Chapter 35: Hello CSE, Goodbye Tommy

The Science of Computer Engineering

March - April 1987

“Good morning, I am Doctor Will Morris,” I said to the assembled class. “I assume you are all here for subject 6.710, Introduction to Computer Engineering. If anyone is in the wrong place, I suggest now you leave quietly, and get to wherever it is you should be.”

I looked around the lecture theatre, there were maybe two hundred and fifty people, none of whom looked like leaving. Most of them had been in the previous lecture, the introductory lecture for Computing 1A. The challenge we had faced was to make sure there was a distinction between the two introductory subjects and not duplicate material across both subjects. I was the one most affected since my portion of the old 6.611 Computing 1 subject material was now included in 6.710, Introduction to Computer Engineering. I would be covering the ‘Computer Applications’ part of the subject; Document Processing, Spreadsheets, Data Bases, Graphics and Communications. But on Thursday, David had asked me to take what would be his lecture, with an introduction to structured programming and program design.

“This subject is designed as an introduction to the new course, Computer Engineering. I assume the bulk of you are taking this course, and if you are successful and graduate, you will be awarded a Bachelor of Engineering. So, what is the course, and this subject all about? It’s not a ‘computer programming’ course; let me emphasise that, if what you are hoping to achieve is a job as a computer programmer, then there are easier and faster ways to achieve that goal. Now, before I get into what this subject will entail, and ways to make sure you pass it – that’s probably what you think is the most important part of this morning’s lecture – let me tell you about the resources and staff we have here to assist you.”

I then spoke about the staff we had in the ‘First Year Unit’, how we were available for help with any issues they encountered at the university. I mentioned how they would have each of us as lecturers or tutors in some of their other subjects, and that if they had any issues with the university administration, Michelle was the best person to see to get them resolved. Normally, that would have been in my first lecture for Computing 1, but we decided to shift it to this subject.

“Okay, on to this subject, and the overall course. The first thing I will talk about is the name of the course and the Department that teaches it. ‘Computer Engineering’; why not ‘Computer Science’, or ‘Computer Programming’, or ‘Information Systems’? And why is the ‘Computer Science’ department part of the School of Electrical Engineering – or as we have become named now ‘The School of Electrical Engineering and Computer Science’? Some history; the university acquired its first electronic computer back in 1956, an English Electric ‘DEUCE’ computer, called ‘UTECOM’ – University of Technology Electronic Computer. It was assembled, tested and operated by some staff of the School of Electrical Engineering, one of whom was my father, working on it as a post-graduate project. So, by that happy coincidence – the school being responsible for the computer – ‘Computer Science’ was included in Electrical Engineering by default.

“The first department within the school that was involved with computers was the ‘Department of Electronic Computation’; and as UTECOM was replaced, first by an IBM 360/50 mainframe – which was located in the original computer centre, on the third floor of this building, right where our First Year unit is located. Many of the ‘Computer Science’ subjects, despite being taught by Electrical Engineering staff, counted as ‘Science’ subjects. Some Electrical Engineering students enrolled in a dual degree – a BE / BSc course, which the Computer Engineering course is designed to replace. The department name changed to the ‘Department of Computer Science’ in 1972, and the IBM 360 was replaced by a Control Data Cyber 72 mainframe.

“Now, in the next few years, the Department of Computer Science will split from the School of Electrical Engineering and Computer Science, to become a school within the Faculty of Engineering in its own right, the School of Computer Science Engineering. So let’s look at that name, as it gives a good background to what our career direction will be.

“Computer – and by that word I mean a stored program, digital computer. Some thirty years ago, a ‘computer’ was a person – usually a woman – who would perform complex mathematical calculations. This was well before electronic calculators; they would use tables of logarithms – often to seven digits – to perform calculations. Actuarial companies, aeronautics, NASA would use these human computers to do repetitive calculations, like what thrust would be required to achieve escape velocity for a certain rocket, things like that. But the computers that we will be dealing with are digital computers, which means they use binary arithmetic, ones and zeros, as opposed to analogue computers that work with continuously varying values.

“Science – and I see a number of you are wearing UNSW tee shirts, and part of the arms of the University is an open book, with the Latin word ‘Scientia’ across its page. ‘Scientia’ is the origin of the word ‘Science’, which means ‘knowledge’, a systematic approach to build and organize knowledge. At high school, you should have been taught about the ‘scientific method’, an empirical method where observation leads to making hypotheses, experiments to test those hypotheses, and communication of the method and results, to allow others to repeat the experiment. A good scientist is sceptical, not accepting anything on face value, not assuming something is true without evidence.

