Highlights of the Link Discussion List

**First Principles**

Skeeve's tale of woe (Re: [LINK] MCSE tale of woe) illustrates a pattern

I think has long existed an academia where practical, rigorous, 

engineering subjects are taught. I wrote about this a few years ago on 

Link, I think. 

The people who run the institutions have to hire teaching staff, but 

they have no idea how to tell whether the potential teacher has genuine 

expertise about the subject they will teach - except by referring to 

paper qualifications. 

Some people who are vitally interested in a rich, deep, important, 

ever-growing and demanding field often couldn't be bothered with paper 

qualifications. There are exceptions, such as in medicine, where you 

rightly have to have the qualifications before being able to practice. 

It is too often true that those who can't do, teach. Furthermore, those 

who can't do, or who don't even really understand the field, are often 

tempted to get paper qualifications as a job ticket. 

Quite a few people have no innate capacity for understanding anything of

genuine engineering complexity. They may not even appreciate what it is 

they don't know about engineering, so they may consider themselves 

qualified and useful if they pass the various formal tests. They may 

be good with word associations, playing the academic game etc. but those

skills are useless when confronted with a complex and unknown situation,

which must be resolved using high level debugging skills and with a 

great deal of knowledge of the many technologies and human factors in 

the situation at hand. 

There seems to be a self-perpetuating cycle of people with paper 

qualifications being less likely to possess the real knowledge of the 

field for which they are supposedly qualified. 

These people with paper qualifications are hired to run the courses, set

the exam questions and therefore control many factors which determine 

who will be the next generation of paper-qualified people, what of value

they will know, and what they will think they know. 

I once encountered a fully academically qualified electronics engineer 

who did not know the resistor colour code. This is like a computer 

engineer who can't instantly tell you that 0011 1110 is 3E in hex or 

that 0D 0A is a newline for Windows machines. 

Real skills with complex technologies are hard to teach, hard to provide

a genuine learning environment for, and hard to examine. It may be 

easier with software, web design etc. But electronic hardware and real 

live computer administration presents many challenges to the educator. 

I used to work at a TAFE college 20 years ago. There were a few really 

keen teachers who were dedicated to making the best of things for 

students, with as much hands-on experience, building and debugging 

electronic systems as possible. Most of those teachers wouldn't cut it 

as a technician - but they were training people to be electronic 

technicians. None of them had any real-life work experience in 

electronics. (The electrical trade section was entirely different - all 

the teachers there were sparkies from way-back, and they did great work 

training apprentices.) 

One of the technician teachers asked me why I squirted water onto the 

cleaning sponge of a soldering iron. Clearly, he had never used a 

soldering iron properly. Another asked me to check the calibration of a 

new oscilloscope. "What sort of errors does it have?" I asked, thinking 

perhaps he was being fussy about one or two percent. He replies that 

some channels seemed to be out by a factor of ten. He had never heard 

of oscilloscope leads with x10 attentuators built into them - as 

virtually all oscilloscopes have. These same teachers would be leading 

a class of calculator-pressing young men (and sometimes an intrepid 

young woman) through phase diagrams, AC and DC equivalent circuits, real

and imaginary currents, Norton and Thevenin equivalents etc. etc. (as 

per the exam requirements) yet if I had walked in with a BC108 

transistor (a common or garden transistor) virtually none of the 

teachers would know it was an NPN, have any real idea of its maximum 

voltage and range of beta (current gain) or be able to tell me which 

leads with emitter, collector or base. The same goes for most of the 

students - with the exception of those students who had an at-home 

interest in electronics. 

None of this high-level mathematical stuff is any use at all unless you 

have a hands-on physical-feel understanding of electronics. Design 

without long experience repairing equipment is madness. Many design 

issues are not related to things which can be expressed mathematically. 

So much depends on "common sense", physical robustness, serviceability, 

thermal design, resistance to dust and vibration etc. 

It takes years of genuine work experience to develop the debugging and 

general knowledge skills to be a decent technician. (I don't care for 

the "technician/engineer" "officer/grunt" distinction at all.) I don't 

see how those skills can be acquired in an academic course alone - and 

most courses do emphasise the importance of work experience. 

