[
Lists Home |
Date Index |
Thread Index
]
Andrew Watt wrote:
>
> When some people indicate that XML is semantic markup what type of semantics
> are being referred to? Is it semantics accessible to a human reader? Machine
> processable? Both?
>
> So ... just what is semantic markup?
I really dislike the term "semantic markup". In fact,
I'd go so far as to say that the very notion is completely
against the philosophy of SGML (see [1], section 4, question 6
for the operative definition of this term).
Raw TeX markup has a well-defined and totally unambiguous semantics;
however, raw TeX isn't the sort of thing most XML Deviants have
in mind when they talk about semantic markup.
I prefer the terms "generic markup" and "generalized markup".
Generic markup is simply a shift in focus, from "what to do"
to "what this is". But even though the vocabulary of generic
markup systems (like LaTeX) may consist of words like "section
heading" instead of "14 point bold Times Roman", they still
have a predefined meaning -- usually defined in terms of
a lower-level procedural markup system (like raw TeX).
In the leap from generic to generalized markup, semantics
are stripped away entirely. The meaning of an SGML document
is the document itself: a sequence of characters, hierarchically
organized into elements, labelled with generic identifiers
and annotated with attributes. Documents using generalized
markup have no inherent semantics; all meaning is emergent,
imposed by external processes.
> Looking at the question from another angle which, if any, of the following
> statements is true? XML is self-describing. Generic XML markup has no
> semantics.
I agree completely with the second statement. As for the first,
XML *could* be considered self-describing in the sense that
what an XML document describes is the XML document itself,
but that's probably stretching things :-)
[1] <URL: http://www.flightlab.com/~joe/sgml/faq-not.txt >
--Joe English
jenglish@flightlab.com
|