“Engineering – again, that word has a Latin origin, ‘Ingenium’, meaning ‘cleverness’, and ‘Ingeniare’, meaning ‘to contrive, devise’. An engineer is someone who uses scientific principles to design and build machines, structures and other items. The earliest engineers, back in ancient times, built structures like the pyramids, aqueducts, major buildings. In the middle ages, engineers would design and build windmills, machines like the cotton gin and spinning wheel, and the pinnacle of the master masons’ skill, massive cathedrals. They would work off empirical rules, like ratios of the height of a wall to the thickness of its base, designs for supports, buttresses and the like.

“So, what this course – ‘Computer Engineering’ – is all about is the design, construction, implementation and management of an end to end computer system. It will encompass computer architecture and hardware design, digital systems, software engineering, data management, computer networking, human interfacing ... the whole ball of wax, so to speak. Computer programming is but one small part of that; there are many other skills you will need to master to be successful in this field. How to work in a team, technical writing, project management, some of the subjects that you will do you will wonder at the time why are we wasting your time teaching them to you.

“This subject – at least I hope so – will not be one of those subjects. As well as having to listen to me, we will have some guest lecturers, graduates from the computer science course here who are working in the industry, as software developers, network engineers, project managers, and you can learn from them about what it’s like to work in the field, how things are done in the ‘real world’, and not in the artificial environment of a university course. You will be able to ask them questions about what they found to be essential skills in their jobs, and what other skills, things that we don’t cover here, are important once you start working.”

With the introduction out of the way, I talked about reading assignments for the subject, and told them the first one, on the history of the ‘Enigma’ code-breaking machine and Alan Turing’s work with early computers was out the front, and they could collect it on the way out. It was then into the meat of the lecture; an overview of a ‘typical’ computer systems installation. I used the System/38 installation we did for Bitola as a case study and explained how to manage a successful implementation, there are skills other than those directly related to computers that are required – logistics, building design, environmental issues, project management. I was able to project some photos that I had taken of the installation site to illustrate various points, particularly photos of the loading dock and freight elevator, plus some of the building to show how the area was modified to suit the equipment.

That pretty much wrapped up the first lecture; time to have my lunch, and shift my frame of mind to teaching the second year, Computing IIA. At least this subject was unaltered from how it was the last few years, and I could use the same set of lecture notes as I had done previously.

The rest of the week went pretty quickly, getting back into the swing of giving lectures, leading tutorials, talking to thesis students about their projects. It looked like being a good year, despite the bullshit programming language decision, I still couldn’t wrap my head around Lisp and Lambda functions, the silly thing was we would still have to teach them the rudiments of C as part of the second and third-year subjects. However, despite whatever programming language was being used, the fundamentals of problem-solving and program design remained the same, things that I would cover when I took David’s class on Thursday morning.

“Good morning everyone, I’m sorry to disappoint you, but Doctor Carrington won’t be able to take the lecture today, so you are stuck with me again,” I said to the class. At least no one groaned!

“As I said in our first meeting on Monday, this subject is split into three parts, an introduction to computing applications, which I teach, an introduction to computer systems, which Doctor Sammut teaches, and you would have had the first part of that yesterday. The third part, where you get into writing some computer programs, is an introduction to software engineering, and this will be taught by Doctor Carrington. Now, ‘Software Engineering’ is more than just writing programs; it covers the full spectrum of developing software, from defining problems, logic, abstraction, how to write specifications, algorithms, data structures, and, yes, writing programs.

“Now, in yesterday’s lecture, Doctor Sammut would have talked about, at a fairly high level, the hardware structure of a computer, and mentioned the Central Processing Unit – the CPU – along with other components, and how in the CPU, the actual machine instructions that make the ‘program’ are executed. Now those machine instructions are pretty basic, things like load a register from a memory location, store the value in a register to a memory location, add two values, compare a value, do some bit manipulation, set the memory address for the next program instruction. Writing programs at that level – which you will get to do in your third year – is slow and tedious, and the vast majority of programs are written in what we call ‘high-level languages’. Some of these languages – FORTRAN, COBOL, for example – have been around for quite a while, some have come and gone, while new languages seem to appear all the time.

