Mathematics at Flinders

In the last two decades at Australian universities, departments have primarily been funded using formulae that are based to a large extent on the number of Australian and foreign students that are enrolled in their majors, or are taught by their service courses. As such, there is a tremendous pressure on administrations and departments to increase enrollments at any cost – for instance by lowering course requirements, or removing requirements involving other departments and replacing them with in-house substitutes. As one consequence on this focus on enrollments, other factors such as research and teaching quality, or the needs of industry or the community, are considered secondary or are neglected altogether. Another effect is that department funding is subject to short-term fluctuations based on changes in student demand. [In contrast, in other countries such as the US, funding is usually not determined by rigid formulae, but instead changes slowly from historical levels in response to a combination of all the factors listed above.]

These factors have played a significant role in the recent USQ crisis, and as indicated in a letter to the most recent issue of the Gazette, have also had a detrimental effect on mathematics at the University of New England. In this post I would like to describe how this structure had largely destroyed mathematics at my old alma mater, Flinders University, in which the number of full time mathematics staff has dwindled from 19 during my time there in 1989-1992 to just three today, although a rebuilding effort is now being planned there, as reported last week in the Australian.

I graduated from Flinders with a B. Sc. Hons in 1992. At the time there were 19 mathematicians and 5 statisticians in the School of Information Science and Technology, which also included computer science. My graduating honours class was small – there were only two other students in it, and it would in fact end up being the last honours class in maths Flinders would have – but the maths program was otherwise very healthy at the time for such a small university, with a constant stream of post-graduate students, and many advanced courses.

In 1992, there was some movement in the Computer Science division to decouple itself from mathematics, in particular reducing the mathematics requirements in that discipline. Because of this, the School was eventually split in 1995 into three separate schools of Mathematics, Statistics, and Computer Science – and the latter school promptly lowered their maths requirements, thus pulling a significant number of enrolments (and therefore income) from the Mathematics department. (Nowadays, the CS major at Flinders has no mathematics or statistics requirements whatsoever; for comparison, the CS major at UCLA requires six quarters of lower-division mathematics and one quarter of statistics.) [Update, 18 June: The B. Sc. (Computer Science) degree has recently been revised to include more maths and stats requirements, see comments below.]

This situation contributed to a significant funding shortfall in mathematics, leading to voluntary redundancy packages being offered to many of the faculty. By 1996 there was just eight mathematicians and one statistician at Flinders. Other departments such as Engineering and Physics also began reducing their maths requirements at about this time, exacerbating the situation. (To give one example, the Bachelor of Engineering (Computer and Electronics) degree offered at Flinders nowadays requires just three maths courses: two semesters of single-variable calculus, and one 3-unit course in probability and signal processing. A comparable degree at UCLA requires seven quarters of mathematics. The degrees at Flinders are currently being reviewed, however.) [Update, 18 June: The Engineering degrees have also been recently revised to incorporate more maths content.]

The shrinking mathematics and statistics departments were eventually recombined in 1999 with the computer scientists and engineers into the School of Informatics and Engineering (but now recently renamed the School of Computer Science, Engineering, and Mathematics). Nevertheless, the funding structure continued to give each individual department a strong incentive to lower requirements from other departments (with “hard” mathematics courses being particularly vulnerable to being cut), and mathematics and statistics continued to experience dropping enrollments and fiscal deficits. In 2006-2007 several more redundancy packages were offered, leading to the current situation with just three mathematicians and no statisticians as full time academic staff in the School. Coupled with similar losses in other departments, it now seems that many of the degrees offered by Flinders are no longer viable at their desired levels of quality, simply because there is not enough capacity to teach the required courses.

As the above news item indicates (see also this announcement), there now does seem to be some efforts at Flinders to try to repair the situation, for instance by reviewing the degrees offered and the organisation of its departments, but it seems that there are serious structural issues, particularly the budgetary incentives that encourage departments to lower their course requirements at the expense of other departments (and of the quality of the degrees being offered), that need to be addressed.

[Update, 18 June: Some corrections in view of comments.]

[Update, 3 July: Some more information about the review of the maths department can be found here.]

About these ads

67 Responses

  1. Hi Terence,

    I’ve been reading your highly entertaining blog now for a while but this is the first time I’ve contributed anything. As the recently appointed head of the School responsible for Mathematics at Flinders, I didn’t want to promise anything that Flinders wasn’t prepared to deliver.

    Much of what you say is correct. The various funding models for mathematics in Australia have given at times more and at times less for educating a maths student (normally less) but almost all of the time funding has been based on a k_1.n model where n is the number of students and k_1 is some level of funding per student. Unfortunately, the cost to teach a topic is more accurately s + k_2.(n-1) where s is the cost of teaching the first student and k_2 the cost of teaching the rest. The upshot of this is that large classes pay the bills for small classes. So, as you so rightly say, when maths is withdrawn from awards that have high numbers (such as computer science) the result is often that the axe falls on the smaller maths topics now left unsustainable.

    However, Maths is different. It’s fundamental to almost every science, engineering, medical and commerce course. Had this funding reality happened to any other area than mathematics then I guess that we wouldn’t now be looking at rebuilding it. And we are. We have an external review happening right now. As the VC said, it may well have two parts – a central core of mathematicians and statisticians (we do, by the way, still have statisticians left on staff) plus those who are embedded in areas such as econometrics, epidemiology, biostats and so on. Once the final model is worked out, I promise I’ll let you know.

    Just a comment about maths in computer science. The new Bachelor of Computer Science has three core maths topics plus the ability for students to take more. And unlike many CS courses, it requires year 12 Maths for entry.

    The CS major in the Bachelor of Science (your link was to the old graduate entry degree – now discontinued) has been revised to include statistics, as have the new Engineering awards (which now require two 4.5 unit maths in first year – four if their maths is lower than normal), a 6-unit signals and systems topic in second year and a further 6-unit signal processing topic in third year. However, there is also much maths embedded in other topics.

    For me to claim that Maths isn’t at something of a low ebb at Flinders would be unrealistic. Most of what you say is true and most of your comments I agree with. But a path to the future is being mapped out, we have strong institutional support for mathematics and we hope to have something concrete to report by the end of the year.

    I’ll keep you posted…

    Cheers

    John.

    http://flinders.edu.au/mathematics

  2. A few quick comments:

    (1) At the time Flinders cut maths out of computing, it did not need to. Student numbers in computing were quite high (dot-com boom), and there was something of a thriving computer science degree. The actual reasons, from my viewpoint were quite different and, in some ways unprofessional; but that will have to wait for another day. Briefly, there were also personality issues involved, in the form of people who were allowed to wield power but who ranged from “clueless about maths and its relevance elsewhere” to “resentful about high research productivity of the maths stuff”.

    (2) It is true that to some extent, Flinders was seduced by the promise of even larger numbers. That was readily accomplished by cutting out anything the students thought was too difficult (and maths was a prime candidate) or simply didn’t like. The remaining topics were just dumbed down.

    (3) [2] above was extremely short-sighted and a sad reflection on some of the “managers” involved. Many of the graduates of the “information technology” program are people who should never be allowed near a computer, unless it is to play video games. Fortunately, the bust of the dot-comm bubble has been a remedy: with employers actually wanting serious graduates, even Flinders is now restoring computer science. Ironically, serious universities who did not get carried away with the “information technology” thing have continued to do well or even better: in South Australia, Uni Adelaide is a good example, but many other examples exist across the country (e.g. UNSW).

    (4) In relation to your comments on maths in engineering at Flinders, I am amazed that these programs continue to get accredited, perhaps a sad reflection on the quality of engineering in Australia.

    Roddick is off to a good start, but he has his work cut out for him—years of serious damage by managerial types who were left in their positions for far too long. On the bright side, two of have had their “successes” at Flinders rewarded by other places (one a joke, the other a semi-university) and the third should be encouraged to move on if the university is serious about its research profile.

