Requirements for Gutenberg DTD's

by "Frank Boumphrey" <bckman(at)ix.netcom.com>

 Date:  Sat, 11 Mar 2000 16:37:22 -0500
 To:  <hwg-gutenberg-dtds(at)hwg.org>
  todo: View Thread, Original
We are beginning to collect requirements for the Gutenberg DTD's. Here are
some of the suggested requirements.

If you have requirements that you think need to be addressed or discussed
(especially  if you disagree with any of the following), please get them in
within the next 2-3 weeks.  We would like to try and publish a formal
requirements document by the middle of April.

Frank

The DTD's shall have common top level model.
===========================================
  Rational:
  Ease of use for the elements containing the documents
  meta data such as the markup blurb, the Gutenberg blurb etc.

The DTD's shall be simple to use.
================================
  Rational:
  Gutenberg DTD's will be used by used by Volunteers to
  mark up Etexts and other existing electronic documents.
  For the most part these volunteers will have had no
  formal training in XML, SGML, or other document authoring
  processes. It is therfore necessary that DTD's should
  be useable with only minimal formal training.

Markup shall be intuitive.
==========================
  Rational:
  This goes hand in glove with requirement 1. The element
  names should reflect real world structures and content.
  Thus <para> is preferable to <p> and <date> is preferable
  to <dt>.

  Also it is possible that at a future date some author may
  come across a marked up document without the benefit of
  documentation. The markup will only be useful to this user
  if the meaning is intuitive

DTD's shall reflect Structure and Content.
=========================================
  Rational:
  The majority of markup will by necessity reflect document
  structure. However as an aid to researchers the DTD should
  be capable of expressing rich content structure. See also
  the extensibility requirement below.

DTD's shall be capable of reflecting style
where this is important.
============================================
  Rational:
  The DTD's must not be used to relect style. However in certain
  documents the original formatting had some semantic meaning.
  Alice Shrinking in "Alice in wonderland" springs to mind.
  Less fanciful examples reflects the original authors use
  of Italics etc. for emphasis.

  Also some Etexts contain ASCII art (See the Etext of "Two
  years before the mast" by Dana.) It we would be useful if
  some means were discovered of maintaing this art.

Markup should be capable of distinguishing between
original content and added content.
==================================================
  Rational:
  Many older books have aquired footnotes and annotations
  added by subsequent editors. It is important to be able
  to distinguish this content from the original content,
  i.e. an editors footnote from an original authors foot
  note. (Ed Note: This can possibly handled by a class
  called 'editorial)

Clear distinction between Meta content and original content.
============================================================
  Rational:
  This requirement is self evident.

DTD's shall be transformable
============================
  Rational:
  It is likely that a future researcher will want to use his
  own Schema for analysis (e.g. TEI). Gutenberg DTD's should
  be capable of conversion to other formats(Ed Note: Actually
  this is a feature of any good DTD design)

They shall be non-restrictive
=============================
  Rational:
  This is part and parcel of the use that these DTD's will be
  put to. The idea is not to enforce an author to adopt a
  'house style', but rather to markup in a meaningful manner
  an existing document. Any restrictions imposed should only
  be used to discourage 'tag abuse'.

They should be modular
======================
  Rational:
  Although this is not an actualy necessity for the DTD use,
  an adoption of a modular pattern for the DTD will encourage
  it's reuse and will ease future mainainance.

They should be extensible
=========================
  Rational:
  To be of use to researchers the DTD's should be capable of
  easy enrichment and increased content granularity. (Ed Note:
  We may want to provide a couple of 'out of the box'
  extensions.)

They should be subsetable
=========================
  Rational:
  Many documents can be marked up with a very small subset of
  a DTD (e.g. a title, a chapter, and a paragraph container).
  If this is all that is required the DTD should not impose
  burdemsome additional requirements on the marker

They should be stable
=====================
  Rational:
  The DTD's should follow the DocBook policy of giving at
  least a versions notice of the introduction of a change
  that will reflect on backwards compatibility.

I am adding this further requirement to my list 10:16 PM 3/10/00
following the e-mail thread on XML-Dev.

They should, where possible, be transformable
=============================================
  Rational:
  If possible the DTD's should be capable of being transformed
  into other DTD's such as TEI.

HWG: hwg-gutenberg-dtds mailing list archives, maintained by Webmasters @ IWA