Having been told repeatedly that I was wrong and should shut up,
it provides some gloomy pleasure to look at the SGML markup (not
the content) in Professor Sampson's post. Here is an experienced
editor, who knows much about tagging text but is unfamiliar with
SGML, using SGML markup to tag quotations and emphasized text only,
in an otherwise unaltered document.
Internal evidence suggests that he typed the file on a text editor
without SGML markup and then he (or Lou) used a one-at-a-time search
and replace procedure to put it in. Whoever it was, he
made two mistakes in only 261 lines. Some automatic
conversion procedures, attempting to reverse the markup, would be
completely foxed. How much erroneous markup can we expect from a
less experienced user preparing a much longer text and needing many
more markup codes? By the time Erik Naggum's software has trickled
down to them, there are going to be a great many corruptly marked
texts out there. It may be years before anyone finds out, by which
time manipulators of the texts will have to apply elaborate criticism
to work out what the markup should have been, even if they have software
tools to flag apparent errors.
To repeat, the only way to stop this is to provide easy-to-use
software for automatic conversion into, and out of, SGML markup
as soon as possible.
Although I hesitate to enter the lists at this point, I do not
want Christoper Currie to feel put down. He has put his finger on
a very practical point. The successful implementation (i.e. broad
usage among humanities scholars) will in fact be dependent upon the
availability of software for input, for export, and for import.
But as others have pointed out, we must have an agreed upon standard
if the software is to come. My own experience is in the editing of
historical documents which are tagged for typesetting. As my electronic
files now exist, they would be of little value to other scholars for
either research or instructional purposes. (Some tags are intelligible,
others are not.)
In the States, there are at least another 50 historical editions which
have similar files. And there are new editorial projects waiting in the
wings (mostly for lack of funding). What I personally envision is the
eventual creation of electronic editions from which scholars can draw for
both research and teaching. But the only rational way for that to happen
is with the use of an agreed upon standard.
The recent TEI workshop at Chicago pointed up a number areas in which further
work will have to be done--not the least of these will be the development of
simple, concise user manuals. But before we get to that point, we will have
to address the more technical issues involved in making sure that the basic
framework encompassed by the TEI guidelines will at least meet the minimal
needs of scholars.
As someone not well versed in SGML, trying to use the guidelines as they
stand is painfully time-consuming. But live markup is the only way I
know of evaluating them. And that's another topic altogether.
David Chesnutt, Editor
The Papers of Henry Laurens
University of South Carolina