Why didn't he consider the needs of other sciences?
Installed: 27 Aug 2011. Updated: 31 Aug 2011; 28 Jan 2012
Summary
Most people discussing the importance of teaching computing in schools think only of
three (related) reasons:
1. not enough good school leavers choose to study computer science at university
2. industry is desperately short of good programmers/computer systems engineers.
3. computing products are very important for the economy and for our quality of life.
Some people add a fourth (also related) reason:
4. designing and testing working computer programs can be challenging and great fun.
What they nearly all miss is the need to teach scientific (not just numerical!)
computational thinking (including (a) designing, building, testing, debugging,
analysing and comparing working computer programs and (b) using such systems to model
and explain natural information processing systems, e.g. minds, brains, ecosystems,
and social systems.) This need is desperate for many disciplines, and in the long run
just as important as producing new computing engineers. But the education required is
subtly different.
So choices about programming languages, syllabus structures, tasks, projects and modes of
assessment should not be restricted to what serves aims (1) to (3) (or 4) above.
NOTE ADDED 12 Jan 2012
I recently discovered that Paul Nurse, a biologist who is the current President of the
Royal Society, has been presenting ideas closely related to this note: e.g. claiming
that major future progress in the biological sciences will depend on understanding uses
and management of information in biological systems, from cells upwards, e.g. in his
2010 Royal Society lecture on "Great ideas in Biology"
http://royalsociety.org/royalsociety.tv/
Revised and Updated: 12 Jan 2012 After Michael Gove's speech on Digital Literacy:
http://www.guardian.co.uk/education/2012/jan/11/digital-literacy-michael-gove-speech
This page is: http://tinyurl.com/EricOmission
Or:
http://www.cs.bham.ac.uk/research/projects/cogaff/misc/eric-schmidt-lecture.html
A partial index of discussion notes is in
http://www.cs.bham.ac.uk/research/projects/cogaff/misc/AREADME.html
NOTE:
UKCRC is the UK Computing Research Committee http://www.ukcrc.org.uk/
One of its achievements, providing part of the background to my comments, was setting up the
Computing Grand Challenges Initiative (led by Prof Tony Hoare, and Prof Robin Milner, in 2002):
htp://www.ukcrc.org.uk/grand-challenge/
The inspiration behind that initiative has been lost sight of in recent discussions of requirements
for computing education in schools. See for example:
Grand Challenge 1: In Vivo - in Silico: The Virtual Worm, Weed and Bug
Breathing Life into the Biological DataMountain
A GRAND CHALLENGE FOR COMPUTATIONAL SYSTEMS BIOLOGY
Grand Challenge 5: Architecture of Brain and Mind
Grand Challenge 7: Journeys in Non-Classical Computation
Those are not primarily concerned with building useful systems.
Their aims are deeper: more like the aims of theoretical physics, or theoretical biology.
Revised version of Message Posted to UKCRC 27 Aug 2011
Date of original message: Sat, 27 Aug 2011 17:21:22 +0100
From: Aaron Sloman
Subject: Re: MacTaggart lecture by Eric Schmidt in Edinburgh
To: UKCRC (UK Computing Research Council) Mailing list
A member of the UKCRC mailing list recommended this link:
http://www.guardian.co.uk/media/interactive/2011/aug/26/eric-schmidt-mactaggart-lecture-full-text
Yes, it's a great read.
Eric Schmidt came quite close (on page 8) to making the point that computing is not just
about useful technology but also includes science. But he seemed not to notice how
important the educational need for computational thinking is in a range of scientific and
other research disciplines.
It's a great pity that most people talking about computing education in schools focus only
on the technological, economic, and social importance of applications and not on the
deep and growing significance of computational concepts, techniques and methods in the
many areas of science and knowledge (pure and applied) concerned with naturally occurring
information processing systems, varying in scale from cells and microbes to ecosystems or
cultural and socio-economic systems. This point was emphasised recently by the president of
the Royal Society, as noted above. But he did not draw out the educational implications.
There are also deep computational ideas, not encountered by people who merely learn to use
computers, that are directly relevant to some of the oldest problems in philosophy, for example
problems about the nature of causation, about the relationships between mind and matter, and
about how mathematical discoveries differ from discoveries in the empirical sciences.
I have a growing, somewhat disorganised, collection of papers and presentations illustrating
these points here:
http://www.cs.bham.ac.uk/research/projects/cogaff/talks/
http://www.cs.bham.ac.uk/research/projects/cogaff/
Alas, most of the people who do research, teaching and learning in areas (e.g. biology,
psychology, animal cognition, education, child development, social sciences, philosophy,
and others) that require the ability to think deeply about complex information processing
systems have used computers (e.g. for preparing documents, accessing the web, using email,
using databases) but have had little or no experience of building working systems of
any kind.
I.e. they have not learnt how to specify requirements and goals, produce designs, then
build, test, debug, criticise, extend, analyse, explain and compare designs.
So they lack the intellectual resources required for formulating good explanatory
hypotheses in their fields.
In that sense they cannot be competent at their jobs, and are restricted to scratching the
surface (e.g. often merely doing experiments, collecting data and running statistical
packages to find out what correlates with what, but never understanding how anything
actually works, or can go wrong, or might be improved). Building powerful explanatory
theories requires more than the ability to use natural language and draw flow-charts and
other diagrams. It also requires more than the mathematics developed for use by
physicists, e.g. differential and integral calculus. In particular it requires the ability
to model processes in which complex structures (e.g. molecules or thoughts) are assembled,
modified and interact with other structures.