  3. Dear John: Thanks for the comments and corrections! It seems some of my information was a little out of date; it’s encouraging to hear that things are finally moving in the right direction there.

  4. Dear Terry: Encouragement is always good, but it never hurts to have a little basis for any “uplifment” one might have. So, here we go.

    Yes, Flinders appears to have some sensible directions, but, as the Big Mac ad from some years go put it, where’s the beef? Beyond the nice words of better things to come, we should ask one or two questions:

    (a) Has Flinders advertised any positions in maths? When do they intend to?

    (b) What does one make of the fact that right now Flinders has decided that an excellent mathematician should be no more than a Senior Lecturer while a much, much better uni has decided that the very same person is worthy of a chair?

    No, I don’t expect you to have answers, but I will try to shed a little light. Before I do that, a quick comment: In one of you other blogs, you expressed a distaste for “ad hominem” comments. As far as it goes, that’s all very well—and, indeed, extremely admirable—but managers, especially senior ones, are expected to show leadership and to take the heat when appropriate. That’s why they get paid so much. That being so, there is no reason why failures in leadership should not be pointed out. Now, here comes the light.

    1) I wish to make clear that my earlier remarks should in no way be taken to mean that the problems with maths at Flinders were solely due to clueless middle managers. In another of your blogs, a smart former student (Scott Vallance) notes, quite correctly, that the university’s senior management were pretty happy (my words) to let things go that way. Bill Moran (see comment on your other blog) is also spot-on with his comments, and he should know—he was at the top of the heap, but he cut out before the serious rot set in. My story is very good—I know this because I am writing it—but for another good, and probably more subtle, and maybe even better and truer version—you need to get Moran to write something.

    As in most Australian universities, all of the issues that have to do with the stuffing-up of maths at Flinders properly fall under the remit of the Deputy Vice-Chancellor (Academic). In that regard, the university has not been well-served. In the last decade, that position has had no fewer than 4 occupants. They are/were probably good for something—just what that might be is hard to imagine—but, as far as that portfolio goes, the university could not have had a more out-to-lunch lot. Inter alia, any clued-on person would, at least, have had some idea of the impact that shafting maths would have on several other disciplines.

    2). In regard to the other rhetorical question under (b) above, I should like to make a few remarks:

    There is a great deal to be said for the fact that most of the maths staff managed to avoid being bitter and extremely pissed-off at a time when they were getting shafted every which way. Several of them doing very good work (and some still are) had to f***k around with the Promotions Lottery while all sorts of clowns were getting ahead on very dodgy grounds—e.g. administering, and managing, and generally bothering people who are/were trying to do some real work.

    I am all for “community engagement”, and “administration”, and “service to the profession”, and other such-like good stuff, but at the end of the day, we all know that a university’s reputation is determined by research. I never heard of anyone trying to get into Oxford or MIT because of the level of “community engagement” or whatever in those places. You know what I mean, Terry. This being your blog, I will say that you are undoubtedly a fine and inspiring teacher, engaged at all times with the community, forever making strong contributions to the community, etc. etc. etc. Plus, you are a mensch. But, still! We all know that even if you couldn’t mumble your way out of a wet paper-bag, your Fields Medal guarantees an indefinite supply of excellent students and, I imagine, a tidy pile of $s from UCLA.

    One of the problems with the rudderless ship that Flinders has been in the last decade is a forgetting of what actually determines the reputation of a university. There are quite a few “world university rankings” popping up these days, and one of the most bizarre things about Flinders is the number of academics—where have they been?—who express astonishment that Flinders is not among the top 100 in the world. I suspect that they have come to believe some of the propaganda being peddled locally and by the “Innovative Reseach Universities” (IRU) group. (What exactly does IRU do, other than gangin-up because everyone else is ganging-up?) I rather doubt that IRU does more research or innovation than other similar universities—-in your description of Flinders, “small and little-known” but with some pockects of past excellence [that's mine outside the quotation marks]—but I get a good chuckle at the “world-class universities” claim. As I write this, I am visiting some of the best universities in Asia. By whatever ranking system you choose, all of them are way, way, above the IRUs, but, oddly enough, they merely claim that they are trying to be “world-class”; none of them claims to be world-class. For another day: as soon as someone says they are world-class, you can be sure thet are not, and they know they are not, and we know that they know …

    To be continued … But I will close with this: Things are looking up. A bit. The new Vice-Chancellor appears to be on the ball—actually appears to know what it will take to get Flinders forward, i,e. back to where it used to be. Check out his “State of Flinders” report. And subtle too—I could learn stuff from this guy—: “Wonderful university, glad to be here, gratitude to all you folks doing one heck of a job …. but still!, relative decline, and so forth …” Like Roddick, he has his work cut out for him.

  5. I missed Roddick’s bit about the current “review”. The way Flinders works, that’s probably good news. Here’s how it went with Engineering:

    (i) A few people decide to f**k up Engineering.
    (ii) The University Council issues a press-release on the imminent closure of Engineering.
    (iii) Academic staff in Engineering get “consulted”. (They are not amused and decide to fight back.)
    (iv) The university has a “review”.
    (v) Engineering is back, “restructured” to fit the needs of the “new South Australia”.

    I confess that I am not a leadership kind of guy, but I would have proceeded in semi-reverse, thus: consult the staff, have a review …

  6. *COMPULSORY* UNDERGRAD MATHS (AND RELATED TOPICS) ONE UNI I JUST VISITED:

    Computer Science: Maths and Computers, Discrete Maths, Linear Algebra, Probabilty and Statistics, Theory of Computation, Algorithm Analysis. (In my view Numerical Analysis should also be compulsory.)

    Electrical/Electronic Engineering: Probability and Statistics, Signal Processing, Communication Theory, Control Theory. These are from second-year on; in the first year, the “common engineering” students has other maths topics for all students in the College of Engineering”. ((In my view Numerical Analysis should also be compulsory.)

  7. Terry, a couple of additional comments.

    (a) It’s all very well that Engineering at Flinders is now being “restructured to meet the needs of the new South Australia”. But, as with Maths, bizarre actions have made the task much harder than it should have been: in their inscrutable wisdom, the powers decided to stop student-intake and Engineering has not had any for 2 years. Why they did not continue taking students while “consulting” and “reviewing” is a mystery that defies common-sense.

    (b) It being a slow afternoon, I have dug up my notes made at various departmental, school, and faculty retreats over the last decade. Page 3: “Nice talkfest; try to stay awake and think of the excellent food at lunch.” Page 7: “Great to know that we are engaged in best practice. Hope they will tell us what that means, seeing as we are doing it.”. Page 14: “Pity that *our most valuable resource* has so many elements that are unmotivated, unwilling to learn, and apparently semi-illiterate.”. And so on. Lunch was excellent. Grand plans were made. Fantastic futures were mapped out. Never to be heard of again until the next strategic retreat.

    My point again, and this is not restricted to just Flinders (where the bungling has been extreme) is management, management, management. Take Uni Adelaide, for example; in the last decade, they have not been shy about ejecting a Vice-Chancellor and a Deputy Vice-Chancellor who were simply not up to the jobs.

    On a more forward-and-upward note: In my opinion, one of the ways in which moribund Australian maths departments could revive themselves is is to get into more applied mathematics and to use the procceeds thereof to subsidise the pure maths. Examples of applied maths that I have in mind are “financial” maths (the folks in those industries can’t seem to be able to get enough good mathematicians and are willing to pay big bucks), bioinformatics/computational biology, cryptography, and mathematical modelling in many areas.

  8. Dear Egghead

    Some of the things you suggest have been considered at Flinders. In particular, bioinformatics and mathematical modelling. Neither ever had any serious support within the university.

