A markup language that is extensible (unlike HTML).
Extensible because it is not really a markup language.
-
A markup language has
- a vocabulary, and
- a grammar
By markup language
I mean a vocabulary (i.e., the
set of elements that have meaning), and a grammar (i.e. how
they relate to one another).
DocBook is a markup language:
OFX (open financial exchange) is a markup language:
Not a language, but rather a meta-language
XML has no tags of its own, but instead defines the syntax of tags; it defines no vocabulary or grammar of its own, but does tell you how to define a vocabulary and grammar
-
methods of defining markup language
- rules for declaring what the vocabulary of a markup language is, and for writing the rules of their grammar
- e.g., rules for saying that
title is a thing in my markup language, andparagraphs are not allowed in them.
-
syntax for expressing markup language
- rules for how to incorporate a markup language into a document, a stream of text.
- i.e., rules for differentiating markup from content
- e.g.,
when you see
title that means atitle thing has begun
- Many different purposes (financial data, linguistics, literary texts, technical manuals…)
- Many different kinds of markup (structure, content, interpretation, appearance…)
- Many different user communities (IRS, airplane manufacturers, literary scholars, librarians…)
XML is
Have to admit that when I say When people say "XML is hard", they usually do not
mean "XML 1.0 is hard" but "XML 1.0 + namespaces in XML +
XPath + DOM + XSLT + W3C XML Schema + XML Base + xml:id +
XInclude + XPointer + ... is hard" and the proportion of
criticism that goes to XML 1.0 itself is usually pretty
low.
XML is a metalanguage
Text is divided into
- elements by
start-tags andend-tags <heading></heading> - start-tags by
< … ><> - end-tags by
</ … ></> - special case: short-hand for an element with no content
</>]]>
of great charm, which was grander than anybody
else's, or seemed so to Bear, because it had both a
knocker
Elements have attributes (sort of as nouns have adjectives).
- feature-value pairs
- syntax:
name="value"orname='value' - always specified in the start-tag (or the empty-element tag)
Warp Speed, Ms Bright!There was a young lady named Bright, Who travelled much faster than light, She departed one day, In a relative way way, And returned on the previous night.
To refer to a character that is not on your keyboard,
delimit its ISO 10646 (or Unicode) code-point with
&# and ; for
decimal values, or&#x and ; for
hexadecimal values.’tis nobler in the mind to suffer</l>
A way to handle hard-to-type characters (e.g., 西
and €
are not on an American keyboard) and
boilerplate text
- Delimited by
∧ - five are pre-declared:
amp,lt,gt,apos, andquot - some have standard names: e.g. é = é
- some are locally defined: e.g. ©right; = "this text is copyrighted by the Women Writers Project…"
'tis nobler in the mind to suffer</l>
Comments are sections of text that are ignored by the XML processor.
<!-- Famous lost book -->
Comments start with <!-- and are terminated by -->.
Any character sequence, including markup, is allowed
inside a comment, “--”
Simple set of rules on document syntax:
The following example has
<l>
<s>
</l>
<l>
</s>
</l>
<l>
<s>
</l>
<l>
</s>
</l>]]>
The following example uses the TEI
<l>
<s part="I">
</s>
</l>
<l>
<s part="F">
</s>
</l>
<l>
<s part="I">
</s>
</l>
<l>
<s part="F">
</s>
</l>]]>
Of course software that comes along and counts how many sentences you have in your poem> elements now will need to be smart enough to not count this as 4 sentences.
There is an example of how to do this in my stylesheet to count metrical lines that is on the TEI wiki.
Software will also need to know the sentence isn't
... down the walkHe did not ...
.
A valid XML document follows the rules of a schema that
describes a particular markup language:
- In order to be valid, an instance must be well-formed;
- A well-formed document need not be valid.
- A way to use tag vocabularies from different markup languages
- Allows for specialization of markup languages (by discipline, by function)
- Good for metadata: can use TEI header in a METS record
- Good for specialized markup: e.g. MathML, MusicML
- No need for every markup language to handle everything
Instance syntax:
XML thinks of texts as trees (with links).
-
some advantages
- represents hierarchical text structures very naturally
- many texts actually have a native tree-like hierarchical structure
- use of tree structures makes processing easy
- yields some natural advantages, e.g. context-sensitive searching, or variable granularity of text retrieval
-
some disadvantages
- does not represent non-hierarchical structures naturally
- many textual features do not naturally fit into a tree-like hierarchical structure
- use of tree structures is sslloooww
- yields some natural disadvantages, e.g. verbosity
Because elements are always nested inside each other, XML
can be thought of as representing a
-
some advantages
- represents hierarchical text structures very naturally
- many texts actually have a native tree-like hierarchical structure (think sentences inside paragraphs inside sections inside chapters)
- use of tree structures makes processing easy (computer scientists know how to deal with trees)
- yields some natural advantages, e.g. context-sensitive searching, or variable granularity of text retrieval
-
some disadvantages
- does not represent non-hierarchical structures naturally
- many textual features do not naturally fit into a tree-like hierarchical structure (think of a verse line split between two characters in a play)
- use of tree structures is sslloooww
- yields some natural disadvantages, e.g. verbosity