OK, so it's been a while since I posted a big wall of text. So here goes...



The name of the game is data compression. If you already know how that 
works, this is going to be a VERY boring post. If you don't know, it 
probably seems like magic; you take a file, throw it at WinZip, and 
suddenly the file gets 4x smaller.

Now MP3 does that to sound files, but it does that by DELETING DATA. The 
compressed file has lower fidelity than the original. It's not that 
surprising that you can make a file smaller by removing data from it. 
But LOSSLESS compression algorithms do something seemingly impossible: 
they take a file, somehow make it smaller, and yet the process is 
EXACTLY REVERSIBLE. When you decompress the file again, it is IDENTICAL 
to the original. How is that even possible?

Well, it ISN'T always possible. Sometimes when you zip a file it gets 
tiny, and sometimes it doesn't get that much smaller. You may not have 
experienced this, but occasionally a file will get LARGER. For example, 
zip a file - any file - and then go zip the zip file. It WILL NOT get 
any smaller. I promise. (Think about it; otherwise you could take an 
arbitrary file and zip it over and over until it gets arbitrarily small. 
But that really IS impossible!)

Rather than dwell on this further, I'm going to show you some 
compression algorithms now. We're going to start with Huffman coding.

As you probably know, the American Standard Code for Information 
Interchange ("ASCII") assigns an 8-bit code to every letter of the 
alphabet. (You may or may not know that technically ASCII is actually a 
7-bit code; any text encoding which actually *uses* the 8th bit for 
something technically isn't ASCII, it's an ASCII extension.)

You might also be aware that Unicode defines zillions of characters as 
24-bit codes. In UTF-32, every character is stored as 32 bits. In 
UTF-16, almost all characters are stored as 16 bits, but particularly 
unusual characters take 32 bits. And in UTF-8, many characters take only 
8 bits, a few take 16 bits, the really rare ones take 24 bits and a tiny 
few take 32 bits. (The extra bits are necessary to signal how many bits 
make up the character.)

The various UTF encodings are of course designed to make encoding and 
decoding as painless as possible, while not wasting absurd amounts of 
space. UTF-32 is particularly easy to deal with, but very space-hungry. 
UTF-8 is the most frugal, but quite awkward if there are lots of 
multi-byte characters.

But suppose we want to assign codes to characters, not so that 
processing is easy, but so that space is minimised. Ease of use is not a 
requirement; we just want to crunch the space down as much as possible.

It seems obvious that the fewer bits per character we have, the less 
space this is all going to take up. If you have a plain ASCII file, just 
dropping every 8th bit already gives us 1/8 compression ( = 12.5%).

We can go one step further: Scan the whole file, and see which 
characters are *actually used* in this file. Codes 0 to 31 in ASCII are 
all "control codes", and with the exception of the newline character(s), 
they're all pretty much obsolete now. By not assigning any code at all 
to characters we aren't actually using, we *might* be able to get away 
with 6 bits per character. Maybe.

But even if we do, that's still "only" 25% compression. And if there are 
64 characters used then 6 bits of character will do, but if there are 65 
characters used then we have to have 7 bits - which is annoying.

But now, suppose we assign different code lengths to different 
characters. By assigning shorter codes to common letters and longer 
codes to rare letters, we can potentially store the same data in very 
much reduced space.

Suppose, for example, that we assign a 2-bit code to the letter E, and 
7-bit codes to everything else. Then every single time the letter E 
appears, we're saving 6 bits (compared to the original ASCII file). 
Since E probably appears quite a bit, that could add up to substantial 
savings. (Hint: The character " " probably appears *way more* than ANY 
actual letter of the alphabet, but hey.)

It might seem tempting to assign the shortest possible codes to 
everything. However, for X number of bits, there are only 2^X possible 
codes. If you make one code shorter, you have to make other codes 
correspondingly longer. So if you take a character that isn't *that* 
common and give it a really short code, you'll have to make all the 
other codes longer to compensate. Because the character isn't *that* 
common, the savings from the short code don't make up for the losses in 
all the other longer codes.

In summary, for every character there is a mathematically "perfect" code 
length, which will give best possible compression. And that length is 
equal to [the negation of] the binary logarithm of the probability of 
that character appearing:

   bits(X) = -log2 Pr(X)

If X takes more bits than this, you're wasting space. If X takes fewer 
bits than this, you're wasting bits somewhere else. So this is the 
"perfect" length of X.

(Note that if we were trying to encode characters with decimal codes, it 
would be log10 rather than log2. The 2 is because computers are binary.)

You can work out the "probability of X appearing" by simply counting how 
many X characters there are in the entire file, divided by the *total* 
number of characters. So that's easy enough.

Now, how to assign codes to characters? And how to figure out where one 
character ends and the next begins? In a normal (ASCII) text file, you 
know that each byte is one character. But if different characters take 
different numbers of bits, how are we to know where the start and end of 
each character code is?

The answer is to use a binary tree. For example:

       +
      / \
     +   C
    / \
   A   B

Starting from the top, go left for 0 and go right for 1. If we start at 
the top and go right, we immediately reach C. So C = 1. If we go left 
and then left, we get to A. So A = 00. And B = 01. So from this tree, we 
have

   A = 00
   B = 01
   C = 1

As we read bits from the input file, we can walk down the tree, going 
left when we read a 0 and right when we read a 1. When we get to a leaf, 
we spit out that character, and go back to the top of the tree. Easy. 
[But very inefficient. There are tricks to speed this up, though.]

So to build a code which can be unambiguously decoded, we just need to 
build a suitable code tree. How can we do that?

One way is to build a "Shannon tree". We count the characters in the 
file, and end up with a table like

   Char  Prob
    A      7%
    B     12%
    C      3%
   ...    ...

Now we need to build a tree. What we do is to try to split the set of 
all [used] characters into two sets who's total probability is as close 
to 50% as possible. All the items in the left subset get codes starting 
with a 0, all the codes in the right subset get codes starting with a 1. 
Now recurse. When a subset contains only one item, we have determined 
its code.

Trouble is, splitting a set into two subsets with equal sum... That's 
the Partition Problem, and it's NP-complete. Which is a technical way of 
saying that this is REALLY FREAKING HARD to do well. Worse, there may be 
more than one partitioning that yields roughly 50%, and which one you 
pick affects how easy subsequent partitionings are. This looks really, 
really hard...

Worse still, it can be proven that Shannon trees are sometimes 
unavoidably sub-optimal. So this algorithm is really hard to problem, 
really slow to execute, and doesn't even give optimal results. Great. :-/

The thing is, you don't need to do any of this. Because there's a way, 
way easier algorithm: You can build a Huffman tree instead.