    At the time the Dean of the faculty was a biologist, and there was some support (at the faculty level only) for bioinformatics. But it was almost entirely all talk.

    Mathematical modelling never had any support, and it did not help that one of the people who played a key role in emasculating maths at Flinders had by that time become the university’s research manager.

    Nevertheless, in a way the people in both groups share some blame, in that they did not accurately read the “political” climate at the university. By way of contrast, take nanotechnology. That group did a great job in selling a bill of goods to people within and without the university. Students were told long, tall tales about the wonderful future awaiting them if they enrolled. A rag called the Adelaide Review ran a long story on how Flinders had started nanotechnology and the rest of the world was rushing to catch up. A legislator got up in State Parliament and demanded an immediate infusion of funding so that SA could maintain this fragile lead. The people on the top floor of the Registry building fell over themselves with delight.

    New programs might work, but it would require real support from top floor of Registry. One of the things I hope will be considered for the new Computer Science program is one of two topics in computational biology. That would be a plus for maths, computer science, and biology.

  9. At the risk of causing distress on a mathematics blog, let me note that reducing the mathematics requirements for entry to computer science is not in itself, necessarily, an evil or inherently stupid action. Large parts of contemporary computer science, including some fundamental, theoretical areas, are not mathematical, and the skills and mind-sets demanded are those more commonly found in the domains of (for example), biology, economics, sociology, lingustics or philosophy, This is not because of the interactions that computer science has with these other disciplines, but because contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities. Contemporary computer science, both theoretical and applied, concerns a great deal which is unrelated to the mathematical study of computional complexity, abstract Turing machines, etc, and it is quite possible to make major contributions (theoretical, as well as applied) to the subject without knowledge of mathematics.

  10. Dear Peter,

    You raise a good point, and I would agree that at the postgraduate level and beyond it does make sense to specialise one’s studies to a subset of one’s field, which in the case of computer science may well be a subset with no particularly strong mathematical content. But I would argue that at the undergraduate level it is strongly preferable to have a broader education that encompasses all the basic aspects of one’s field (and also some exposure to related fields) – in part because it is too early at the undergraduate stage of one’s career to decide exactly what to specialise in, and so one should keep one’s options as open as possible, but also because employers may expect a certain amount of knowledge associated with a Bachelor’s degree in a given discipline, and having an institution offer such degrees while only teaching a subset of that knowledge may eventually damage the reputation of the credentials offered by that institution.

    At a bare minimum, I would expect a Bachelor degree in CS to be sufficient to allow two computer scientists to be able to communicate professionally with each other, even if they work in completely different subfields such as (say) complexity theory and networking. It’s hard for me to see how a CS degree with no mathematical component could offer that.

  11. Dear Peter

    Reducing entry requirements is one thing; reducing requirements while in is another. Pray, tell me how you would deal with, say, the following examples, all of which I have witnessed:

    (a) Student tries to write a program for an unsolvable problem.

    (b) Student with zero knowledge of error analysis writes a program that bombs because it is badly structured for floating-point arithmetic.

    (c) Students want to understand contemporary cryptography algorithms but has no use basic number theory, fields, modular arithmetic, etc.

    (d) Think of all the cases where discrete maths, numerical analysis, basic logic, etc. come in.

    Etc. etc. etc.

    I have seen computing programs devoid of maths, and I , like Egghead, would not (if I were looking to hire a computing person) employ any of the graduates of such programs.

    I am a senior editor of a very respectable computer-science journal, and once in a while (most recently, 2 months ago) I get papers with strings such as your “contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities” . (What on earth that does mean?) At one point, I tried to be “fair” and send them out for review. These days I see no reason to ask others to spend time on babble that I myself don’t understand: I immediately return the paper with a suggestion that the authors try a social-“science” journal, or that they get down to real work if they want to avoid giving computing a bad name.

    I have no problem with people trying to do social “science”, after all, we all need to eat, but why drag computer science into such untidy business? great deal of attempts to drag computing into “sociology, linguistics …” has resulted in a great deal of atrocious nonsense.

    Can you give some examples of these contemporary computer systems and science that you have in mind and try to give a better justification of your claims?

  12. Dear Peter:

    Not every course that involves studying something about computers is computer science. You may be somewhat confused there.

    The world’s largest professional organization for computing professionals is the Association for Computing Machinery. Among other things, the ACM issues standards for undergraduate CS curriculum. ACM devotes a lot of effort to this, and you may want to read one their many lengthy reports. Included in what they consider a minimal CORE for undergrad CS are

    * Functions, relations, and sets
    * Basic logic
    * Proof techniques
    * Basics of counting
    * Graphs and trees
    * Discrete probability
    * Basic computatbility
    * Basic algorithm analysis

    In their most recent reports, the ACM recognizes that computing and its applications have changed tremendously. But nothing in their reports leads me to believe that they are quite prepared for “ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities”. ACM’s idea of what constitutes contemporary computer science appears to be something like 50 years behind yours. Indeed, at this afternoon’s School’s Staff Tea, none of my 34+ CS colleagues present (we are a very large school) could explain what you might possibly have in mind, although a few were fascinated by the idea of a “self-interested” computer and “societies” of computers.

    On a more general view, I always found something to be said for the approach taken at two universities I have worked at. That view is that very student in a Faculty/College of Science has to take some mathematics courses, a couple of common ones and then some that depend on the particular subject.

  13. The notion that computer science can or should be taught without some mathematics seems to have appeared in the 1990s, in weak departments with weak students. I am not aware of any serious computer sience department, anywhere in the world, where such a notion would be entertained in any way.

    Curiously, I note that here in New Zealand such a notion never really took root but it got a lot of currency I Australia. Coincidentally or not, it is interesting to observe that both computer science and mathematics departments have been reasonably stable and there are no signs yet of any impending doom. I would be interested to read comparsions of mathematics in Australia and New Zealand. Perhaps the Aussies could learn something from us little Kiwis!

    Two people have commented on Peter’s statements. I suspect that this was all tongue-in-cheek. I doubt that any serious person would describe computer systems in terms more applicable to newely-found tribes in the Amazon rainforest.

  14. My comment seems to have generated some intense responses, many of which seem to be based on lack of knowledge of contemporary computer science or contemporary computer software engineering. I don’t have time to reply right now to everything, but let me quickly say the following:

    1. My comments were not tongue-in-cheek. The UK EPSRC, the main funding agency for computer science, has spent over GBP 10 million in the last 4 years on the use of methods from biology, ecology and economics in the design and management of complex computer systems. (Disclosure: I have one of these projects, valued at GBP 1.5 million.) The EC has spent perhaps 50-100 million euros over the last decade on research projects in the very same domain, and continues to do so. To give just one example of such a project, Britain’s leading telecoms company, BT, employed biologists, sociologists and economists in their research lab to develop (eg) algorithms based on ant-like behaviours for routing telephone calls. The USA’s DARPA has spent even larger amounts, including over USD 100 million on a single project to develop a multi-agent computer system to manage military logistics in Iraq.

    2. This is a very active research area in CS, so I am surprised that two CS commentators (Flinders guy and AI) do not know about it. The annual AAMAS conference series (Autonomous Agents and Multi-Agent Systems), for example, which has been running since 2002, usually attracts 500-600 full-paper submissions each year, and 700-800 people usually attend. The 2003 AAMAS conference was held in Melbourne.

    The recent papers of AAMAS are available online at:

    http://www.ifaamas.org

    Further information about this whole area is also available at the AgentLink site:

    http://www.agentlink.org

    (Disclosure: I was the administrator of this project in 2004-2006.)

    This research is both theoretical and applied, and is in the centre of computer science, not on the margins. It draws on ideas and methods from biology, ecology, economics, sociology, philosophy, linguistics etc to make contributions to CS and to software engineering. It is not social science, nor biology, etc. Nor is it “babble” or “atrocious nonsense”. Such statements reveal more about the speaker than about CS.