“But, no matter what language we are writing our programs in, the process we use to design and develop those programs is the same. Here, we will be emphasising a style called ‘Structured Programming’; the aim is not just to write programs spewing out line after line of code, but to approach things systematically and logically. When you write a program, it’s not sufficient that it works correctly, that it produces the correct results, although that’s pretty important! There are other factors; can someone other than you understand how your program works, or can someone else make changes when the user’s requirements change? Does it make efficient use of the computer hardware resources? Do the people who use it find it easy to use and understand?”

I then use the analogy of an architect designing a house for a client, starting with a rough sketch, a basic floor plan, and gradually refining the design. Initially, there may be a vague square titled ‘kitchen’, which eventually gets completely specified with the location and layout of cabinets, appliances, plumbing and so on.

“Now the way we approach the design is gradually split the overall design into smaller and smaller parts, and there are several common models that we can use as a starting point. We will start with what I’m calling a traditional batch processing model; this goes back to the times when programs were all on punched cards, they would be scheduled and run, sometimes overnight, working on lots of individual records simultaneously, in a ‘batch’. Consider a bank processing loan interest rates at the end of the month, or the Government sending out pension cheques for example. The high-level model is three sequential steps:

“Firstly, we do some initialization, secondly, read each item and process it, until there are no more items left, and finally, any necessary finalization steps. Then we look at each of those steps, and add details.”

I showed a succession of diagrams of how the design of a program to work out the annual lowest and highest temperature based on a file of daily minima and maxima would be done. I mentioned how flowcharts, Nassi-Schneiderman diagrams, or pseudo-code could be used, all were appropriate tools, and which one a person chose was a matter of personal preference.

“What is important is to first solve the problem, then write the code.”

I mentioned that there are three basic patterns of program statements; the sequence, selection and repetition, and the example showed how these would be designed. The critical aspect of a good design is simplicity, I emphasised, tossing in a quote from Edsger Dijkstra, ‘simplicity is prerequisite for reliability.’

“This can be enhanced by the use of ‘program blocks’, where groups of statements are treated as if they were one statement; that’s part of a programming language design. The other feature, which just about all languages have, are ‘subroutines’, these might be called functions, methods, subprograms, depending on the language. These are callable units of code that allow a sequence of statements to be referred to by a single statement. Now, the reason why we go to what some might say is unnecessary effort is to produce maintainable code; if it is easy to read through, if the purpose and functionality the code is quickly apparent, then in a year, after you originally wrote it, you will – or someone else will – be able to modify it when user requirements change. Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live.”

That last comment caused a few laughs, and I finished the lecture with a discussion of their first tutorial assignment, to design a solution for one of our favourite programming exercises, the ‘collate the student results and print out a grade for each student’. I stressed that we didn’t want to see program code, but an outline of how the program code could be written, the solution breakdown, done in pseudocode, flowcharts or structured English, their choice.

Over the summer break, Michelle had prepared a document outlining project management techniques for our thesis students; it described how to use Microsoft Project to set up a project plan for their thesis, the sorts of reports that we – the thesis supervisors – would want to see regularly, and reference to some standard LaTeX document templates they could use to help with the production of their documents. She had circulated the document and the templates around all of the academic staff, to make sure that all potential thesis supervisors would be happy with their students using it. She would also offer tutoring session in Project and LaTex for those who wished to learn how to use that software.

That wasn’t the only development over the summer, David had been managing the work for WTM Consulting to the point where we needed to take on an additional two consultants to meet the demand. We were starting to get projects other than just software installation, configuration and upgrades; there was one project we had been asked to bid on, a complete computer room design and fit-out, managing the installation of a new IBM System/38, associated local and remote networking. For the building design and construction work, we had Paul and Chris Ramos with their company to call on, and their regular subcontractors for electrical, HVAC and fire protection. Since this client was in the financial industry, security and access control was important, so we could call on Frank Bubalo and Safe-T-Home Security Systems. Any architectural tasks we called on Ian Craig and Ross Trembath, nothing like keeping the work among friends and colleagues. That was something that Alexa had stressed to me; you look after your family first. The big advantage with using those related companies was that we knew their capabilities, and they wouldn’t try to take advantage of us, their pricing would be fair and we knew we would get their priority.

A key part of our project proposals was the inclusion of the initial project plan, Microsoft Project was essential here since we could produce detailed GANTT charts. The feedback that we had received from other clients was including the project plan, showing task durations, dependencies and resource requirements convinced them that we had a sound understanding of their problems and how to address them. It wasn’t always the lowest bid that won the jobs but giving the client confidence that we could complete the project, on time and within budget.

