A construct in the strict language, which consists in a production
tree leading to a terminal production consisting in a sequence of
symbols such as 'bold begin symbol' 'skip symbol' 'bold end
symbol', must be represented somehow so it can be read by either a
human interpreter or some mechanical interpreter such as a computer
program. A representation language assigns some particular
representation to each symbol.
It is thus possible to represent programs in the Algol 68 strict
language in different ways, tailored to different purposes. A
publication language will likely use rich text, fonts and/or
graphical features in order to represent symbols such as 'bold
begin symbol', 'bold letter a symbol' or 'brief case
symbol'. A programming language to be used by programmers and
text processing programs would typically use some stropping regime,
resulting for example in BEGIN SKIP END. Finally, a
hardware language could use a compact binary representation to
ease the storage, transmission and automatic processing of the
programs.
The Revised Report suggests and uses a particular representation language, which is the reference language. Implementations are encouraged to use representations that are reasonably closed to the reference language whenever possible.