    3. Prominent research groups in this domain are in the CS departments at the University of Southampton, the University of Liverpool (where I am), the Spanish National AI Institute in Barcelona (IIIA), UPC in Barcelona, the University of Utrecht, the University of Bologna, Carnegie-Mellon Univ, Brooklyn College, Melbourne University, the University of Southern California, Stanford University and MIT, to name the main places. With respect, none of these CS departments are engaged in “babble” or “nonsense”.

    4. Flinders guy has listed 4 situations ( (a) though (d) ), none of which apply in this domain.

    5. I am co-editor-in-chief of a respected CS journal, The Knowledge Engineering Review, published by Cambridge Univ Press. Our journal welcomes papers in this domain. Other journals in this area include (all published by leading publishers): the Journal of Autonomous Agents and Multi-Agent Systems, Artificial Intelligence Journal, ACM Transactions on Evolutionary Computation, ACM Transactions on Adaptive and Autonomous Systems, and the Journal of Agent-Oriented Software Engineering.

    6. It would be nice if people who have an opinion on the undergraduate entry requirements for Computer Science in fact knew something about Computer Science.

  15. I overlooked something Helen said:

    “The notion that computer science can or should be taught without some mathematics seems to have appeared in the 1990s, in weak departments with weak students. I am not aware of any serious computer sience department, anywhere in the world, where such a notion would be entertained in any way.”

    Most British universities offer programs in computer science to students with only GCSE (School Certificate) mathematics, ie maths just to age 16. Usually these programs comprise courses in software engineering or information systems, etc, rather than in algorithm theory. These departments include the best in Britain.

  16. Dear Peter

    “You wrote: Most British universities offer programs in computer science to students with only GCSE (School Certificate) mathematics, ie maths just to age 16. Usually these programs comprise courses in software engineering or information systems, etc, rather than in algorithm theory. These departments include the best in Britain.”

    (a) If you point us to such best i Britain, we could have a better discussion.

    (b) Earlier I wrote,

    “Not every course that involves studying something about computers is computer science. You may be somewhat confused there.”

    Uder computing, ACM recognizes four different categories:

    * Computer Science
    * Computer Egineering
    * Software Engineering
    * Information Systems

    Note that all four could be offered (And sometimes are) by a single department; you should not confuse the name of a depatment with what it actually teaches.

  17. Dear Peter

    A couple of quick ones:

    (a) The fact that people are getting moey to do somethig in no way means that they are doig something useful. I can think of numerous examples.

    A quick example, the journal that I mentioned bills
    iteslf as inter-disciplinary, which tends to get all sorts out of the woordwork to annoy us. The paper that I mentioned was in fact the product of a project that had been very well-funded under the EU FP6. Unlike you, they did not use social-“science” jargon” but instead string together, seemingly at random, some technical computer-science jargon. My 4-line review was not kind. In my view, people who do nothing other than pollute the intellectual environment should be strongly discouraged.

    (b) I still have no idea what “contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities” means. In my view, the Emporer has acquired a new tailor. May I ask: Are these computers friendly, even though self-interested? Would they like to meet our leader?

    (c) You wrote: “It would be nice if people who have an opinion on the undergraduate entry requirements for Computer Science in fact knew something about Computer Science”. You have been pointed to the ACM and its curriculum in computer science. Which parts of that displays an ignorance of computer science? (Keep in mind that all accredited CS programs in the US, and many elsewhere, follow the ACM).

  18. Dear Peter

    I will address your concerns “one at a time”. But first, some preliminaries:

    (a) Can you tell me what, if anything, this has to do with undergraduate computer science; that is, how many CS degrees actually get devolve much time to these things?

    (b) I am aware of some of the journals you mention and also IEEE Transactions on Evolutionary Computation. I am also aware that some of them contain articles that cannot be understood by a person without a non-trivial grasp of mathematics. Is that not a sufficient reason to teach some mathematics to some of those people?

    (c) The number of people submitting papers to or attending a conference, or the amount of money some people have obtained should not be taken as a measure of quality. Think about that one? In any case, if you want to use that a a measure of worth, what percentage of computing money is actually going to these folks?

    (d) I agree with Former Flinders Guy. Regardless of where it is it is being done or who is doing the deed, ““ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities” is babble. What, for example, is a “self-interested” computer? What are the other kinds?

    Every researcher becomes moribund sooner or later; real work is hard work, and ideas eventually dry up. I have noticed that quite a few respectable people, in some very respectable places decide, when they reach that state but want to keep publishing, get into some cakey and flakey actvity. And since, they can’t do that within computer science proper, inevitably the ants and the sociologists and whetever else they can find get dragged in.

    It’s all very well for people to do such things in the privacy and comfort of their own homes, but it is going too far to dump it on an unsuspecting public. If I could, I would line up all such people and strangle them one by one. And I’d do it real nice and slow—to make sure they got the point. Hey, just kidding!

    (e) Regarding your comments on British universities—are you counting the post-92 ones, many of which are jokes and in some cases have been lambasted (see THES, for example) for issuing “junk degrees?— you should read Peter Hall’s comments. You should also read in THES all those articles of how British standards are falling as they let in more punters. Anway, which best British universities do you have I mind?

    Lastly, you seem to confuse entry requirements with what should be studied in a particular program. The fact that some British universities lets in people with only GSCE maths does not and should not mean that the CS programs do not or should not contain any maths.

    (f) It is utter nonsense to say that the stuff is at the centre of computer science. What percentage of computer science conferences, journals, etc. are devoted to the stuff? How come ACM, the oldest, largest, and most respetable organization in computing has missed the centre? How many computer science departments have even a single person doing such stuff?

    (g) The four cases that Former Flinders Guy are cases that a person with a good undergraduate degree in CS should be able to deal with, and employers of CS graduates should reasonably be able to expect that they ca. How do think they will be able to do that if they have been fooling around with ants and sociologists instead of studying the necesary maths?

    In particular, would you care to address Terry’s comment: “You raise a good point, and I would agree that at the postgraduate level and beyond it does make sense to specialise one’s studies to a subset of one’s field, which in the case of computer science may well be a subset with no particularly strong mathematical content.”

    Keep in mind that we are here discussing the content of an *undergraduate computer science program*.

  19. Peter, as I mentioned, ACM distinguishes between Computer Science, Computer Engineering, Information Systems, and Software Engineering. I need not go into maths requirements for the former two.

    For Informtion Systems, the curriculum guideline states that the syllabus should include, “prerequisites or interleaved, quantitative and qualitatives analysis, … including such topics as”:

    * discrete mathematics
    * introduction to calculus
    * introductory statistics

    For Software Engineering: similar to Information systems but with an additional course in discrete maths.

    I forgot to mention that as of 2008, there is a curriculum development underway for Information Technology as a fifth area.

    2008 Draft for Information Technology includes;

    * Applications of maths and stats to IT
    * Functions, relations, and sets
    * Basic logic
    * Graphs and trees
    * Discrete probability
    * Sampling and descriptive statistics
    * Hypothesis testing

    Now, how about you read the 1000s of pages of the five guides and then come back and explain how ACM has managed to get it so wrong—with maths in computing, self-interested computers, etc.

    ———————————————————————–

    PS. Peter, all the curriculum guides are for undergraduate degrees.

    To those offering dubious “information systems” and “information technology”, please pay attention.

  20. Dear Peter

    Now that I have some time, I was going to start with this;

    “many of which seem to be based on lack of knowledge of contemporary computer science or contemporary computer software engineering”

    But in light of preceding comments, I will await your explanation of why you think the ACM (and by extension, millions of computing people, thousands of computing-education programs, etc.) have managed to get it so wrong. In particular, what do you think of their view that no undergraduate computing degree (of any sort, seemingly) should be devoid of maths? Over to you!