Instead of starting with the set of all characters and trying to 
recursively split it to build a tree, we start with single characters 
and recursively join them to build a tree. The algorithm is:

   1. Start with a set of trivial 1-element trees containing just a 
character.
   2. Pick the two lowest-probability trees and remove them from the set.
   3. Combine the two trees, and insert the new tree back into the set.
   4. If the set has more than one tree, to go step 2.

It's difficult to come up with an example big enough to see what's 
happening, but small enough to not take thirty pages of printout. Let's 
try this:

   Char  Prob
    A      2%
    B      2%
    C      5%
    D     20%
    E     68%
    F      3%

Our initial set of trees is therefore:

    2%    2%    3%    5%    20%    68%
    A     B     F     C      D      E

The two lowest-probability trees are A (2%) and B (2%). 2% + 2% = 4%. So 
now we get:

    3%    4%    5%    20%    68%
    F     +     C      D      E
         / \
        A   B
        0   1

Now we join F to AB:

    5%    7%    20%    68%
    C     +      D      E
         / \
        F   +
        0  / \
          A   B
         10   11

Now we join C to that tree:

     12%    20%    68%
      +      D      E
     / \
    C   +
    0  / \
      F   +
     10  / \
        A   B
      110   111

Add on D:

       32%    68%
        +      E
       / \
      +   D
     / \  1
    C   +
   00  / \
      F   +
    010  / \
        A   B
     0110   0111

And finally, our Huffman tree:

        100%
          +
         / \
        +   E
       / \  1
      +   D
     / \  01
    C   +
  000  / \
      F   +
   0010  / \
        A   B
    00110   00111

So we have

   Char   Prob   Code    Bits   Log2(Prob)
    A      2%    00110   5      -5.643
    B      2%    00111   5      -5.643
    C      5%    000     3      -4.321
    D     20%    01      2      -2.321
    E     68%    1       1      -0.556
    F      3%    0010    4      -5.058

As you can see, the actual number of bits is quite close to -Log2(Prob). 
(Only C is a bit off.) If E really does make up 68% of the characters in 
the file, then by encoding it with just 1 bit, we're saving a huge 
amount of space.

The trouble with all this, of course, is that in order to *decode* the 
data, you need a copy of the Huffman tree used to encode it. So you have 
to store that somewhere. Fortunately, if you pick the codes in a 
suitable systematic way, it's possible to reconstruct the codes just by 
knowing how many bits were assigned to each character.



Alright, so that's nice and everything. But look at that table again:

   Char   Prob   Code    Bits   Log2(Prob)
    A      2%    00110   5      -5.643
    B      2%    00111   5      -5.643
    C      5%    000     3      -4.321
    D     20%    01      2      -2.321
    E     68%    1       1      -0.556
    F      3%    0010    4      -5.058

Ideally, we want to assign half a bit to E. Well, that's *obviously* 
impossible, right? I mean, a bit is a 0 or a 1. What the heck would half 
a bit look like?

It turns out you *can* actually do this. (!) Suppose we want to use 
exactly 1.5 bits per character. If you ask the encoder to encode 1 
character, then 2 bits come out. But if you ask it to encode 2 
characters, only 3 bits come out. Not 4, just 3. And this is what it 
means; the number of bits output is always a whole number (obviously), 
but it's the fractional number of bits, rounded up. (It's a bit like 
buying a box of 7400s priced at 21.225p each. Obviously there's no such 
thing as 0.225p; if you buy one chip, they change you 22p. But if you 
buy 1,000 chips...)

Huffman encoding can't do fractional bits. And there are some 
limitations on the combinations of bit-lengths you can get. (E.g., C 
should be 4 bits or 5 bits, but we've given it 3 bits.) But there is an 
algorithm that lacks these limitations: range coding.

I'm going to demonstrate this in decimal, but the principle in binary is 
just the same. We start off with an "active range", which begins with 
0.0000 to 0.9999:

   Max = 0.9999
   Min = 0.0000

We then split that range into several subranges, one for each possible 
character. The width of that range equals the probability of that 
character appearing next:

         0.9999
    F    ||||||
         0.9700
         0.9699
    E    ||||||
         0.2900
         0.2199
    D    ||||||
         0.0900
         0.0899
    C    ||||||
         0.0400
         0.0399
    B    ||||||
         0.0200
         0.0199
    A    ||||||
         0.0000

Suppose the first character to be encoded happens to be a C. Then that 
becomes the current range:

   Max = 0.0899
   Min = 0.0400

and we split this new subrange up as before:

         0.089999
    F    ||||||||
         0.085800
         0.085799
    E    ||||||||
         0.054500
         0.054499
    D    ||||||||
         0.044500
         0.044499
    C    ||||||||
         0.042000
         0.041999
    B    ||||||||
         0.041000
         0.040999
    A    ||||||||
         0.040000

This process continues for each and every character. Finally, at the end 
of the calculation, we have a really, really tiny range. For example:

   Max = 0.04487515156722040876510168073524657069870453400
   Min = 0.04487515156722040876510168073524657069870453399

We then output one of the two values (it doesn't matter which), together 
with some indication of how many characters are supposed to be decoded. 
(E.g., if the input was just lots of As, the min value would stay at 
zero forever, and we couldn't tell how many As to decode.)

This is great and all, but it looks like we need to calculate with 
*ridiculous* precision. If I compress a 2GB file and it becomes a 200MB 
file, that's 200MB for *one number*. We have to do all calculations with 
200MB of precision?! That doesn't sound fun...

Fortunately, this is not the case. As the range gets smaller and 
smaller, soon the leading digits all become the same. Look at the table 
above: every single code begins 0.0xxx. So we can output the 0.0 part, 
and shift it out of our calculations. The result is that when we get to 
here:

   Max = 0.448737
   Min = 0.448611

we can output "0.448" and change the numbers to

   Max = 73799999
   Min = 61100000

We can then continue our calculation as before. So we only need to hold 
in memory the handful of digits that are *actually changing*. To the 
left are digits we've already output into the compressed file. To the 
right, the max is 999 recurring, and the min is 000 recurring.

(There is an issue called "underflow", where the min gets stuck at 9999 
and the max gets stuck at 0000 and the digits don't meet before you run 
out of precision. You can work around that by "remembering" how many 0s 
and 9s there are, and cancelling them out temporarily. When the digits 
do align, you just need to remember to reanimate the cancelled digits. 
It's fiddly, but not hard.)