Since academia is often poorly paid compared to those who really know 

their stuff, especially in IT (there was a phrase like "gold collar 

workers" or similar), why would anyone with genuine expertise become a 

teacher? Few do - but there are people who like the social contact of 

teaching, the different working environment, and who have both a passion

and an aptitude for inspiring and helping people to learn. But even 

then, they have to teach by someone else's syllabus - probably the work 

of a paper-qualified committee. As teachers, their main task is to help 

students pass the exams, not to help them really learn what they need to

be useful in the real world. The teacher's performance, for official 

purposes, will be judged on their student retention and pass rates, not 

on how well those students actually perform in the real world in the 

years and decades to follow. Full time teaching typically means that 

they do not get to extend their skills or remain up-to-date with 

changing technology. 

I think it is really difficult to reliably test people's real skills in 

electronics, computing etc. So much of electronic and computing work 

involves debugging really tricky situations that it is impractical to 

set up such situations in an exam situation. Even for a skilled 

person, the time to resolution may vary enormously due to perfectly 

sensible decisions about what path to pursue first. Resolution may 

take much more time than would be suitable for an exam. Also, 

resolution involves access to web sites, discussions with colleagues, 

discussions with the people who run the system etc. It involves 

personal skills and a well developed capacity to look beyond the 

purported problem space into other related matters. For instance, 

the customer may report a problem in some specific way, but the real 

problem may lie in other equipment, in their usage of this or other 

equipment or in their faulty expectations and understanding etc. 

Many problems are intermittent. I once had a DX7 musical synthesiser 

which crashed and clobbered its battery backed up memory - but *only* 

when first turned on after being not used for a week! Eventually, by 

good fortune, the problem became more prevalent and I was able to use 

heat to prove it was one custom LSI chip, which must have had a faulty 

pin driver which was corrupting the data bus when first turned on. 

That resolution took a month! 

You can see the results of paper-qualified people being isolated from 

the real world in the many security vulnerabilities found in Microsoft 

software. When Outlook (Express) hands an HTML email to MSIE, and MSIE 

finds an attachment with a mime type indicating it is for Media Player, 

and Media Player looks at the file and decides it is not the right type,

and then hands it to some other thing which decides it is an executable,

and very helpfully *runs* it, then you know that this whole system was 

put together ("designed" seems too strong a term) by people who had not 

the first clue about security. 

Another problem is the vastly expanding need for technical expertise. 

In the early to mid 20th century people with a technical electronic bent

did "radio". Then it became electronics. Then computers and software 

was developed. Now we have various aspects of electronics, programming, 

system administration, computer building and repair, web site design, 

database management etc. Each of these fields keeps growing more and 

more complex, and being useful in one field often means having a working

knowledge of several others. Yet there is a constant supply of people 

with minds which can *really* do this stuff well. 

I think that the increasing complexity of everyday things acts as 

a real impediment to anyone wanting to learn many aspects of 

technology. 

My first electronic devices were things such as a 1940 Mullard valve 

radio. With a little theory, it was perfectly possible to understand 

the whole thing without a schematic. Even without a databook, you could 

see what the valves were (rectifier, beam power tetrodes, pentagrids, 

pentodes etc.) and see their pin connections. Then you could draw the 

circuit in an afternoon, and with a multimeter and perhaps an 

oscilloscope you could fix faults and modify it. 

Now, with LSI chips, surface mount etc. etc. most everyday items are not

at all amenable to understanding, debugging, repair or modification. 

Also, I think the educational system has been run by humanities types 

and an increasing proportion of women, who generally have no 

understanding of or interest in engineering. There was even a fashion 

against science and engineering at one stage - as if these were the 

roots of war, power elites and environmental destruction and so should 

be avoided. 

I don't know any teenagers now who are anything like as advanced in 

electronic understanding as I and my radio club colleagues were. (The 

Camberwell Grammar Radio Club no longer exists - we used to do all sorts

of things there, including ballistics research and 40 metre amateur 

radio.) Some young people in their late teenage years may be getting 

into computer programming - which when I was 14 meant hand-punching 

cards in a cut-down version of Fortran, to be transported to Monash 

University and run on the Control Data 3200 and returned with a 

print-out a week later! 

