UNIVERSITY OF HERTFORDSHIRE
		 COMPUTER SCIENCE RESEARCH COLLOQUIUM

		   "Generic Programming in the Context
		      of Array Processing Languages"

			     Stephan Herhut
	    (Centre for Computer Science and Informatics Research,
		     University of Hertfordshire)

			10 January 2007 (Wednesday)

			  Lecture Theatre E350
		       Hatfield, College Lane Campus
			       3 - 4 pm

	       Coffee/tea and biscuits will be available.
			  [Catering Permitting]

		    Everyone is Welcome to Attend
			  [Space Permitting]



Abstract:

Given the steadily rising complexity of today's software applications,
software or component reuse has become more and more important over the
last few years. Different means of boosting software reuse have been
proposed and some have already made their way into mainstream programming
languages. One of the more recent examples is the introduction of Generic
Programming into Java 5.

The term Generic Programming subsumes different concepts that are all
aimed at abstracting over certain aspects of data structures and thus
allowing for more versatile component reuse. While the generics in
Java focus on abstraction over the element types of container classes,
Datatype Generic Programming, for example, allows us to abstract on the
type of data in a more general setting; it supports the specification
of algorithms on the structure of data, thereby enabling us to specify
boiler-plate code like pretty-printers or parsers in a data type
independent way. Mainly functional languages like Haskell or Clean have
been extended to support this kind of generic programming.

In SAC (Single Assignment C), a functional array processing language,
a different concept of generic programming has been implemented. By
abstracting over the dimensionality and shape of arrays, SAC allows us
to write shape-invariant, i.e. dimension and shape generic, programs on
n-dimensional arrays; it utilizes subtyping based function-overloading
as a glue for generic and non-generic code.  Although this approach has
proven beneficial, it is primarily focused on abstractions over shape
and does not cover other structural attributes of arrays like nesting
structure or symmetry.

This talk presents the first results we obtained in trying to extent the
generic programming capabilities of SAC by amalgamating various approaches
towards generic programming into a unified framework. After a brief
introduction into the existing facilities for generic array programming
in SAC, I will focus on an integration of Data Type Generic Programming
into SAC that has been developed in the course of my PhD studies.

--------------------------------------------------
Hertfordshire Computer Science Research Colloquium
http://homepages.feis.herts.ac.uk/~nehaniv/colloq