Notice that the smaller a range is, the quicker the leading digits start 
to match, and the sooner we write digits out to the compressed file. In 
other words, narrow ranges take more precision to specify. To summarise, 
rare symbols take more digits than common ones - just like with Huffman 
codes.



Huffman coding and range coding are both examples of "entropy coding". 
In either case, how much compression you get depends on:

   1. How much difference there is between the probabilities of the symbols.
   2. How accurately you can predict the probabilities.

#1 is a property of the file itself; you can't change that. Some files 
just compress better than others. #2, however, depends on how you 
estimate your symbol probabilities.

The simplest thing to do is to count how many times each character 
appears in the whole file, and divide by the total character count to 
get a probability for each character. That *works*, but the decoder 
doesn't have access to the uncompressed file to compute the statistics, 
which means you have to *tell* the decompressor what to do.

Fundamentally, there are two kinds of compression:

* In "static" compression, you scan the entire file, decide how to 
compress it, and then do the actual compression.

* In "dynamic" compression, you decide how to compress the next item 
based only on the data you've already processed so far.

Static compression has the advantage that you can "see" all the data, so 
you can make the best compression decisions. But the decompressor can't 
see that data, so you have to "tell" the decompressor what you decided 
to do (e.g., tell it what Huffman table you built), or it won't be able 
to decompress. "Telling" the decompressor takes up disk space, making 
the file larger and partially negating the potentially superior 
compression ratio.

Dynamic compression means you have limited information to work with, so 
you can't make compression decisions quite as well, resulting in a lower 
compression rate. On the other hand, the decompressor can "see" exactly 
what the compressor saw, and by running the same algorithm it can come 
to the same conclusion. In other words, you don't have to "tell" the 
decompressor what you did; it can work it out automatically from the 
compressed data itself. So you get lower compression rates, but there's 
no extra data tables to send, so there's less file overhead.

In the end, it's swings and roundabouts. But I will say this: some 
algorithms work better static, and some work better dynamic.

You can do dynamic Huffman compression. You can start estimating all 
characters with equal frequency, and AFTER compressing each character, 
update your frequency estimates and rebuild the whole Huffman tree. The 
decompressor can do the same, and hence recover the Huffman tree without 
needing any tables. However, rebuilding the Huffman tree takes time; 
more likely you'll update it every dozen characters or something, 
otherwise speed will suffer too much.

But in general, people usually do static compression when they use 
Huffman coding. It's quicker to build the Huffman tree once, and then 
send it once.

You can do static range coding. You calculate all your probabilities, 
write them into the header of the file, and then encode everything. 
It'll probably take a bit of space (it's not just bit-counts like 
Huffman, it's limited-precision probability values for each character), 
but you can do it.

The key advantage of range coding, though, is that it *costs nothing* to 
completely change all your probability estimates for every single 
character. So range coders are usually dynamic.

Whether Huffman or range coder, static compression of an English text 
ASCII file usually gets about 50% compression. That's good, but not 
fantastic. I just tried some random text files on my PC, and by zipping 
them I got around 80% compression. So how do we do that?

A key thing you can do is to use a higher-order model. U is a fairly 
uncommon letter. There aren't too many words with U in them. So the 
probability of seeing a U might seem fairly low. But if the previous 
letter happens to be Q... surely the probability that the next letter is 
a U ought to be *a lot* higher now!

If you calculate a fixed probability for each character, that is a 
"1st-order Markov model". If you calculate a different set of 
probabilities for each letter depending on what the previous letter was, 
that's a "2nd-order Markov model". And it can give substantially 
improved compression.

Of course, coming up with probabilities for 256 different byte values is 
one thing. Coming up with 256 tables of 256 probabilities, that's 65,536 
probabilities we need to compute... If each one takes 1 byte to store, 
that's 64KB just to store the probability tables! If you're trying to 
compress a 5KB text file, adding 64KB of overhead is going to give you 
some serious negative compression! (I.e., the "compressed" file will 
actually be *bigger* than the original.)

Suffice it to say, nobody does static 2nd-order Markov. It eats too much 
space. But dynamic? Sure. And *that* probably means you're using a range 
coder.

But why base your probabilities only on the *last* character? Why not 
the last two? Or three? Or then?

In fact, if you do this, you start hitting the problem that if you've 
never seen a certain sequence of characters before, how do you estimate 
probabilities following it?

The usual solution is "partial predictive matching" (PPM). The 
probability tables [of which there are a hell of a lot!] are sparse. If 
you haven't seen a certain 10-character sequence before, you ignore the 
oldest character and look up the corresponding 9-character sequence. If 
you haven't seen that, you try 8-character, until you finally get a 
match. In this way, for each character you can make the best possible 
probability estimate. And better probability estimates = better 
compression. (And because this is all dynamic, there's no need to store 
huge tables in the compressed file. The decompressor will run the same 
algorithm and compute the same estimates.)

Compressors such as PAQ use range coding and PPM (amount other things) 
to top the compression ratio ratings. (At the expense of speed and 
memory usage, it must be said.)



Actually, PAQ is interesting, since it uses context mixing. That is, 
half a dozen different algorithms (not just PPM) estimate the 
probabilities for the next character. It then uses a left-learning 
artificial neural network (!!) to combine these probability estimates to 
get a final estimate.

It does all of this PER BIT!

The result is an excruciatingly slow compressor that gives astonishing 
compression ratios.

Post a reply to this message

On 14/02/15 15:33, Orchid Win7 v1 wrote:
> OK, so it's been a while since I posted a big wall of text. So here goes...
> 

Aha! The real Andrew is back.

An interesting read - have you ever thought of taking up teaching?

John
-- 
Protect the Earth
It was not given to you by your parents
You hold it in trust for your children

Post a reply to this message

On 14/02/2015 09:10 PM, Doctor John wrote:
> Aha! The real Andrew is back.

Will the real Mathematical Orchid please stand up...

> An interesting read - have you ever thought of taking up teaching?

Erm... I've thought about it. But I've yet to see a class of people who 
actually *want* to learn stuff.

* Nobody at school wanted to learn a damn thing.
* The people at college were only there so they didn't have to get an 
actual job. None of them seemed that bothered if they get their 
qualification, never mind what grades they get.
* At university, everybody wanted the best grades, but nobody cared 
about the actual subject. They just wanted a nice piece of paper.
* I went on a Microsoft technology course. Everybody wanted the bit of 
paper, nobody really cared about the actual course content. Most of them 
were only there because their employer paid for it.

In short, I haven't come across a lesson where people are there because 
they actually *want* to learn something. Most of them are just seem to 
want the bit of paper that makes employers think they're awesome.