But even in computer programming there are pervasive problems. There is 

a generally low standard of documentation of the code inside the source 

files - which I put down to it being done mainly by young men under the 

influence of young warrior/hunter instincts, which have them strut their

stuff, showing what they can do, but not giving anyone any clues at all 

about how they do it. This is a field where some more feminine, English 

sentence, communicative aspects of humanity are sorely needed! Another 

possible factor in the too-terse, badly expressed, badly laid out or 

non-existent comments is that it takes up too much space in printed 

books, which constitute a guide for many programmers. 

There are a number of vicious circles at work here, including: 

1 - Increasing complexity and pervasiveness of an increasing number 

    of technologies while there is a finite number of people
with the 

    mental makeup, skills and passions to do it properly.

2 - Widely deployed systems written in lousy ways (Microsoft, for 

    instance) firstly setting low standards by which others
are 

    judged and secondly requiring vast armies of people to
mollycoddle 

    the second-rate systems into working as desired. 

3 - The vicious circle of academics being chosen only by paper 

    qualifications, of there being many reasons why good technical

    people do not become teachers etc. 

4 - Many people with good technical skills being hobbled by their own 

    lack of human communication skills, including their inability
or 

    unwillingness to clearly document their work for the next
poor 

    sod who tries to understand it, which will probably be
themselves. 

    So, when debugging or extending the software, they spend

    excessive time trying to understand what was in the original

    worker's mind (probably themselves) rather than using
that knowledge 

    from good documentation and therefore going straight into
the fray 

    fully informed. 

5 - Complexities of everyday electronic items being not at all conducive

    to teaching or understanding by beginners. Compare a PC
or video 

    game, to a crystal set or superhet radio. 

6 - Pervasive and generalised factors mitigating against patience and 

    long attention spans, including massive exposure to TV
and 

    video games, and also potentially problems with nutrition,

    mercury in vaccines (http://www.autism-mercury.com) etc.
which makes 

    many people too scattered and impatient to spend years
learning 

    something difficult like electronics or computer programming.

I just got an email from a journalist friend. The company she works 

has ten or twenty staff or so. She regularly uses free web-mail and 

personal email accounts, because of unreliability of the office email. 

She just wrote to me that she has had no email on the office system for 

two weeks, and virtually none for the two weeks before that. She also 

wrote that some co-workers can't access their hard drives. I don't 

have to mention which company's software is involved in this - and how 

most of the trouble seems to have arisen due to the Klez worm which 

relies on this company's software's well-known vulnerabilities. 

Complexity, when badly done, turns into a terrible mess. 

 - Robin 
Bernard Robertson-Dunn wrote:

> There is a world of difference between a technician and an engineer.

> It is the same distinction as that between a nurse/doctor, 

> architect/builder, analyst/programmer, DBA/designer etc. They are 

> different skill sets and responsibilities and have dfferent 

> education/training needs.

I agree in principle.  But questions like these can be answered in a

number of ways.  There is the formal, ideal, definition of something,

and there is the day-to-day experience of it.   For instance, science
-

depending on which philosophy of science you adhere to - is ideally a

rather different thing from what is often practiced.  Likewise religion

and many other fields.

My objection to the engineer/technician distinction is that I see no

pattern of qualified engineers being more capable at real-world things

then qualified technicians or unqualified people such as myself.  This

is only my experience, and I like to think it is not typical.  
For

engineers to think they can design things without having spent years

fixing design faults and the effects of wear-and-tear on existing

equipment is madness - and only perpetuates the need for more people to

fix the things they design.

Likewise, I think its nuts for people to become qualified doctors

without having worked extensively as nurses.   Similarly officers
in the

armed forces.   In BDSM, most people say that one should never
become a

dominant without first being a submissive.

> Having said that, I fully agree with Robin's assessment of the current

> state of academia and the way industry uses the products of our 

> current training systems. In fact I would go as far as to say that it

> is even worse than he claims.

The fact that a publishing company can pay good money for what it (and

most other companies) think is the best available software, and have a

full-time staff-person (no doubt with Microsoft paper qualifications) to

keep it running smoothly, whilst also engaging outside consultants, and

still have no email for two weeks and a number of completely

non-functional PCs indicates that things are very bad indeed.  These

people are just trying to run an office - not to nuclear physics.  I
the

productivity cost of the Klez worm is probably several percent of GDP of

affected countries - and this is primarily due to a stupid weakness in

Microsoft software which no-one in their right mind should ever have

foisted on customers.  It is also due to people in charge of

Internet-connected computers being clueless enough to click dodgy

attachments.