  21. Peter wrote “… ACM Transactions on Evolutionary Computation, ACM Transactions on Adaptive and Autonomous System…”

    Did anyone else find it funny that ACM journals are being given as examples of why undergraduate computer science needs no maths at the very time that ACM insists that all good computing programs have some maths?

    Peter, you wrote: “Britain’s leading telecoms company, BT, employed biologists, sociologists and economists in their research lab to develop (eg) algorithms based on ant-like behaviours for routing telephone calls.”

    So, Peter, what would it take to develop a good understanding of algorithms? (Keep in mind that ultimately they must run on conventional computers.)

  22. *******I overlooked something Helen said:

    “The notion that computer science can or should be taught without some mathematics seems to have appeared in the 1990s, in weak departments with weak students. I am not aware of any serious computer sience department, anywhere in the world, where such a notion would be entertained in any way.”

    Most British universities offer programs in computer science to students with only GCSE (School Certificate) mathematics, ie maths just to age 16. Usually these programs comprise courses in software engineering or information systems.****

    Dear Peter

    As has been pointed out to you, there a difference between entry requirements and what one studies once in. Also, note that I refer to “computer science”, whereas you have in mind “software engineering” and “information systems”.

    kind regards
    Helen

  23. Peter, regarding this: “With respect, none of these CS departments are engaged in “babble” or “nonsense”.”

    I have looked at webpages. Obviously the *departments” are not engaged in babble or nonsense. But there are some individuals apparently doing a roaring trade in recycled snake-oil.

  24. Dear Peter

    When you first posted, I did not realize that you are working in these areas. Now that I know you do, I can see that they must be the centre of computer science. And if you have no use for maths, then, certainly, we should knock it all out of computer science. All that being so, this is my last comment on the matter. But I would still be interested in your opinions as to why ACM etc. have got it all so wrong.

  25. Flinders guy wrote:

    “(b) I still have no idea what “contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities” means. In my view, the Emporer has acquired a new tailor. May I ask: Are these computers friendly, even though self-interested? Would they like to meet our leader?”

    Perhaps an example will help you understand. The New York Stock Exchange requires all transacations over a certain size (about USD 1 million) executed by software programs to be reported. The NYSE then publishes each week the proportion of such trades (called “program trading” in financial circles). Most weeks it is around 30% of all trades by value, but in some weeks, it is over 70% of all trades by value.

    These programs, called agents or softbots in computer science, are goal-directed, in that they typically seek to achieve some end state for their human principal (eg, to maximize profits, or to maximize volume of transactions, or to spread risk); they are autonomous, at least within the constraints set by the programmer; to a greater or lesser extent, they are intelligent, in that they may (eg) automatically parse and process text-based news bulletins or CNN video broadcasts before deciding whether and how to act; and they may interact with one another, insofar as they respond to actions undertaken in the market by other such agents. They may also interact directly, for example, with agents from the same human principal co-ordinating their actions or agents from multiplpe principles colluding with one another.

    From the point of view of the designer of the system in which these transactions take place (ie, from the point of view of the NYSE software developers), this is an ecosystem or a society of interacting, self-interested, goal-directed, intelligent entities. The NYSE team cannot assume that these agents are benign: indeeed, more likely than not, there will be malicious agents among them. Even if all these agents were benign, the NYSE team cannot necessarily predict the consequences of their interactions, and even if they could, the consequences of benign individual actions are not themselves necessarily benign. Some people have arged that the October 1987 stoct market crash was the result of negative interactions of rational individual-level decisions by such program agents.

    The design, development, testing, verification and management of such software systems is fundamentally different from previous software systems, as indeed is the theory underlying them. Some theoretical computer scientsts (eg, Peter Wegner) have argued that it is not possible to model such interactive systems with anything like the traditional Turing machine models.

    If you think this is just hype, and not something new, then as co-editor of KER, I look forward to seeing your article arguing your case.

  26. Peter, I will later look at that and get back to you. In the meantime, how about the ACM stuff?

  27. Dear Peter

    First, a general remark: A person with serious cancer will try just about anything. So the fact that some seemingly serious person buys something should not be the sole measure of its worth, especially if the person can affors to hedge his bets. In this case, the cancer appears to be greed for money.

    You stated: “They are intelligent, in that they may (eg) automatically parse and process text-based news bulletins or CNN video broadcasts before deciding whether and how to act; and they may interact with one another, insofar as they respond to actions undertaken in the market by other such agents.
    *This* is what is considered intelligence? No wonder many Artificial-Intelligence did so well selling us a bill of goods, with all sorts of hype and promises that never came to anything. By the way, they too got a lot of money from a quite a few places. (These days, when I here “AI”, I reach for my gun.) Reminds me: Not too long ago some people approached me to join a grant-proposal that had something to do with “intelligent” software and hardware. After reading the proposal carefully, I asked them where the intelligence was. They admitted there was none but the would-be funders were interested in “intelligent” systems, so there. I did not join them, but they did come away with a nice chunk of change. By the way, how is what you describe different from a system consisting of finite automata, each of which processes input in certain ways and communicates with other finite automata based on some conditions? After all, automatic parsing and processing of text are tasks that cen be done by rather simple finite automata.”

    You wrote: “From the point of view of the designer of the system in which these transactions take place (ie, from the point of view of the NYSE software developers), this is an ecosystem or a society of interacting, self-interested, goal-directed, intelligent entities.”

    I would not dispute that they are interacting and possibly goal-directed. Self-interested, I still do not understand. The “intelligent” bit is laughable at best.”

    You wrote: “Some people have arged that the October 1987 stoct market crash was the result of negative interactions of rational individual-level decisions by such program agents.”

    I can well believe that. And some people have also argued that the moon is made of green cheese. Look, in both the case of the 1987 US crash and 1997 Asian crash, there about as many theories and arguments as you can find “some people” out there. The mere fact that an argument has been made does not establish anything factual. If you wish to sell that line, you will need to come up with real hard evidence.

    You wrote: “The design, development, testing, verification and management of such software systems is fundamentally different from previous software systems, as indeed is the theory underlying them. Some theoretical computer scientsts (eg, Peter Wegner) have argued that it is not possible to model such interactive systems with anything like the traditional Turing machine models.”

    Now, here you finally have some meat that we can sink our teeth into. I will, for now, skip the first part and wait until you tell me how the design, testing, verification, etc. all different. But I will give you a hint of what is to come: There are nowadays many sorts of real-time embded systems in which multiple processors do there own thing while communicating with other. The folks who do such systems do design, development, etc, with get bots (selfless or self-interested, friendly or malicious) involved. Once you tell me how all that is different for your systems, I will explain to you why it is not. Please, in your explanation avoid all that funny language.

    The second bit can be tackled right now.

    First, I will be pedantic: At the end of the day, your self-interested (perhaps friendly) thingies end up running on perfectly ordinary computers. Regardless of how many computers you have, the thingies and computers are in total a finite system. Therefore, the whole schmoogle can be described by a single Turing machine or equivalent. If you disagree with that, then I am keen to know why.

    Now, I will cut you some slack and assume that each of your computer system has an infinite amount of memory. In such a case, simply go with a non-deterministic system, Turing or whatever.

    You wrote: “If you think this is just hype, and not something new, then as co-editor of KER, I look forward to seeing your article arguing your case.”

    The impression that it is snake-oil comes from the language you are using. Perhaps if you just used English, but might achieve a better understanding.
    I will think about the article and get back to you by the end of the week. The stuff doesn’t seem to be terribly difficult or very diffierent from many systems I have seen, so why not.

    LASTLY: Please note the “Former” in my nome de plume. As it is, I suspect that my former colleagues might well prefer “***FORMER !$#%!” to “Former”. Lets be sensitve to their sensitivities.