The fun part is when the *teacher* doesn't particularly want to teach. 
I'm sure I've told you about Mr Apathy, who I met at uni. I prefaced a 
lecture on binary by saying "of course you don't *really* need to know 
any of this. It's pointless really. I mean, the Java virtual machine 
handles all of this for you. But it's in the exam, so I'm supposed to 
teach it to you". Damn, I'm sold! Sign me up!!

Mind you, this was the same guy who spent ten minutes disinterestedly 
explaining how "pointless" the rebuilding of Colossus is. "I mean, it 
won't be the *actual* machine that they used. It's just a replica. And 
they don't make half the parts any more, so they'll have to change the 
parts and redesign the circuitry to compensate, and at the end of all 
that, it'll be so different from the original that there's just no point."

Not my favourite lecturer. :-P

Post a reply to this message

On 14/02/15 23:29, Orchid Win7 v1 wrote:
> On 14/02/2015 09:10 PM, Doctor John wrote:
>> Aha! The real Andrew is back.
> 
> Will the real Mathematical Orchid please stand up...
> 
: -)

> 
> Erm... I've thought about it. But I've yet to see a class of people who
> actually *want* to learn stuff.

<snip>


> In short, I haven't come across a lesson where people are there because
> they actually *want* to learn something. Most of them are just seem to
> want the bit of paper that makes employers think they're awesome.
> 

I was thinking about you going for a doctorate; you are as capable as
the Kat and I. All you need is an original project.

<snip>

> 
> Mind you, this was the same guy who spent ten minutes disinterestedly
> explaining how "pointless" the rebuilding of Colossus is. "I mean, it
> won't be the *actual* machine that they used. It's just a replica. And
> they don't make half the parts any more, so they'll have to change the
> parts and redesign the circuitry to compensate, and at the end of all
> that, it'll be so different from the original that there's just no point."
> 
> Not my favourite lecturer. :-P

Yeah. Come the revolution, he'll be the first against the wall. [I am
not joking]

John
-- 
Protect the Earth
It was not given to you by your parents
You hold it in trust for your children

Post a reply to this message

On 15-2-2015 0:29, Orchid Win7 v1 wrote:
> * Nobody at school wanted to learn a damn thing.
> * The people at college were only there so they didn't have to get an
> actual job. None of them seemed that bothered if they get their
> qualification, never mind what grades they get.
> * At university, everybody wanted the best grades, but nobody cared
> about the actual subject. They just wanted a nice piece of paper.
> * I went on a Microsoft technology course. Everybody wanted the bit of
> paper, nobody really cared about the actual course content. Most of them
> were only there because their employer paid for it.
>
> In short, I haven't come across a lesson where people are there because
> they actually *want* to learn something. Most of them are just seem to
> want the bit of paper that makes employers think they're awesome.
>


You are too pessimistic I believe (or the world has changed too much). 
Except for the last point (the Microsoft course) with which I agree, at 
all other levels I have been with people who genuinely wanted to /know/ 
and to /learn/. Not all of them of course but a substantial part, yes.

And I agree with John: you would be an excellent teacher.

-- 
Thomas

Post a reply to this message

>> In short, I haven't come across a lesson where people are there because
>> they actually *want* to learn something. Most of them are just seem to
>> want the bit of paper that makes employers think they're awesome.
>
> I was thinking about you going for a doctorate; you are as capable as
> the Kat and I. All you need is an original project.

My mum wanted me to do a PhD. I have no idea why. As best as I can tell 
from reading Stack Overflow, if you have a PhD then you are basically 
over-qualified for 99% of all employment opportunities ever, and you'll 
never be able to work outside academia ever again.

>> Not my favourite lecturer. :-P
>
> Yeah. Come the revolution, he'll be the first against the wall. [I am
> not joking]

Dude, if you hate your job so much, WHY ARE YOU STILL DOING IT?!?

Still, at least he could speak English. Many of my lecturers couldn't 
(or at least not very well). Seriously, if your JOB is to teach people 
who only speak English, not being able to speak English is surely an 
instant disqualification. No?!?

Post a reply to this message

On 15/02/2015 08:34 AM, Thomas de Groot wrote:
> You are too pessimistic I believe (or the world has changed too much).
> Except for the last point (the Microsoft course) with which I agree, at
> all other levels I have been with people who genuinely wanted to /know/
> and to /learn/. Not all of them of course but a substantial part, yes.

I haven't met them. Everybody I've met regards having technical 
knowledge as only something "nerds" and "losers" have - including the 
people I work with, who are all professional computer programmers. They 
all know how to write C#, yet they seem to think that understanding 
relational algebra or knowing how floating-point arithmetic works is 
only for lamers.

> And I agree with John: you would be an excellent teacher.

Again, I don't know. I'm good at writing long monologues that nobody 
will ever read; interacting with a room full of live humans is quite 
different. It requires a different skill set.

My Dad tells me [so I don't know how true this is] that Bach was a 
virtuoso organ player, but a HORRIBLE teacher. He just couldn't 
understand why nobody else could play like him...

Post a reply to this message

On 15/02/2015 10:33, Orchid Win7 v1 wrote:
>> I was thinking about you going for a doctorate; you are as capable as
>> the Kat and I. All you need is an original project.
>
> My mum wanted me to do a PhD. I have no idea why. As best as I can tell
> from reading Stack Overflow, if you have a PhD then you are basically
> over-qualified for 99% of all employment opportunities ever, and you'll
> never be able to work outside academia ever again.
>

Maybe Mother knows best. :-)

Going from your mathematical ramblings. I think that you would be 
eminently suitable to do a PhD at a good university. Not one of those 
who were ennobled, just to take the overflow from schools. When 
politicians decided it would keep them off the dole for another three years.


>>> Not my favourite lecturer. :-P
>>
>> Yeah. Come the revolution, he'll be the first against the wall. [I am
>> not joking]
>
> Dude, if you hate your job so much, WHY ARE YOU STILL DOING IT?!?
>

Yeah! Go on and tell him. I'm all ears. :-P

And while you are at it. Which revolution or counter revolution, would 
that be? ;-)


> Still, at least he could speak English. Many of my lecturers couldn't
> (or at least not very well). Seriously, if your JOB is to teach people
> who only speak English, not being able to speak English is surely an
> instant disqualification. No?!?

It would be at a a school. Universities are different, they employ 
lecturers for what they know and publish. Their world and ours touch but 
occasionally.