> Having done a couple of post graduate degree courses in Control 

> Engineering (MEng and PhD) I decided to update my library and bought
a 

> recent book on the subject. It covered most of the topics that I 

> remember doing thirty odd years ago, but there was one major 

> difference. It assumed that you had MatLab, a simulation package.

> 

> In my day, you either worked out the maths by hand - and gained a 

> thorough understanding of the problem, or you built devices and 

> learned from reality. Now, and this seems common in other areas of 

> electronics, students use computer simulation.

This does not seem unusual to people raised on video games.

I once lent a two valve (2 x 12AX7 I think) hand-wired computer module

from a 1950s computer to a friend who was in her final year of

engineering or its technician equivalent (it shouldn't matter at all

which, for the purposes of my argument).  I had never bothered to write

out the circuit, but I figured it was probably a dual flip-flop, since

it had a few metal film resistors, mica capacitors and compact selenium

diodes.  I suggested she draw out the circuit and try to figure out
what

it did.   She was keen and bright and a few weeks later came back
with a

CAD drawn circuit which may have been technically correct, but which

bore no resemblance to common-sense ways of arranging components between

ground and the positive supply.  Her approach to trying to figure out

what it did was to turn it into a SPICE model and simulate it on the

computer!!!   This is wholly deficient and useless.  She did
not have

any viable theory as to what it was.   Anyone with genuinely useful

electronic skills should be able to draw the circuit by hand and by

looking at it and then by thinking alone, figure out what it is meant to

do.

Computer simulation is used for all sorts of things now, and I

understand that there are ways of integrating almost everything into the

simulation / design process.  You can have the innards of CPUs, all
the

chips, and programmable logic devices simulated along with all analogue

and RF components, complete with the capacitance and transmission-line

characteristics of the circuit boards all simulated and linked with the

various PCB, PLD etc. design programs.  I think they can even simulate

electromagnetic emissions and susceptibility to interference!  

By the way, I am always on the lookout for things which used to be

impossible firs becoming concrete, weird, expensive artefacts, and then

becoming cheap, standardised commodities, and ultimately being reduced

to a pattern of bits in a computer.  Now this has happened to CPUs! 
You

can get a logic design for a complete 100 MIPS CPU, with RAM and all

I/O, to put in one part of a programmable logic device.  You might even

program two such CPUs into the one chip, together with all the other

logic guff to interface them.  Thus, a complete CPU is reduced to a

pattern of bits in an Electrically Erasable (Flash) PROM to be loaded

into a RAM-based programmable logic chip at power-up.  Then simulation

would involve a PC running a program to simulate a logic chip which was

implementing a CPU - and then simulating how that CPU executed the

software which was loaded into its program memory.   

While there is good cause to use this stuff to try to eliminate problems

before committing to a ten layer surface mount PCB design, often with

the pinouts of PLD chips set by the particular logic design and the

chip's inability to route any signal to any pin, the whole thing sounds

like a nightmare.  I am always impressed that anyone can do this and

bring mobile phones, video cards, hard disk drives and motherboards into

reliable existence before they become obsolete.

Firstly, you would have to know that all the models of all the

components were correct - but how could you know?  Secondly, you would

have to be sure the simulation software was free of bugs.  Thirdly,
the

simulation would have to run fast enough not to bog down the design

process with overnight and weekend runs.  Finally, the real test for
the

device being designed is the real world, and I can't see how one can

reliably simulate a GSM phone, without having also a simulation of a

base-station and all the vagaries of radio-wave propagation.  

Companies sell this super elaborate software - but I can imagine a team

of engineers taking months and months to learn it, to debug their

component definitions etc. etc. - all of which detracts from actually

designing the thing.

> Computer simulation is a powerful technique, but it is no replacement

> for real-world experience. 

Exactly.  One remembers the smell of a 1 watt resistor which dissipated

ten watts - and the burnt finger!   What about the way a tantalum

capacitor connected the wrong way round will work quite OK until at some

stage in the next seconds, minutes, days or years, suddenly in less than

a second, burst into flame, emit a puff of smoke (purple smoke perhaps -

I have seen it a few times) and then resemble a piece of burnt corn with

two leads attached.

