Markup: Why Bother?

Markup: Why Bother? Syd Bauman 2003-10-25/26

What is Markup?

Markup is a method of making explicit one's understanding of a text.

In the Beginning …

Scriptio Continua NOTEXACTLYTHEMOSTREADABLESTUFFINTHEWO RLDBUTITDIDSUFFICEFORSOMETIMEDESPITE…

Punctuational Markup

Added whitespace and punctuation NOTEXACTLYTHEMOSTREADABLESTUFFINTHEWO RLDBUTITDIDSUFFICEFORSOMETIMEDESPITE…

Not exactly the most readable stuff in the world, but it did suffice for some time despite…

Presentational Markup

Categories of Markup

Once could think of markup as having two qualities, mood (imperative vs. indicative) and domain (logical or structural vs. renditional). Indicative Imperative Logical <head> Renditional <hi> .ce .bd

Markup has Meaning

WORDSWORTHLESSORBITSUNBOUND

Words — Worthless; Orbits — Unbound

Wordsworth Lessor Bit Sun-bound

Words worth less, or bits unbound

What We Mean

So far only rudimentary, vulgar markup.

TEI indeed any decent XML application, uses indicative logical markuppreviously called descriptive markup to indicate the structure of the text, labeling what various chunks of text are rather than what to do with them.

This allows for multiple kinds of processing on a single text. Most obviously different looking output from the same input with different layouts or stylesheets.

Descriptive (Indicative, Logical) Markup Describes the structure and function of a textual feature, not its appearance. Allows for multiple presentations and uses of the same structural data Permits independence of output mechanism (software, hardware, etc.)

In the end, it should be clear that descriptive markup is not just the best approach of the competing markup systems; it is the best imaginable approach. Coombs, Renear, and DeRose, Markup Systems and the Future of Scholarly Text Processing, Communications of the ACM 30.11 (1987-11): 933–947.

Why Use (Indicative Logical) Markup? To separate data from metadata To separate form from content To facilitate retrieval and analysis To achieve data longevity

XML

The extensible markup language (XML) is a standardized method for marking up texts with indicative logical markup, and, optionally, for explicitly declaring the markup language used.

Why XML?

XML is easy to understand; non-proprietary plain-text: human readable, software independent, hardware independent; (relatively) easy to write a parser for; represents hierarchical text structures very well, non-hierarchical not so well; ubiquitous, very well supported with both commercial and free software.

XML compared to SGML

XML is a subset of SGML, a simplification Eliminates the most complex features Removes requirement for a DTD Decreases range of options (tag omission, choice of delimiters)

Many encoding languages exist in both: [X]HTML TEI DocBook

XML Basics

XML is a metalanguage No tags or attributes of its own Instead, a set of rules for defining tags and attributes Imposes no constraints on elements and attributes in document Instead, defines how rules for such constraints are written

Metaphors XML as a "tree structure" XML as an "ordered hierarchy of content objects" Renear, DeRose, Mylonas XML as a tree-like representation of other, more complex structures

Parts of an XML Document Instance

Everything is delimited: elements by start-tags and end-tags tags by < … > and </ … > special case: an element with no content may be represented as < … /> as short-hand for < … ></ … > entity references by & … ;



      
        Everything's Delimited: elements
        Text is divided into elements (the nouns of
          the encoding — content objects).
        
          Elements are delimited by tags: a start-tag at the begining, and an
            end-tag at the end.
          Start-tags are delimited by < and >.
          End-tags are delimited by </ and >.
          Special case: for an empty element start- and end- tags are combined — delimited by
              < and />.
        
      

      
        Example Elements
        
          Sara Schmidt]]>
          The year that Buttercup was born, the most
beautiful woman in the world]]> … </para>
          <line>']]>
          We've had  hits since last
updated on  at .]]>
        
      

      
        Everything's Delimited: attributes
        Elements have attributes (sort of as nouns have adjectives). 
        
          Attributes have names: rend, type, id, n,
              target
          
          Attributes have values: slant(italic), chapter, P0001
          
          Attributes are specified in the start (or empty) tag
          The attribute name is preceded by whitespace and followed by =You can have whitespace before the =
          
          The attribute value is delimited by either single or double quotation marks ("s or
              's). No curly quotes!
        
      


      
        Example Elements
        
          Philadelphia]]>
          UIUC]]>
          'Twas the night before Christmas]]>
          ]]>
        
      

      
        Everything's Delimited: character references
        To refer to a character that is not on your keyboard, delimit its ISO 10646 (or Unicode)
          codepoint with 
            
              &# and ; for decimal values, or
            
              &#x and ; for hexidecimal values.
          
        
      

      
        Everything's Delimited: entity references
        Names delimited by & and ; are references to the string
          of characters or bits declared as associated with that name.
        How to declare such an entity is not covered here.
        Five entities come pre-declared: amp, lt, gt,
            apos, and quot.
      

      
        Entity references
        A way to handle special characters (i.e. not lower ASCII) and boilerplate text
        Delimited by & and ;
        &eacute; = é
        &copyright; = "this text is copyrighted by the Women Writers
          Project…"
      

      
        Well-formedness
        Simple set of rules on document syntax.
        single root element
          every element has a start- and an end-tag (or empty)
          no element overlap
          all entities and attributes properly delimited
        
      

    
      Validity
        Validity: follows the rules of a DTD
        
          which elements can go where
          which attributes can be (or must be) on which elements
          minimal constraint on values of attributes
        
        
      


      
        Samples
        <p><q>Curiouser and curiouser</q> said
<name>Alice<name> to herself.</p>
      

      
        Samples
        <lg>
   <l>A bird came down the walk</l>
   <l>He did not know I saw,</l>
   <l>He bit an angleworm in half,</l>
   <l>And ate the fellow raw.</l>
 </lg>
      

      
        Samples
<lg>
  <l><s>A bird came down the walk</l>
  <l>He did not know I saw.</s></l>
  <l><s>He bit an angleworm in half,</l>
  <l>And ate the fellow raw.</s></l>
</lg>