary to store
documents in SGML, converting them as necessary to HTML, a doable but still
cumbersome process.  If XML were here, XML would be the underlying
implementation.  Because one requirement is that we use COTS products and
not build in-house systems and software, we are limited by what is
available on the market--or that is clearly and demonstrably on the way. 
We can't deliver on a kiss and prayer.

I won't go into the limitations of HTML or SGML here except to say that
HTML lacks the flexibility we need to support the motley environment we
must respond to, although it will be incorporated as an important feature. 
Our taskers also balk at both the complexity of SGML and the derth of
trained experts.  We need HyTime linking features, but until we saw the
draft linking spec for XML Jon distributed at the WWW6 conference, we
weren't sanguine about seeing COTS implementations of useful linking beyond
that provided by SGML any time soon.  XML therefore holds a great deal of
promise for us and we applaud its development and wish to help in its
formulation to the extent we can.  Thus, while most of the WG traffic I've
been reading the past few days talks about XML and the Web, our company
sees much broader application, such as in efficient, long-term storage of
large volumes of stable documents, a functionality that should be looked
upon with favor by any large commercial enterprise.

In any event, I hope this gives you an idea where I'm coming from.  I have
some specific comments on the WD-xml-link spec Jon distributed at the WWW6
conference that I'll be posting as soon as I can.

Eric

Received on Saturday, 17 May 1997 10:07:06 UTC