-- 

Regards
     Stephen

Post a reply to this message

On 15/02/2015 10:44, Orchid Win7 v1 wrote:
> On 15/02/2015 08:34 AM, Thomas de Groot wrote:
>> You are too pessimistic I believe (or the world has changed too much).
>> Except for the last point (the Microsoft course) with which I agree, at
>> all other levels I have been with people who genuinely wanted to /know/
>> and to /learn/. Not all of them of course but a substantial part, yes.
>
> I haven't met them. Everybody I've met regards having technical
> knowledge as only something "nerds" and "losers" have - including the
> people I work with, who are all professional computer programmers. They
> all know how to write C#, yet they seem to think that understanding
> relational algebra or knowing how floating-point arithmetic works is
> only for lamers.
>

That is because you have academic interests. Most professional computer 
programmers, I work with. Do it because it is a job. Not for love.


>> And I agree with John: you would be an excellent teacher.
>
> Again, I don't know. I'm good at writing long monologues that nobody
> will ever read; interacting with a room full of live humans is quite
> different. It requires a different skill set.
>

Unless you can use your competitive dancing skills to be in front of an 
audience.

> My Dad tells me [so I don't know how true this is] that Bach was a
> virtuoso organ player, but a HORRIBLE teacher. He just couldn't
> understand why nobody else could play like him...

And his throw away scales became the Brandenburg Concertos.

-- 

Regards
     Stephen

Post a reply to this message

>> My mum wanted me to do a PhD. I have no idea why. As best as I can tell
>> from reading Stack Overflow, if you have a PhD then you are basically
>> over-qualified for 99% of all employment opportunities ever, and you'll
>> never be able to work outside academia ever again.
>
> Maybe Mother knows best. :-)

I doubt it. :-P

As I say, it seems to just preclude you ever getting employed, ever again.

> Going from your mathematical ramblings. I think that you would be
> eminently suitable to do a PhD at a good university. Not one of those
> who were ennobled, just to take the overflow from schools. When
> politicians decided it would keep them off the dole for another three
> years.

Fun fact: They stopped giving out university grants just a year or two 
before I started university. I AM STILL PAYING FOR MY TUITION! More than 
a decade later... *sigh*

>> Still, at least he could speak English. Many of my lecturers couldn't
>> (or at least not very well). Seriously, if your JOB is to teach people
>> who only speak English, not being able to speak English is surely an
>> instant disqualification. No?!?
>
> It would be at a a school. Universities are different, they employ
> lecturers for what they know and publish. Their world and ours touch but
> occasionally.

Favourite anecdote: So we're in a Network Systems lecture, and the 
lecturer writes the Shannon channel coding theorem on the board. [That's 
the same Shannon who invented Shannon codes, so we're drifting back 
on-topic now...] The formula states that the maximum data-rate over a 
noisy channel is given by

   C = B log2 (1 + S/N)

And everybody is like "ZOMG! WTF is a logarithm?!" And the lecturer is 
like "oh, really? I'm sorry, I just assumed that THIRD-YEAR 
UNDERGRADUATES would already know that".

So it turns out that only two people in the room actually know what a 
logarithm is. Obviously one of them is me. And the other one... is the 
crazy beer-drinking rugby thug from Manchester. o_O And I'm serious; he 
actually knows what he's talking about. First and last time I ever heard 
him say something sensible...

Post a reply to this message

>> I haven't met them. Everybody I've met regards having technical
>> knowledge as only something "nerds" and "losers" have - including the
>> people I work with, who are all professional computer programmers. They
>> all know how to write C#, yet they seem to think that understanding
>> relational algebra or knowing how floating-point arithmetic works is
>> only for lamers.
>
> That is because you have academic interests. Most professional computer
> programmers, I work with. Do it because it is a job. Not for love.

This.

Everybody I met at university was like "OMG, this is SOOO boring! Just 
give me the piece of paper so I can start earning thousands of pounds 
per hour as a top London computing consultant."

Everybody seemed to think that computing is *excruciatingly* boring, but 
also that there's billions of pounds in it. Like, if you have any 
computer skills at all, you're going to be a millionaire. (Still, it 
*was* right around the height of the dot-com bubble I suppose...)

>>> And I agree with John: you would be an excellent teacher.
>>
>> Again, I don't know. I'm good at writing long monologues that nobody
>> will ever read; interacting with a room full of live humans is quite
>> different. It requires a different skill set.
>
> Unless you can use your competitive dancing skills to be in front of an
> audience.

Uh, I do competitive dancing; what makes you think I have *skills*? :-P

Unless you mean I give the lecture while dancing the quickstep. ^_-

>> My Dad tells me [so I don't know how true this is] that Bach was a
>> virtuoso organ player, but a HORRIBLE teacher. He just couldn't
>> understand why nobody else could play like him...
>
> And his throw away scales became the Brandenburg Concertos.

Hey, I think I have the score for that somewhere...

Post a reply to this message

On 15/02/2015 11:17, Orchid Win7 v1 wrote:
>>> My mum wanted me to do a PhD. I have no idea why. As best as I can tell
>>> from reading Stack Overflow, if you have a PhD then you are basically
>>> over-qualified for 99% of all employment opportunities ever, and you'll
>>> never be able to work outside academia ever again.
>>
>> Maybe Mother knows best. :-)
>
> I doubt it. :-P
>
> As I say, it seems to just preclude you ever getting employed, ever again.
>

For Arts graduates maybe. I have met so many PhD doctors at work. I only 
look up to Misters. ;-)
Having a higher degree in maths type subjects would open doors to 
better/more interesting jobs.

>> Going from your mathematical ramblings. I think that you would be
>> eminently suitable to do a PhD at a good university. Not one of those
>> who were ennobled, just to take the overflow from schools. When
>> politicians decided it would keep them off the dole for another three
>> years.
>
> Fun fact: They stopped giving out university grants just a year or two
> before I started university. I AM STILL PAYING FOR MY TUITION! More than
> a decade later... *sigh*
>

Scandalous!


>
> Favourite anecdote:


It only saddens me it does not surprise me.


-- 

Regards
     Stephen

Post a reply to this message

On 15-2-2015 12:24, Orchid Win7 v1 wrote:
>>> I haven't met them. Everybody I've met regards having technical
>>> knowledge as only something "nerds" and "losers" have - including the
>>> people I work with, who are all professional computer programmers. They
>>> all know how to write C#, yet they seem to think that understanding
>>> relational algebra or knowing how floating-point arithmetic works is
>>> only for lamers.
>>
>> That is because you have academic interests. Most professional computer
>> programmers, I work with. Do it because it is a job. Not for love.
>
> This.
>
> Everybody I met at university was like "OMG, this is SOOO boring! Just
> give me the piece of paper so I can start earning thousands of pounds
> per hour as a top London computing consultant."
>
> Everybody seemed to think that computing is *excruciatingly* boring, but
> also that there's billions of pounds in it. Like, if you have any
> computer skills at all, you're going to be a millionaire. (Still, it
> *was* right around the height of the dot-com bubble I suppose...)