  28. Dear FFG:

    I do not have the time at present to respond to each of your comments. In any case, I object very strongly to the tone and language of your comments (“snake oil”, “atrocious nonsense”, “babble”), which indicate a complete lack of respect. I see no reason to participate in a debate with someone showing no respect for me, for my work, or the work of entire discipline, particularly when your stated position is clearly one based on complete ignorance of the discipline in question.

  29. Dear Helen

    You said:

    “As has been pointed out to you, there a difference between entry requirements and what one studies once in. Also, note that I refer to “computer science”, whereas you have in mind “software engineering” and “information systems”.

    In my original post, and my subsequent posts, I used the term “computer science” in the sense in which I believe this term is commonly understood (at least in Britain), to include (eg) software enginering, information systems, e-commerce, etc. Perhaps some confusion has arisen if you are using the term with a more specific denotation.

    In the case of multi-agent systems, the traditional division between so-called theoretical and so-called applied topics (such as “software engineering”) is IMHO quite problematic. The very subject matter of multi-agent systems are software systems which are human artefacts (eg, the trading platforms used by electronic stock markets), and so there can be no theory in this domain without reference to some application — by definition here, any theory is a theory about something applied.

    There is thus a deep and intense interplay between theory and practice in this area. The creation of these artefacts has been informed by the theoretical abstractions used to describe them (usually borrowed from economics, sociology, political science and/or biology) and vice versa — the activity of theorizing about such systems is informed by the practical experiences gained in developing and managing real or prototype systems.

    Accordingly, computer science in this domain is a mix of both theoretical and applied activities, and I do not think it makes sense to separate agent-oriented software engineering from theoretical development of agent systems. It makes no sense for undergraduate or postgraduate students to learn just one of these. Some of the science is driven by experiences gained in building prototypes and the results of simulations conducted on software models. The theoretical developments thus created in turn generate concepts needed to be tested and refined via prototypes and simulation models.

    As it happens, I believe the same dynamic has driven traditional computer science, except that we have forgotten or not been aware of this. The subject of computing began with applications — eg, Jacquard’s looms, Babbage’s machines,etc — and theoretical abstractions were only developed a century later. These in turn led to new practical applications and methods, which in turn fed further theoretical developments. The invention of reverse polish notation by the Australian philosopher Charles Hamblin in the late 1950s is a good example: he was led to develop RPN as a clever way to overcome tight memory constraints on the early DEUCE computer he was programming. But he was able to develop RPN only because, as a philosopher, he had earlier studied propositional logic, including Lukasiewicz’s notation (so-called “Polish Notation”), something most mathematicians and electronics engineers at the time (perhaps even still) knew nothing about.

  30. Peter wrote:

    “I do not have the time at present to respond to each of your comments. In any case, I object very strongly to the tone and language of your comments (”snake oil”, “atrocious nonsense”, “babble”), which indicate a complete lack of respect. I see no reason to participate in a debate with someone showing no respect for me, for my work, or the work of entire discipline, particularly when your stated position is clearly one based on complete ignorance of the discipline in question.”

    Very funny, about the lack of respect, considering your statements abot people being ignorant about contemporary computer science and wishing they knew something about computer science before commenting on it. Look, here are some basics:

    (a) You have been challenged with regard to the ACM curriculum, and not a peep has been heard from you.

    (b) You brought up the matter of your bots in the NYSE, claiming that the design, testing, verification, … modelling are different from those of anything we have hitherto known. You have been challenged on the matter, and not a peep has been heard from you.

    These are tough matters, and I was wondering how you would get out of them. “You are ignorant about what am doing, and you don’t respect me” is probably as good a way as any.

    Somewhere at the back of my mind, there’s a Decent Guy saying I should apologize, but it’s hard: I really think that “contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities”. Low quality, recycled snake oil at that.

    I have just read a bunch of papers in the areas you mention. They, as well as your latest reply to Helen, remind me of a fish I once tried to eat. Monkfish or something, I think it was (but can’t be sure). Anyway, the fish looks real good and juicy on the plate, but stick a fork into it and all you find is a bunch of bones.

  31. Dear Peter

    I am sorry that your feelings have been hurt, and I agree that some people have been quite rude to you. But you were not very polite in how you barged in, and, by your own admission, you knew you might cause distress.

    I suspect that what people have problems with is just the language, e.g. computer systems as ecosystems. Why not try to be clearer?

    kind regards
    Helen

  32. Dear Flinders Guy —

    On your two so-called “challenges”:

    (a) I did not reply because I fail to see why anyone’s academic teaching or research should be pre-approved by, or conform with, a trade association, let alone a foreign one.

    (b) I gave the example of the NYSE because you told me you were unable to understand my statement about contemporary computer systems being viewed as ecosystems. I gave this example in good faith, assuming you were keen to become educated on the matter. I see now that this is not the case.

    A person either understands this position (of computer systems as ecosystems), or they do not, and you apparenty do not. Lots of people do understand it, which is why national and international research agencies, such as the EC and DARPA, and major companies such as IBM, HP, EDF, Boeing, British Telecom, etc, have put money towards researching these problems, and why the topic is currently so important in computer science.

  33. Peter wrote:
    (a) I did not reply because I fail to see why anyone’s academic teaching or research should be pre-approved by, or conform with, a trade association, let alone a foreign one.”

    Peter, a couple of things:

    (a) (i) First, ACM is like the British Computer Societ; so your sneering at it as a “trade association” is meaningless. Almost every academic computer person in the USA belongs to it, as do millions outside the USA. (ii) More to the point, I noted that when you wanted to support your case, you did not hesitate to pull up a couple of ACM journals. Trade magazines, are they? (iii) The Americans are much better at computing, so it would pay you to listen to them.

    I’ll give you this, though, this new attempt to wiggle out is definitely much better than the last one.

    (b) You gave the example to try and convince me of something. I was not convinced, and I indicated why and left asking for more clarifications from you. I am a reasonable fellow and fully understand why you do not want to go further with your case; so let’s leave it there.

    (c) You are right: I do not understand it. And doubt that I would “understand” something like ” contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities” without first ingesting a copious amount of wine, inhaling a big fat joint, and rounding it off with three tablets of LSD.

    Look, I am a man of the world. I understand that you folks too have to make a living. But why not do it quietly—on the sly, as it were? Why risk public embarassment? Rushing here to set us straight only exposes you to well-deserved missiles.

    You stated that you risked distressing people. Well, you did distress people, and they distressed you right back—five times as hard, just to make sure you learn from it. So, there having been distress all around, how about we call it a day?

  34. Helen — Two reports which may be useful as an introduction to the idea of computer systems as ecosystems:

    1. Seth Bullock and Dave Cliff [2004]: “Complexity and Emergent Behaviour in ICT Systems”. A report completed for the UK’s Department of Trade and
    Industry. Available as a HP technical report from:

    http://www.hpl.hp.com/techreports/2004/HPL-2004-187.html

    Cliff is now Professor of Computer Science at Bristol University and Bullock at Southampton.

    2. The AgentLink Agent Technology Roadmap [2005], available from:

    http://www.agentlink.org/roadmap/

  35. Also, I just remembered a new, multi-million euro computer science research project, which has just started in Spain, on the topic of “Agreement Technologies”. (Search on “agreement technologies Spain” for the website.) For those who still doubt that our work is computer science, the home page of this project includes the following text:

    “Computer science has moved from the paradigm of an isolated machine to the paradigm of a network of systems and of distributed computing. Likewise, artificial intelligence is quickly moving from the paradigm of an isolated and non-situated intelligence to the paradigm of situated, social and collective intelligence.