There are so many interactions in devices which cannot be simulated. 

What about the metal fatigue on solder joints of inductors in the

horizontal drive circuits of video monitors or likewise in switching

power supplies.  Only experience as a service technician would make
you

learn that there are powerful forces on the wires in inductors and that

these are capable of fracturing solder joints.  Likewise expansion and

contraction due to temperature changes.

> I have blown more fuses, transistors, transformers etc than most 

> people will ever see. Am I a better engineer for it? probably. 

For sure!

> Am I appreciated for it in the work force? As it happens I work

> for a company that does value experience over youth and trendy 

> knowledge, but not all do. Does that have anything to do with my age?

> Probably, but that's another subject altogether.

Assuming that you are working effectively in a field, then you should

get better at it.  But what of the MSCE who wasn't working at all -
just

stuffing around.  In what way would that person be better at his work

after making a total mess of something because he lacked even the basic

clues required to resolve the problem?

> What can we do about it? Now there's the real question.  Has anyone

> ever been able to change society in a predicatble way?  IMHO the

> answer is no.  It is easy to change society, bombs, guns and power

> hungry politicians do it all the time. Is the resulting change what

> was intended?  Not in my experience.d

It may be what they intended.  I don't think things are so bleak, but
it

is the failures which tend to stand out.  Paul Keating had the right

idea putting Arts in with Communications. 

The appears to be a lack of proper skills and motivation amongst quite a

few political leaders and technical teachers.  Perhaps the solution
is

something between conscription and jury duty.   If someone actually

wants to be prime minister, then they are insane and/or a megalomaniac

and so are unsuitable for the job.  There was a US movie about

dragooning some bright unfortunate into the job of president.  
Even a

committee could do a better job this way than what unfettered democracy

and the TV generation did in the US.

Likewise, it might make some sense to have compulsory service and/or

high-level cash inducements to get practising, competent

engineers/technicians etc. into the teaching ranks for a year or two. 

Any longer than that, unless they do it part time, and they would lose

touch with their real field of work.   The teaching of doctors
and

nurses is, or at least used to be with nurses, tightly integrated into

real-world work situations with real practising professionals.  
I know

institutions do try to get work experience for their students.

Glen Turner wrote:

> And just straight funding.  Your typical modern electronics

> company has at least $0.5m of software to do FPGA design,

> PCB layout, EMC analysis, and so on.

Ouch!   Then everyone has to try to make the packages talk to each

other, and read all the manuals, updates etc.  Then they have to learn

to use them all, find out their bugs and limitations and somehow

construct a giant video game which is worse then useless unless it truly

predicts the behaviour of the thing they are designing, but have not yet

built.  This is really daunting, but it is a fact that designing complex

equipment is an extremely challenging task.  This is especially true
if

it is not just software and electronics, but involves servo systems, RF,

vibration, usability for non technical people, physical ruggedness etc.

> Also, electronics components prices are very sensitive to quantity.

> It's simply unaffordable to have a high-density PCBs made in

> quantities of one.  This is turn makes it hard to give undergrad

> students a 'real-world' project.

Indeed - it is very hard to provide a real-world experience in

electronics courses.  But before they are designing things, they should

be fixing and modifying things, I am adamant.

> This also happens in computing -- 2,500 lines of code is about

> all you can expect from a student in a semester project.  So they

> onyl start to glimpse the complexity of real-world software

> systems.

I like to think that this is 1500 to 2000 lines of comments, and the

rest actual code for the compiler to chew.   2,500 lines of actual
code

with few comments is a very large project which virtually no-one can

understand, including the person who wrote it, just days or weeks later.

My own experience when I have occasionally written a few lines of code

without proper comments (the comments come first) is that I often can't

understand what I was trying to do, or be sure whether the code does

that or something else, ten minutes later.

Craig Sanders wrote:

> 7.  The Cult of Mediocrity.   This insidious cult has
infected

> educational systems since the 1970s - it became more than 

> unfashionable to recognise that some individuals had talents 

> greater than other individuals(*).  recognition of merit was 

> forbidden.  streaming was abandoned, forcing the smart kids to

> learn at a slow pace that the average and dumb kids are just 

> capable of keeping up with.