I am not a computer professional so that world is indeed foreign to me, 
but what you and Stephen are saying truly appals me... But then I may be 
from an alien planet: I have never been /really/ interested in money... :-)

>
>>>> And I agree with John: you would be an excellent teacher.
>>>
>>> Again, I don't know. I'm good at writing long monologues that nobody
>>> will ever read; interacting with a room full of live humans is quite
>>> different. It requires a different skill set.

That last is indeed true but can be learned with practice (like 
dancing). Writing skills, which you have already, is more difficult to 
learn. I must say that I enjoy your long monologues btw.

>>
>> Unless you can use your competitive dancing skills to be in front of an
>> audience.
>
> Uh, I do competitive dancing; what makes you think I have *skills*? :-P
>
> Unless you mean I give the lecture while dancing the quickstep. ^_-

Seriously, in a good lecture before an audience, some kind of 
choreography (even body language) is of paramount importance. Whatever 
the subject.

>
>>> My Dad tells me [so I don't know how true this is] that Bach was a
>>> virtuoso organ player, but a HORRIBLE teacher. He just couldn't
>>> understand why nobody else could play like him...
>>
>> And his throw away scales became the Brandenburg Concertos.
>
> Hey, I think I have the score for that somewhere...

Great stuff indeed.

-- 
Thomas

Post a reply to this message

On 15/02/2015 11:16 AM, Stephen wrote:
> That is because you have academic interests. Most professional computer
> programmers, I work with. Do it because it is a job. Not for love.

Having said that... The company I work for employs a grant total of 9 
people. The company has sold less than 100 units in its entire 10-year 
history. I cannot *imagine* that anybody here is particularly well-paid.

And yet, if you look through source control, you will find the 
occasional commit performed at 2AM. Nobody gets *paid* to do that. 
People occasionally choose to do that out of... well, something that 
isn't money anyway. (And no, it's not deadline pressure; we don't have 
deadlines. Or pressure of any kind, really.)

Still, while my colleagues genuinely want to make the software the best 
it can be (and not just because they get paid to do that), I am 
apparently the only person in the building who thinks that computer 
science is a worthwhile endeavour. Everybody else just hits Google until 
the problem goes away, without bothering to understand why. I find that 
a little disturbing.

But not as disturbing as me getting paid to do C++ when I know nothing 
about it! o_O

Post a reply to this message

On 15-2-2015 13:14, Stephen wrote:
> On 15/02/2015 11:17, Orchid Win7 v1 wrote:
>>>> My mum wanted me to do a PhD. I have no idea why. As best as I can tell
>>>> from reading Stack Overflow, if you have a PhD then you are basically
>>>> over-qualified for 99% of all employment opportunities ever, and you'll
>>>> never be able to work outside academia ever again.
>>>
>>> Maybe Mother knows best. :-)
>>
>> I doubt it. :-P
>>
>> As I say, it seems to just preclude you ever getting employed, ever
>> again.
>>
>
> For Arts graduates maybe. I have met so many PhD doctors at work. I only
> look up to Misters. ;-)
> Having a higher degree in maths type subjects would open doors to
> better/more interesting jobs.

I second that.

>
>>> Going from your mathematical ramblings. I think that you would be
>>> eminently suitable to do a PhD at a good university. Not one of those
>>> who were ennobled, just to take the overflow from schools. When
>>> politicians decided it would keep them off the dole for another three
>>> years.
>>
>> Fun fact: They stopped giving out university grants just a year or two
>> before I started university. I AM STILL PAYING FOR MY TUITION! More than
>> a decade later... *sigh*
>>
>
> Scandalous!
>
>
>>
>> Favourite anecdote:
>
>
> It only saddens me it does not surprise me.
>
Sadly, I second that too.

-- 
Thomas

Post a reply to this message

On 15/02/2015 11:24, Orchid Win7 v1 wrote:

> This.
>
> Everybody I met at university was like "OMG, this is SOOO boring! Just
> give me the piece of paper so I can start earning thousands of pounds
> per hour as a top London computing consultant."
>
> Everybody seemed to think that computing is *excruciatingly* boring, but
> also that there's billions of pounds in it. Like, if you have any
> computer skills at all, you're going to be a millionaire. (Still, it
> *was* right around the height of the dot-com bubble I suppose...)
>

"They" forgot to mention that the high pay would drop when the demand 
was filled by themselves.



>> Unless you can use your competitive dancing skills to be in front of an
>> audience.
>
> Uh, I do competitive dancing; what makes you think I have *skills*? :-P
>


Performance skills before an audience. I meant.
I did a job a couple of years ago where I had to deliver training 
courses. I had never done anything like it before. When I got my sea 
legs. It was SHOWTIME!!!
I really enjoyed it. But somehow I don't see you being comfortable in 
front of an audience, yet.

> Unless you mean I give the lecture while dancing the quickstep. ^_-
>

Tap would be more appropriate. IMO. ;-)

>>> My Dad tells me [so I don't know how true this is] that Bach was a
>>> virtuoso organ player, but a HORRIBLE teacher. He just couldn't
>>> understand why nobody else could play like him...
>>
>> And his throw away scales became the Brandenburg Concertos.
>
> Hey, I think I have the score for that somewhere...

The story goes that he gave them to his patron to play with his amateur 
orchestra.


-- 

Regards
     Stephen

Post a reply to this message

On 15/02/2015 12:26, Thomas de Groot wrote:
>
> I am not a computer professional so that world is indeed foreign to me,
> but what you and Stephen are saying truly appals me... But then I may be
> from an alien planet: I have never been /really/ interested in money... :-)
>

 From a Dutchman! We will keep that quiet. ;-)
During my time in Aberdeen I met a lot of geologists who only studded it 
to get a job in the Oil Industry.


>> Unless you mean I give the lecture while dancing the quickstep. ^_-
>
> Seriously, in a good lecture before an audience, some kind of
> choreography (even body language) is of paramount importance. Whatever
> the subject.
>

Good body awareness is important. You don't want to knock your coffee 
over your notes.