    Agent Technology is the latest paradigm of software engineering methodology. The development of autonomous, mobile, and intelligent agents brings new challenges to the field. Agent technologies and multi-agent-systems are one of the most vibrant and active research areas of computer science. At the same time commercial applications of agents are gaining attention. The construction of artificial (agent) societies leads to questions that already have been asked for human societies. Computer Scientists have adopted terms like emerging behavior, self-organization, and evolutionary theory in an intuitive manner. Multi-agent-system researchers have started to develop agents with “social” abilities and complex “social” systems. However, most of these systems lack the foundation of the social sciences.”

  36. Peter

    A lot of this stuff is like self-pleasure: Some people do it, and there is nothng fundamentally wrong with that. There are even people, some of them “upright pillars of the community”, who will pay to watch others do it; that might raise an eyebrow, but it may be argued that people should be free to spend their money as they wish as long as they do no harm. But, it is hardly the sort of thing one should do a lot of or publicize widely. And no number or reports or lengthy books can justify the claim that it is good, normal, healthy exercise.

  37. Dear Peter

    Thank you for those pointers. I will look them up when I have some time. I think, though, that you may have missed part of my point. I think your community should be seeking to communicate with others, especially other computer scientists, and not just among themselves. If you want to convince your detractors that what you are doing is computer science, then you should phrase an argument that is different from “here, another person in our community has said it is”. Many remember the undelivered promises of artificial intelligence and are, somewhat rightly in my opinion, wary of new claims phrased in language that is hard to understand.

    You may disagree, but I really do think language is very important.

    kind regards
    Helen

  38. In my first comment above, I said that many contemporary computer systems are best understood as ecosystems or as societies of interacting,
    self-interested, goal-directed, intelligent entities. These software entities are typically called “agents” and the systems “multi-agent systems”
    (MAS). We lack an adequate mathematical account of such multi-agent systems. (That is, as far as I am aware — I would be delighted to be corrected.)

    There are probabilistic theories of systems of interacting entities, for example in statistical physics and in population biology (eg, epidemiological models), but these tend to be “top-down” in the sense that they usually assume the agents involved and their interactions are very simple.

    By contrast, models from mathematical economics (such as Game Theory and rational expectations models in macro-economics) are typically “bottom up”, in considering intelligent entities. However, these models tend to make implicit assumptions (eg, about epistemic states) and ignore constructability issues which reduce their applicability to MAS. One key difference is that, unlike economics, designers of computational systems cannot assume that all the participants are utility-maximizers: Open MAS may include malicious or whimsical participants, or participants with otherwise inexplicable behaviours, because software code may always contain bugs.

    It is possible that a completely different approach is required to create adequate mathematical theories of MAS. If any mathematician is still reading these comments, I would welcome hearing of any alternative approaches or suggestions.

  39. Dear Peter

    I am not a mathematician, so I cannot comment on that aspect. I am a computer scientist, and from that point of view, I would suggest that you avoid words like “malicious”, “whimsical”, and “rational”. A piece of software might be written to do something bad or might have a bug. In such cases, what it does can be explained as such. Again, I see language as being the problem. For example, I do not understand what a “whimsical” piece of software is. I admit that I do not know anything of your area. I am interested, but I think the language will be very challenging.

    I will agree that sometimes a different way of looking at things can yield new insights and results, but sometimes new terminology can, rightly or wrongly, be taken as obfuscation.

    kind regards
    Helen

  40. “I did not reply because I fail to see why anyone’s academic teaching or research should be pre-approved by, or conform with, a trade association, let alone a foreign one.”

    Peter, perhaps you don’t have it where you are, but here in the USA we have something called accreditation for many types of degrees, including computer science. Accreditation is very important and does play a big role in what is taught at the undergraduate level, in that what is taght must be conform with what the accrediting body directs. In the USA, ACM is responsible for computer science, and we do not see it as just a “trade association”. As for research, the leading journals in computer science are from ACM or IEEE (another very influential “foreign trade association”).

    For American and Canadian computer scientists, ACM is a very influential body and not just a “trade association”. If you list 100 of the world’s top scientists, you will find that most belong to the ACM. The ACM also gives out the Turing Award, which is like the Nobel Prize or Fields Medal of computer science, and for many North American scientists being made a fellow of the ACM is very high honor indeed.

  41. Thanks, Helen, for your comments regarding language and communications within computer science. I agree with you that language is important., although once again I do not feel able to change the language already used in a research domain.

    Perhaps I should have mentioned that the area of multi-agent systems within computer science is over 15 years old, and (as I said in one of my comments above) now has a global community of researchers. This community can certainly be counted in the hundreds of researchers and is possibly more than 1000 strong. It would be inaccurate to consider the subject simply as part of AI, as its concerns are much wider than just machine intelligence — eg, software engineering, interaction protocols, network security, resource allocation in distributed systems, etc. The three AgentLink projects, funded by the EC between 1998 and 2006, spent almost 2 million euros developing publicity and communications materials, publishing regular quarterly newsletters, organizing and supporting annual summer schools, workshops and conferences, supporting journal special issues, supporting inter-university visits and collaboraitons, collecting case studies of commercially-deployed agent systems, writing technology roadmaps, organizing exhibition stands at conferences, and otherwise generally communicating the domain to people in industry and academia, both inside and outside computer science. The third AgentLink project had over 200 organizations (leading universities, companies, and research institutes) as member organizations. Having myself been part of this immense (and I suspect, unprecedented) communications effort, you may perhaps understand my frustration at being told now that I have not communicated these ideas adequately. If people in other parts of computer science are not already familiar with the ideas of multi-agent systems, then there is very little that I as an individual can do.

  42. Dear Peter

    I suspect then even people who are familiar with multi-agent systems would be put off by the statement that has been picked on by your detractors, i .e. “contemporary computer systems are best understood as ecosystems or as societies of interacting, self-interested, goal-directed, intelligent entities”. I suspect that had you started with something else to give your viewpoint on maths, you would not have encountered such rude derision.

    Similarly, some people may have a problem with software that is “irrational” as opposed to software that has not been properly written. Is it really impossible to use other words? I think you may have valid ideas but need to find a better way to communicate them to outsiders. That is what I mean by a language problem.

    The numbers you have given indicate why you need to do better at communications. You state that you now have a global community of 100s perhaps even 1,000 researchers. That is a very tiny percentage of computer scientists worldwide, and I think it would serve your community to be more outward looking.

    Finally, and please don’t take this the wrong way, but I really disagree with your last statement. As I see it, it is precisely because people aren’t familiar with the ideas that you should try harder.

    kind regards
    Helem

  43. Helen — With all due respect and with sincere gratitude for your courtesy, my experiences on this blog over the last few days have convinced even more of the truth of the final sentence of my previous comment.

  44. I recently started dabbling in cryptography, and today I got a request asking if I could identify some bright sparks for that sort of work. The ad (from DSO) is an example of the way I think Australian maths departments could revive their fortunes:
    ————————————————————

    Information Security Mathematician
    (Singapore[Map])

    Responsibilities:

    You will join a team of researchers to analyze complex cryptographic algorithms and protocols. Your work scope includes practical applied research and development of novel techniques to test the strength and discover the limitations of algorithms and protocols used in operationally deployed information security systems. You should be a motivated, self-learning individual with the persistence as well as strong creative and analytical skills to enjoy solving impossible problems and conducting information discovery in inscrutable systems.

    Requirements:

    Bachelor’s Degree in Mathematics.

    Preferably with concentration in algebra, finite field theory, combinatorics, number theory, probability and statistics.

    Strong creative and analytical problem solving skills.

    Good programming skills.

    Knowledge of Cryptology will be useful.