But the incompetence is invisible because it is nearly universal.
The problem is also invisible to those who do have the required competences (e.g. people
on computing mailing lists, teachers of computing in schools, and the very bright computer
scientists and engineers in industry) because their interest (and most of their knowledge)
is focused elsewhere: on how to make computers do something newer, better, faster, more
reliably, more securely, more lucratively, or whatever. Most of them are not at all
interested in studying or explaining the development and operation of natural information
processing systems (e.g. animal minds).
Unless there is a major change in the priorities of computing education in schools, so as
to include education of students who go into disciplines other than computer science and
computer engineering, the nation will go on producing inadequately educated students,
teachers and researchers studying such information processing systems, e.g. in
developmental, clinical, social psychology, in neuroscience and psychiatry, in various
fields of biology, in linguistics, in philosophy, in social science, in education[*], and
in many non-computing areas of science and engineering -- whose understanding of ways of
thinking about information processing systems falls far short of what they need to do
their jobs well.
[*](Human learning/development is one of the most important and complex forms of
information processing in our lives, but we hand over responsibility for guiding
learning to people who are not at all equipped to understand information processing
systems of any kind -- though fortunately there are a few who have important
educational talents despite their ignorance, thanks in part to biological and
cultural evolution.)
If we continue to focus on how to improve computer science courses in schools,
focusing on CS courses that are mainly concerned with how to build useful applications,
courses that are not aimed at or taken by the majority of very bright learners whose main
interests lie outside computer science, we risk missing a major opportunity to educate our
nation for the future. Not because everyone needs to learn to build and/or use tools, but
because everyone needs to learn some of the new ways to think about both natural and
artificial information-processing systems. These ways of thinking have been developed since
Alan Turing did his seminal work in 1936, and many of them arise out of his work.
What we should be doing is injecting hands-on experience -- of specifying, designing,
implementing, testing, analysing, evaluating, comparing, and extending information
processing systems of various types, and various levels of complexity -- into many
non-computing educational trajectories, including mathematics, science, philosophy,
language, and many others. The focus should be on doing something interesting and directly
relevant to the subject, rather than insisting on all students trying to build something
that is practically useful, evaluated in terms of its actual or potential
usefulness, or its entertainment value. Having fun programming can help learning, but it's
not enough for deep learning.
If we don't redirect the teaching of computing in this way (not excluding building
useful applications, but offering alternatives for those with other interests) then
we'll miss the opportunity to fire up a lot more of the very bright students -- some of
whom might, as a result, consider computer science degrees and careers, while others go on
to become ground-breaking researchers and teachers in other areas of human endeavour,
including other sciences, humanities, the arts, medicine, and education. After all, a
learner is a self-extending information processing system, and if we don't understand the
implications of that we cannot design good learning and teaching systems -- except by
luck.
If learning to think about such matters becomes as wide spread as learning to understand
and manipulate numbers, then maybe even some future politicians will develop a better
understanding of the problems they are trying to solve and how to test and evaluate
proposed solutions, instead of being hoodwinked by captains of industry who promise the
moon, or wishful-thinkers who are sure they know what governments should do because the
results will be so valuable (for them).
Unfortunately, most computer scientists discussing computing at school focus only on
how to get more pupils interested in and prepared for degrees in computing and
careers in computing -- a disastrously narrow focus.
Also most of the computing researchers publish mainly in places where they will be read by
their same-discipline peers and gain prestige for the national research evaluation
exercise, instead of using open access multidisciplinary journals where there's more
chance of cross fertilisation of other disciplines.
Of course, all my generalisations have some exceptions, but the general pattern
(including the misguided exclusive focus in recent discussions on improving computer
science teaching) is disastrous both for computer science and for many other areas of
learning, pure and applied.
I don't know if there are any countries that understand this, but if there are and
they apply their understanding to reorganising school curricula in a wide range of
disciplines, the resulting difference could be very striking after a few decades.
[Yes: I know the existing teachers in other disciplines will argue that they are
already short of time to teach enough of what they think is important (though
in some areas, e.g. mathematics, the importance of computing is already recognised by
exceptional teachers). Even a few philosophy teachers are beginning to understand the
revolution.
Yes: there is a deep chicken and egg problem if nearly all school computing teachers
have not had the education required even to understand the problem let alone
contribute the required teaching. There are ways of solving these problems -- but not
at a stroke. Some draft suggestions are here:
http://www.cs.bham.ac.uk/research/projects/cogaff/09.html#908
Teaching AI and Philosophy at School?
http://www.cs.bham.ac.uk/research/projects/cogaff/misc/alevel-ai.html
A possible Artificial Intelligence/Cognitive Science GCE/A-level Syllabus
Some examples of introductions to 'Thinky' programming can be found here:
http://tinyurl.com/thinky-ex
With a small collection of illustrative video tutorials (to be extended) here:
http://tinyurl.com/PopVidTut
Some example presentations on research overlaps between Computing/AI/Robotics and
both philosophy and the study of biological intelligence and its evolution and
development can be found here:
http://www.cs.bham.ac.uk/research/projects/cogaff/talks/
See also
Darwin Among The Machines: The Evolution Of Global Intelligence,
George B. Dyson, Addison-Wesley, 1997,
Marvin Minsky on the One Laptop Per Child (OLPC) project
http://www.cs.bham.ac.uk/research/projects/cogaff/misc/minsky-olpc.html
]
Maintained by
Aaron Sloman
School of Computer Science
The University of Birmingham