I now had some time to review the material that Bob McMullen had sent to me on proposals to reform the structure of the Federal Public Service, assuming the party won the next Federal Election. It was a rehash of Peter Walsh’s proposal that was shot down at last year’s policy review meeting, and revolved around reducing the number of Federal Departments from twenty-seven to seventeen, of those there would be eight ‘super departments’, along with the restructuring of the administration into three categories of entities. The departments would have a pure policy formulation role; and there would be statutory authorities to carry out regulatory and monitoring functions, as well as separate organizations to perform the ‘service delivery’ functions.

I couldn’t see any issues with the change in the number of departments, there had always been changes like that, new departments being formed, or abolished, having names changed. Apart from a brief period of disruption while these changes were being implemented, things would quickly get back to normal. It was the other set of changes that I saw some problems with. I started to make a list of what I saw as concerns, starting with a lack of ministerial accountability for the actions of these statutory authorities; if they had a ‘board’ or ‘industry committee’ overseeing their operations, would be how to keep the Westminster tradition of the minister being ultimately responsible for the actions of their portfolio. It would be all too easy to pass the buck to the board or committee, pleading ignorance, and how as a minister, he (or she) would keep a ‘hands-off’ approach with the running of the authority.

The other concern would be potential conflicts of interest for the board members if they were drawn from the industry area that the statutory authority was meant to monitor. I could see how easy it would be for this approach to lead to corruption, companies using the government regulatory agency to target their competitors and let their business get away with all sorts of criminal behaviour. Not to mention the potential for ministers to offer board positions to campaign donors; the approach of greatly expanding the number of government boards and committees would increase the opportunities for kickbacks and other corrupt actions.

What was more worrying to me was the proposal to split out the government’s ‘service delivery’ functions into separate agencies. This was the first stage to achieving the Libertarian’s ultimate ‘wet dream’, getting the government out of undertaking any sort of commercial operation, leaving the service delivery to the private sector. No doubt many companies were champing at the bit to take over the profitable part of government business; increasing prices to the public so they run these enterprises at a healthy profit, and reducing the level of service to the public at the same time. Of course, the only things that these private companies would want to take were the ones that would be profitable – I mean, what companies would want to take loss-making public transport services, or providing government services to remote rural areas of the country. And while I’m sure they would be lining up to purchase the right to run the main international airport in Sydney or Melbourne, who would want to take over the Social Security office for Gulargambone or outer Whoop Whoop? The real losers would be people outside of the major cities, all to give a gift to private industry mates of some politicians.

I couldn’t see the Liberal party opposing the changes, and the Nationals would see plenty of opportunities for them to hand out largesse to their party donors, so the only way to kill this proposal would be for the party to see it for what it was – a Trojan horse for the anti-government, libertarian elements. It certainly wouldn’t benefit the average Australian, and in my mind, that fact alone should be sufficient for us, as a party to reject the proposal. If something wasn’t seen as benefiting the average Australian, giving them an advantage, then why should we, as the Labor Party, be considering it?

I started writing my reply to Bob; the challenge would be to come up with arguments that would convince all sectors of the party that this was a bad idea. I read one section again, how service delivery in certain areas of the country could be ‘contracted out’ to certain groups that already offer community services. So, in a particular country town, the current Social Security office, the Commonwealth Employment Service office, and the Department of Veterans Affairs office could be replaced by a single office, run by the Salvation Army, or the Catholic Church, or Anglicare, or any one of a variety of community service groups. I didn’t want to be seen as ‘bashing’ church-run organizations, but the question had to be asked, how the government could ensure that these groups provided quality service to all clients, and not just those from their particular church group. Any supposed cost savings from using external contractors to provide these services would be eaten up with compliance monitoring, for sure.

I looked back at what I had written in my response to Bob McMullen, and after reading through it, I wasn’t happy with my arguments. They read like an ‘anti-right’ whine, which wasn’t the line I was trying to take, that way it would descend into a factional shit-fight, and wouldn’t achieve anything. A better approach would be to refer back to a party resolution from last year, where what we had discussed at a party meeting was the idea that all new policy proposals need to have criteria associated with them to objectively measure the success of the policy. The feeling was that this would reduce the number of pieces of legislation that were introduced for purely ‘feel good’ reasons. Some people fought back against the proposal, saying that there were some policy issues that the party needed to get into law, irrespective of any financial or other criteria, these were just ‘matters of faith’, and not subject to detailed analysis like what I had proposed. I would have thought objections would have come from the hard socialist left, but it was split fairly equally between those members and the right-wing. Both groups were driven by a pretty strongly held ideology, which I guess explained things. At least Bob Hawke, Neville Wran and Bob McMullen supported the idea, which in Bob Hawke’s words ‘should put an end to bullshit ideas, which are just a bunch of feel-good wanking.’ After those comments, the proposal that any policy changes, any new legislation would need to have measurable ‘goals and success factors’ was accepted.