  45. Peter, catch up with your colleagues on foreign trade associations:

    “Wendy Hall, professor of computer science at the University of Southampton, has been elected president of the Association for Computing Machinery (ACM) for a two-year term. This is the first time in 60 years that a person outside North America has held the position. As a past president of the British Computer Society (2003-04), and a researcher with international connections, Professor Hall said she wanted to guide ACM towards more initiatives in India and China. She also said that her goal as president is to help ACM reach its full potential by increasing the number of women in all aspects of computing.”

  46. Hi Terry,
    the announcement from Flinders regarding their review of mathematics (you’ve provided a link to it in the last paragraph of your original post) has been updated: http://csem.flinders.edu.au/news/Item080703_MathematicsReview2/
    Cheers,
    Phil

  47. Dear Philip,

    Thanks for the update!

  48. For quite a number of years two people were very influential in that school. There was a man who was more interested in his career prospects than in the best interests of the school and had his eye elsewhere. The other was a woman who was allowed by the university to go on the basis that she alone knew everything and what was best for everyone. Between them they managed to do very well for themselves but none of the three departments in the school have improved their status.

  49. The university was without a vice-chancellor for 7 years. What did people expect?

  50. Correction to Egghead: Steven Schwartz is not a man to engage in time-wasting activities. He took Macquarie out of the IRU several months ago. But we continue to do innovation, research, and innovative research :-)

  51. Perhaps one of these days I will write an account of my time at Flinders. but it will have to wait till I can feel a little less dispassionate about it – if I live that long.

    I’ve enjoyed reading the interactions of Helen, Peter and FFG. Peter reminded me so much of prevailing attitudes in CS at Flinders when I was there – not by all members of that department by any means, but by the ones in power, one of whom is reputed to have told CS students that “Mathematics is dead”.

    It seems to me, however, that Peter’s latest email is suggesting that there is a need for a serious mathematical analysis of multi-agent systems – and I’m sure he’s right, but who is to do it if CS departments do not require mathematics?

    Finally monkfish is one of the great culinary delights of my occasional visits to Europe – please don’t equate it with a “bunch of bones”!

  52. 1. Sorry – I meant to say “more dispassionate” not less.

    2. Since DARPA has been mentioned let me say that I am currently involved in a project funded by that organization that brings together computer scientists, mathematicians, and engineers. All of the computer scientists in this project understand the language and ideas of mathematics. I don’t think anyone is arguing that there are aspects of CS that are not mathematical – just that not to provide for a fairly minimal mathematics course content as prescribed by ACM is going to significantly limit the choices available to CS graduates.

  53. This is getting to be habit forming – must stop it! One further remark. During the first few years I spent at Flinders I instigated a reform of the Math courses to include topics more relevant to CS. Much of what we (the mathematicians/statisticians) there at the time came up with reads like the list of ACM minimal requirements. As far as I recall the CS department did not recommend any of the new topics for its students, and indeed required a logic course taught by the Philosophy Dept in preference to the one we had specifically designed for CS students. I also recall that some CS students actually did both our logic course (taught at one stage by me) as well as the one by Philosophy – because our course helped them understand the Philosophy course.

  54. Indeed in many ways CS as an academic discipline would be much better served if undergraduate degrees in it were not offered at all, replaced by a minor sequence in pure maths programs. Most of the useful theoretical content as currently taught in undergrad CS could be compressed into a single semester, if only students had adequate maths background.

    Students are wasting time going to university simply to learn programming and it is painful to see. Unfortunately too many faculty members are only too happy to make it appear a meaningful pursuit.

  55. To DS and others: One does not have to be a mathematician to be a good programmer. But good programming requires the sort of logical mind that is most easily developed via mathematics of any sort.

    For more direct links, i.e. where a knowledge of maths is explicitly useful, consider programming for computer games, a huge industry in the UK. The latest surveys of employers in the games industry show that almost all want their employees to have more maths at university, and they are taking appropriate action. See

    http://www.developmag.com/news/30713/Tiga-calls-for-UK-to-improve-British-education

    For yet another example, go to Google and type “computer arithmetic tragedies”. This should give you some examples of “tragedies”, but a better search will yield real tragedies—where there has been loss of life. What is common in all these cases is that the programmers do not appear to have taken Numerical Analysis 101, which should be required for anyone writing scientific software.

  56. PS. I forgot to add that linear algebra is the maths course at the top of many employers’ lists in the computer-ganes industry.

  57. To FFG: You are, of course, right, linear algebra and numerical analysis courses would be infinitely more valuable in the curriculum than yet another course on database design or “intelligent agents”.

    But the bigger question is why would one want to train “programmers” – programming is a commodity skill and in my opinion having a university major upon completion of which students haven’t had exposure to anything else is doing them a great disservice. I have known many people with mathematics and physics training who would run circles around their CS peers.

  58. To DS: “CS as an academic discipline would be much better served if undergraduate degrees in it were not offered at all, replaced by a minor sequence in pure maths programs.” An “interesting” thought, but how about we get back to the reality of what CS actually is?

    People want to train programmers so that they can program. I have known many people with X and Y training who would run circles around the Z peers. How many solutions are there for F(X, Y, Z) = …. ? What does the existence of each of those solutions prove?

  59. Nothing, of course. It’s just a cheap rhetorical device ;)

  60. To Egghead: I recently discovered Google Scholar and decided to check out many who did well at the time when we mathematicians were held back. Interesting. Some people who were declared to be world-class researchers by promotions committee barely register. I take a microscope to Google Scholar and find only 2 or 3 citations.

  61. I am not surprised. When I worked at Flinders, all that mattered for promotions was the *number* of publications. Quality, citations, and so forth were unheard of. Indeed, even today, many of what the university describes as *world-leading researchers* have only a few citations from a large number of papers written over several years. But these people did, and, I imagine, still do sit in judgement over others. One hopes that the new VC will change things. Too late for me in my retirement ….

  62. I AM REALY SURPRISED WHEN I SEE THIS SCHOLARSHIP FROM AUSTRALIA .AND I AM REALY SURE I WILL JION THIS SCHOLARSHIP.

  63. Unless they bring back serious disciplines, such as maths, this is going to be painful. Forget the PR, the numbers of papers published, the number of grants one gets a free ride on, ….. Instead, look at number of citations (Google Scholar, Herzig’s Publish or Perish, etc.), the number of publications in the ARC’s A-ranked journals/conferences, impact beyond dubious claims in the Wed issue of The Australian, etc. etc. etc.

  64. Flinders does not appear to have performed well in the ERA, except for Medicine. Indeed, maths & computing was not assessed due to an insufficient amount of research. There are lessons to be learned there, but one wonders whether anyone here will learn them.

  65. I graduated in statistics from Flinders back in 1977 and it saddens me to think that this once great department has been cut so much. I look back now and marvel at the privilege and indulgence provided to a poor student like me. One-on-one tutoring in Fortran programming from Mike Georgeff, of analysis from Igor Kluvanek, statistics from Noel Cressie and John Darroch, and a lazy hour or so being walked through the foundations of arithmetic from that wonderful gentleman and algebraist, Bill Cornish. Google the names and you’ll see what I mean. Wonderful, astonishing people, breathtaking academic achievers, all laying out the treasures of mathematical history to a youngster who didn’t have the wisdom or experience to understand how outrageously he was being spoiled. The then maths department at Flinders will always be a treasure in my heart and mind, a real education from the finest.

  66. Back in 2008, I posted a contribution to this blog stating that we were in the process of building mathematics at Flinders. With the encouragement of many, including Terry, this process has come a long way since then. In particular, the University has invested in more staff, including our recently appointed Chair in Mathematics, Jerzy Filar, a new Bachelor of Mathematical Sciences award (to be offered from 2013) and the establishment of the Flinders Mathematical Sciences Laboratory. Further details can be found at http://www.flinders.edu.au/science_engineering/csem/courses/maths/ and http://www.flinders.edu.au/science_engineering/csem/research/centres/fmsl.cfm.

Comments are closed.

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: