{html
  {head
    {title S-Expression Inspired Markup Syntax}
    {style
      H1, H2, H3, H4, H5, H6 {
        font-family: futura, sans-serif;
      }}
  {body

    {h1 Introduction}

    {p
      We wanted to have a simple and easy to remember markup syntax
      for SGML-inspired markup like HTML. It should parse to a
      hierarchy of {i nodes} which these days often is called a
      "DOM".}

    {-- we could make a double newline be like <P>, but that would
     -- need the DTD, but would be quite nice to have.
     --}

    Escaping rules should be dead easy to remember: {:tt \\} is the
    sole escape character and strips the next character of any special
    meaning making it behave like any letter.

    Like with SGML or XML, nodes have a {I general identifier} (GI), a
    list of attributes mapping a name to a string and a sequence of
    either text (PCDATA) or other nodes.

    With our syntax both nodes and attributes share a common syntax:

    {pre
      \{{I gi} {I children}...\}
      }

    When the identifier begins with {:tt :} it names an attribute,
    otherwise it names a node. Text doesn't need any quotes and once
    the {I general identifier} has been consumed, the rest until a
    matching {tt \}} is the text of that node. Non-attribute nodes may
    contain other nodes mixed with text.

    {-- something more complex --}
    
    {table {:border 0}
           {:cellpadding 5}
           {:rules groups}
      {thead
        {tr {th SXML} {th SGML/HTML/XML}}}
      {tbody
        {tr
          {td {tt \{b foo\}}}
          {td <B>foo</B>}}
        {tr
          {td {tt \{a \{:href http://example.com/\} click me\}}}
          {td <a href="http://example.com/">click me</a>}}}}
    }}}