Nowhere in any of the proposal documentation was a justification or any indication that the current structure had problems that needed any drastic change in the structure of the Commonwealth Public Service like the ones being proposed. Of course, there was no mention of any way to monitor any possible improvements in the level of service to the community, so how could we know if all of these changes were worthwhile? That would be the way I would work to kill this proposal; it ran counter to the recently adopted policy that unless some change could be shown to deliver measurable advantages to the general public, then we should not consider implementing it. The only reason that seemed to be behind the restructuring proposal was that it fitted into a ‘small and limited Federal Government role’ that was a key part of the right-wing ideology, but that wasn’t sufficient to justify the party making it our official policy.

I was happy with how I had rewritten that part of the response, so now I turned my attention to what some of the implications of going ahead with the policy would involve. I wanted to be diplomatic with any suggestions that this would reduce ministerial accountability or allow for corruption and favouritism in the appointing of board or commission members. I phrased it in a manner that when the opposition regained government, then we would be handing them a gift that they would most likely misuse. One other thing that we would need to consideration was the message we would send to the Commonwealth Public Service staff, and the associated unions by saying we would be privatizing their jobs and handing them to the private sector. The public sector unions were solid supporters of the party, and here we were ‘rewarding’ their support of us by saying that they weren’t needed, and we were finding other ways to get their jobs done. Certainly, we were not rewarding their loyalty in the way one would expect the Labor party to do.

I re-read my finished response, and I was far happier with what I had written. I called Bob and told him that I had gone through the restructuring proposal. He agreed with much of my thoughts, and I told him I would fax what I had written to him. Another task completed.

Sunday was the fifth heat in the Quartet Bowl series, we finished a close second behind ‘Young Generation’, they had picked a faster course back down Broken Bay after we rounded the Juno Point mark. As hard as we tried, we couldn’t make up the difference on the run back down from West Head. However, so far we had two second places and three first places, so with one discard in the series we had first place locked up. For the other series, depending on our results in the final heat of the Retriever Trophy, we would get either second or third overall, but with two heats of the Blue Water Shield yet to race, it was still uncertain, possibly a third, or maybe a second. Overall, a pretty good result for our first season. For the overall results, like the Shaz II and Commodore’s Trophy, which was the fastest overall yacht on scratch and handicap respectively, I wasn’t all that sure how we stood. Our results on handicap were nowhere near as impressive; which didn’t surprise me, it would take us another season, at least, to learn to sail the boat to its maximum potential. The way the club’s handicap system worked was if you sailed better than your previous results, then your handicap improved, there was a formula that was applied to each boat’s rating after every race. I didn’t quite understand the process; Neil Mattes was responsible for that part of the racing program. A thankless task, in my mind; no matter what rating a boat was given, someone would always complain; either they had too much of an advantage, or were being penalised for being good sailors. I thought that whatever handicap was given to you, you worked with that, tried to sail as well as you could, and let the system work it out from there.

It all sounded a rather complicated process, but it meant that the slower boats had a chance of winning races and series, it was the skill of the crew and not the boat itself that determined the results. That was why winning the Commodore’s Trophy was seen as the highest award the club offered; you had to really earn that through superior sailing, and not just go out, buy a fast boat so you can win races. There were still five races to go in the current season, which would finish just before the Easter long weekend. Over that weekend, Lisa and I were debating whether to go with some of the others in the cruising group for a long weekend cruise to Sydney Harbour. It would depend on whether Alberts were planning any ‘post-release’ gigs for the new album. George was planning on coming around next weekend so we could hear the final mix before they sent it off to cut the master disks.

Just before lunch on Monday I had a call from Bob McMullen, he had gone over the response to the Public Service Restructuring proposal that I had faxed him late last week.

“That’s a bloody good response, Will,” he said. “I’ve run it past both Hawke and Wran, and they agree, your approach is the way to go. They both asked if you are interested in a job as a speech-writer!”