Yes.  This has its roots both in the incompetence of those who came
to

control the education system (though they wouldn't recognise it as such)

and in the social-constructivist theories which have assumed the status

of undisputed truth in recent decades.  In this model, at its purest,

children are blank slates and everything which they become is determined

by experience.   There is no stupidity - only someone behaving
and

thinking badly because of their low self-esteem.   The self-esteem

movement has a lot to do with the Cult of Mediocrity.  

Also, passing people with 50% of their often multiple-choice questions

has got a lot to do with the Cult, because this sets people's general

expectations and determines who is regarded as qualified to teach and

set future syllabi (?) and examinations.

> the cult of mediocrity goes well beyond education, too.  it has

> debased the definition of democracy, so most people now believe that

> mediocrity and democracy are one and the same thing - i.e. the truism

> that "everyone is entitled to an opinion" has mutated into the lie 

> that "all opinions are of equal worth"....no recognition of expertise,

> or merit is allowed.

Its true.  We are all going to hell in a handbasket.   Like
Little Black

Sambo's tiger chasing its tail until it turns into butter, MS software,

flaky over-complex web-design programs etc. is damaging companies and

educational institutions ability to operate - and this makes it even

harder for anyone to make any progress.  Meanwhile, in China, they are

bright-eyed and busy tailed, working hard and smart, becoming proficient

in AutoCad shortly after learning to walk and churning out increasingly

astounding products for a fraction of what could be done in the

currently developed nations.

> it's pretty easy to spot this on mainstream mailing lists and 

> newsgroups where, for example people with no experience or knowledge

> in a given field argue vociferously with others who have, e.g., done

> their PhD thesis on that topic, and insist that their opinion is worth

> as much as the PhD holder's opinion.   

Yes, but some PhDs are extremely rigorously based and some are not. 
I

know one person who was doing his PhD in Canada on human communications

- and the process was very extensive, rigorous and resembled at times an

Inquisition.  It involved oral responses to a panel, literally, of

inquisitors - old beaks with PhDs and professorships probing and

challenging him on his work.  I met someone here who was well into her

PhD on the interaction of the Earth's magnetosphere and the solar wind,

and she couldn't even give a consistent explanation of the

magnetosphere.  Maybe she didn't continue or pass.  

> even worse, many of the other subscribers agree because they also are

> not qualified to judge, and have an unfailing belief in the fairness

> of mediocrity.

This is the Internet as a bullshit amplifier.  The mass media and quite

a few really crappy computer books do this as well.

Craig, I agree with what you wrote, but I suggest capitalising the first

letter of sentences in any treatise concerning encroaching mediocrity. 

Perhaps you subliminally influenced Jan to slip up in this regard with

one of her sentences!

From what little I know, apprenticeships make sense in stable

industries, if there is no exploitation, the apprentice's co-workers are

competent and helpful and if the education system can work well with the

actual needs of industry and the apprentice.  Jan and I have worked
on

programs in this TAFE (VET or whatever it is called this year) flexible

delivery business and it is complex and difficult thing to achieve all

these goals.  It assumes there are stable, full-time workforces with

career structures which remain stable for decades.  But web and database

design are new-fangled professions which really began in earnest perhaps

seven years ago - or maybe earlier for database design and management. 

Companies are often trying to be flexible and engage people with

multiple skills on a casual or contract basis, and these workers may be

working from home.  The apprenticeship system seems to be based on a

traditional large workplace stable full-time workforce model which often

does not apply in higher-tech fields.

A big part of the answer with electronics and computer programming is to

do it at home.  Everyone has a PC, and basic electronic tools and

components cost well less than $1k.  You really need a dual trace

oscilloscope, which is another $600 or so.  But these are minor

investments compared to the investment of time for studying electronics

or computer programming.   There is a false expectation that the

institution should provide everything on a plate.   If someone
is not

fascinated enough by computer programming or electronics to do it

themselves, intensively, for their own interest and probably for really

useful practical purposes, at home, then I doubt they will ever be much

use no matter how much formal education they do.

I will make a link to this discussion and probably put up my

contributions on a web page in the "Various technical things . . ."

department of my site http://www.firstpr.com.au .

 - Robin

Highlights of the Link Discussion List

Link discussions of special interest

Academia, technical skills etc.