-- 

Regards
     Stephen

Post a reply to this message

On 15/02/2015 12:26 PM, Thomas de Groot wrote:

> I am not a computer professional so that world is indeed foreign to me,
> but what you and Stephen are saying truly appals me... But then I may be
> from an alien planet: I have never been /really/ interested in money... :-)

There is something deeply worrying about... OK, so let's try this. Come 
on this journey with me.

You see a job advert for an English / French translator. The advert says 
there will be a test during the interview. You don't know a single word 
of French, but you think to yourself "bah, nobody will notice. I'll just 
Google search of the words when I need to. It'll be fine."

WHAT. THE. HELL.

I'm sorry, IN WHICH UNIVERSE IS NOBODY GOING TO NOTICE?! What the HELL 
are you smoking?!!

Nobody thinks this will work. Nobody ever tries this. Because nobody is 
that stupid. But replace being able to write French with being able to 
write C#, and suddenly everybody thinks that somehow "nobody will 
notice". Wuh?!

I went to one interview where there was an SQL test. The job advert says 
that SQL knowledge is required. The job advert says there will be a 
test. There is no time limit on the test. The test is OPEN BOOK; you 
have access to the Microsoft SQL Server help file [but not the 
Internet]. You have a computer in front of you to test your answers with.

According to the interviewer, people turned up with "20 years' SQL 
experience" and failed to answer a single question.

That was the moment when I realised something: People LIE on their CV.

I mean, sure, I know people *exaggerate* their capabilities. But this 
revelation left me stunned. Anybody who claims to have "20 years' SQL 
experience" and yet cannot figure out how to select all customers from 
the customer table is LYING. This is not an exaggeration for dramatic 
effect, this is outright FRAUD. Strictly speaking, this is ILLEGAL.

I used to be worried about overstating my abilities on my CV, lest I 
give people the impression that I have capabilities that I don't. Now I 
only worry about how to tell people that I'm not just MAKING CRAP UP! 
Some people's CVs clearly have no connection to reality. How do you 
prove to people trying to hire you that you're not just talking nonsense?

Also: FizzBuzz exists. Because, from what I've seen, 99% of all the 
people you interview have NO IDEA how to write software. I don't mean 
they're not very good at it, I mean... Imagine hiring a lorry driver, 
and finding out he doesn't know which pedal is the brake and which one 
is the accelerator. This person CLEARLY has never sat behind the 
controls of a vehicle - no matter WHAT his CV claims.

We started doing phone screenings rather than waste our time with 
hopeless people. We find a website where you can watch the other person 
type, in real-time, as they type it. Almost everybody who reaches this 
stage doesn't get to have a face-to-face interview. A few people got 
creative though:

* While talking to one guy, we could *clearly* hear somebody else 
sitting next to him telling him what to type. Er, REJECT!

* Another guy, we asked him to write some code. We can hear typey typey 
typey. Er, can you type into the window please? Typey, typey, typey, 
BAM! Suddenly a huge block of text appears all at once. In other 
words... you just Googled the answer and copy-pasted it, didn't you? REJECT!

Seriously, how the **** do people think we're not gonna notice this 
stuff? Let us suppose, for argument's sake, that you somehow trick us 
into interviewing you... Do you REALLY THINK it won't instantly become 
EXTREMELY OBVIOUS that you're bluffing? How do you actually see this 
playing out? Are you mental?!

I wonder... Does anybody have this much trouble hiring a carpenter?

>>>>> And I agree with John: you would be an excellent teacher.
>>>>
>>>> Again, I don't know. I'm good at writing long monologues that nobody
>>>> will ever read; interacting with a room full of live humans is quite
>>>> different. It requires a different skill set.
>
> That last is indeed true but can be learned with practice (like
> dancing). Writing skills, which you have already, is more difficult to
> learn.

I guess that just leaves us with the problem of nobody wanting to learn 
then. ;-)

> I must say that I enjoy your long monologues btw.

Oh, well, I'm glad somebody does. Sometimes it does feel like I'm 
talking to myself...

> Seriously, in a good lecture before an audience, some kind of
> choreography (even body language) is of paramount importance. Whatever
> the subject.

I'm sure if I give a lecture while doing the Charleston to some 
authentic music, everybody will remember me. No idea if they'll remember 
a damned thing I said, but... ^_^

Post a reply to this message

>> As I say, it seems to just preclude you ever getting employed, ever
>> again.
>
> For Arts graduates maybe. I have met so many PhD doctors at work. I only
> look up to Misters. ;-)
> Having a higher degree in maths type subjects would open doors to
> better/more interesting jobs.

According to the Academia Stack Exchange portal, it seems if you have a 
PhD, everybody immediately assumes you're going to be hellishly 
expensive to hire and summarily drops you from consideration.

Unless you want to work in the finance industry, which only exists in 
London.

>> Fun fact: They stopped giving out university grants just a year or two
>> before I started university. I AM STILL PAYING FOR MY TUITION! More than
>> a decade later... *sigh*
>>
>
> Scandalous!

Hey, I spent 10 years earning minimum wage, so...

>> Favourite anecdote:
>
> It only saddens me it does not surprise me.

In fairness, while he only said a few sentences, it was immediately 
clear to me that he knew what he was talking about. As in, logarithms 
aren't just something he heard about once, he actually *understood* 
them. (As opposed to merely rote-memorising some facts about them.)

Post a reply to this message

Am 14.02.2015 um 16:33 schrieb Orchid Win7 v1:
> OK, so it's been a while since I posted a big wall of text. So here goes...

I wouldn't normally respond to this posting of yours, but just to let 
you know how much interest there is in reading such posts, I hereby 
declare that I did read it with interest. Huffman coding wasn't new to 
me, so I skipped over that part, and I'm also somewhat familiar with 
Markov chains including PPM, but range encoding is an interesting 
concept I hadn't heard of before. Also, I do start to notice that human 
memory seems to be made up of DRAM cells, so a refresh cycle now and 
then is most welcome ;)

Post a reply to this message

On 15/02/2015 01:12 PM, clipka wrote:
> Am 14.02.2015 um 16:33 schrieb Orchid Win7 v1:
>> OK, so it's been a while since I posted a big wall of text. So here
>> goes...
>
> I wouldn't normally respond to this posting of yours, but just to let
> you know how much interest there is in reading such posts, I hereby
> declare that I did read it with interest.

Hey, cool!

> Huffman coding wasn't new to
> me, so I skipped over that part, and I'm also somewhat familiar with
> Markov chains including PPM, but range encoding is an interesting
> concept I hadn't heard of before.

