Lists Home |
Date Index |
True and true. But this outlines an important
point: when it came time to use SGMLS in production,
the ESIS stuff was invisible to the authors. They
saw the error messages. Understanding ESIS did
not help them debug the SGML. Grosso explained
the ESIS to me and that was for me, also, a nice
aha moment, but it also seemed trivial, as if,
ok, that's what I expected it to do, so? moment. A parser
feeding an application through a clean data model
delights a programmer; a validator that clearly
indicates what is wrong in the markup at what
line in the text delights an author. Who do you love?
XML has almost single handedly given markup
the reputation of being "computer friendly
and author unfriendly". Why? SGML was actually
friendlier to authors and hard on the programmers.
We know why. It isn't always a good thing
to insist on the programmer as the primary
beneficiary. HTML succeeded and continues to
thrive because it is friendly to the author.
XHTML is stricter, cleaner, and wallowing. Why?
There is something to the saying that the
middle class never give up their vices.
From: Sean McGrath [mailto:email@example.com]
>I liked his reference to sgmls. He is right
>about the redoing of ARCSGML being a seminal
>event. OTOH, ARCSGML had better error messages.
>There are some things lawyers do better than
Ah, but the key thing about sgmls was that it emitted a simple, line-oriented
implementation of ESIS - an infoset.
I would argue that sgmls's crowning achievement was to add the infoset
to SGMLs core from a programmers perspective and then
expose it in a very programmer friendly way.
This was the biggest "aha!" moment of my markup career when it dawned on
me that the parser allowed me to think purely in terms of a hierarchical
data-model view of the world and ignore syntax.
Although technically speaking, ESIS is abstract and the line-oriented notation
produced by sgmls (and nsgmls, (and PYX)) are purely James's invention,
dollars worth of production code is based on it and the term ESIS has become
essentially synonymous with James's notation for it.
The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
initiative of OASIS <http://www.oasis-open.org>
The list archives are at http://lists.xml.org/archives/xml-dev/
To subscribe or unsubscribe from this list use the subscription