Yeah, it's interesting. I'm slightly fuzzy on exactly what the 
difference between "arithmetic coding" and "range coding" is. Wikipedia 
seems to think there's a difference, but I don't know what it is.

(Then again, I only just figured out the difference between LZ78 and LZW...)

> Also, I do start to notice that human
> memory seems to be made up of DRAM cells, so a refresh cycle now and
> then is most welcome ;)

More like, it's a neural network which "learns" by adjusting the 
connection weights between cells - but *that* is the subject for another 
wall of text, I imagine... ;-)

Post a reply to this message

On 15/02/2015 12:57, Orchid Win7 v1 wrote:

>
> According to the Academia Stack Exchange portal, it seems if you have a
> PhD, everybody immediately assumes you're going to be hellishly
> expensive to hire and summarily drops you from consideration.
>

That does not fit in with my experience. True often PhD's end up at the 
top of the pile but lots of them do lesser jobs. They just have a PhD. 
It does not rule them out from doing jobs they are "over qualified" for.


> Unless you want to work in the finance industry, which only exists in
> London.
>

I am beginning to get the feeling that you are a member of a sect that 
vilifies London as a pit of depravity.


>>> Fun fact: They stopped giving out university grants just a year or two
>>> before I started university. I AM STILL PAYING FOR MY TUITION! More than
>>> a decade later... *sigh*
>>>
>>
>> Scandalous!
>
> Hey, I spent 10 years earning minimum wage, so...

Even more scandalous.

>
>>> Favourite anecdote:
>>
>> It only saddens me it does not surprise me.
>
> In fairness, while he only said a few sentences, it was immediately
> clear to me that he knew what he was talking about. As in, logarithms
> aren't just something he heard about once, he actually *understood*
> them. (As opposed to merely rote-memorising some facts about them.)


-- 

Regards
     Stephen

Post a reply to this message

Orchid Win7 v1 <voi### [at] devnull> wrote:
> I haven't met them. Everybody I've met regards having technical
> knowledge as only something "nerds" and "losers" have - including the
> people I work with, who are all professional computer programmers. They
> all know how to write C#, yet they seem to think that understanding
> relational algebra or knowing how floating-point arithmetic works is
> only for lamers.

they are useful idiots


> My Dad tells me [so I don't know how true this is] that Bach was a
> virtuoso organ player, but a HORRIBLE teacher. He just couldn't
> understand why nobody else could play like him...

Bach (and many other bright composers from the highly specialized and complex
instrumental music of that era) was much like as autistic as we are.  Hence, his
many bouts of fury against his patrons, arrests, etc.  He simply didn't know how
to harmoniously deal with other human beings as he did with an organ.  He did
teach his children pretty well though.  His oldest 2 sons were as capable as
himself as performer and just had a different composing style because times had
changed...

Post a reply to this message

On Sun, 15 Feb 2015 13:17:05 +0200, Orchid Win7 v1 <voi### [at] devnull> wrote:

>    C = B log2 (1 + S/N)
>
> And everybody is like "ZOMG! WTF is a logarithm?!" And the lecturer is  
> like "oh, really? I'm sorry, I just assumed that THIRD-YEAR  
> UNDERGRADUATES would already know that".
>
> So it turns out that only two people in the room actually know what a  
> logarithm is. Obviously one of them is me. And the other one... is the  
> crazy beer-drinking rugby thug from Manchester. o_O And I'm serious; he  
> actually knows what he's talking about. First and last time I ever heard  
> him say something sensible...

I don't get it. We learned about logarithms in high school.

-- 
-Nekar Xenos-

Post a reply to this message

On 15-2-2015 13:53, Orchid Win7 v1 wrote:
> On 15/02/2015 12:26 PM, Thomas de Groot wrote:
>
>> I am not a computer professional so that world is indeed foreign to me,
>> but what you and Stephen are saying truly appals me... But then I may be
>> from an alien planet: I have never been /really/ interested in
>> money... :-)
>
> There is something deeply worrying about... OK, so let's try this. Come
> on this journey with me.

[snip]

As I said, I am truly appalled about what you tell. It is unbelievable 
but I suppose I am naive... ;-)

>
> I wonder... Does anybody have this much trouble hiring a carpenter?

No, but I am starting to worry.

>
>>>>>> And I agree with John: you would be an excellent teacher.
>>>>>
>>>>> Again, I don't know. I'm good at writing long monologues that nobody
>>>>> will ever read; interacting with a room full of live humans is quite
>>>>> different. It requires a different skill set.
>>
>> That last is indeed true but can be learned with practice (like
>> dancing). Writing skills, which you have already, is more difficult to
>> learn.
>
> I guess that just leaves us with the problem of nobody wanting to learn
> then. ;-)

Sad, sad, sad. I believe learning is one of those rare incentives that 
make life worthwhile.

Oh dear, some would say /money/ of course ;-)

>
>> I must say that I enjoy your long monologues btw.
>
> Oh, well, I'm glad somebody does. Sometimes it does feel like I'm
> talking to myself...
>
>> Seriously, in a good lecture before an audience, some kind of
>> choreography (even body language) is of paramount importance. Whatever
>> the subject.
>
> I'm sure if I give a lecture while doing the Charleston to some
> authentic music, everybody will remember me. No idea if they'll remember
> a damned thing I said, but... ^_^

Right. You can overdo it. Watching lecturers is a great school for do's 
and dont's. I have seen hilarious examples of dont's... ;-)

-- 
Thomas

Post a reply to this message

On 15-2-2015 13:53, Stephen wrote:
> On 15/02/2015 12:26, Thomas de Groot wrote:
>>
>> I am not a computer professional so that world is indeed foreign to me,
>> but what you and Stephen are saying truly appals me... But then I may be
>> from an alien planet: I have never been /really/ interested in
>> money... :-)
>>
>
>  From a Dutchman! We will keep that quiet. ;-)

I am afraid I am already ostracised ;-)

> During my time in Aberdeen I met a lot of geologists who only studded it
> to get a job in the Oil Industry.

Oh yes, there are those. To tell the truth, back in the sixties in NL at 
least, those where mostly concentrated in another university than the 
one I was in (Amsterdam) which was considered more 'leftist' and less 
oil-minded :-)

>
>
>>> Unless you mean I give the lecture while dancing the quickstep. ^_-
>>
>> Seriously, in a good lecture before an audience, some kind of
>> choreography (even body language) is of paramount importance. Whatever
>> the subject.
>>
>
> Good body awareness is important. You don't want to knock your coffee
> over your notes.
>
Oh dear, no! Nor drop all your overhead sheets on the ground as I saw 
happen :-)

-- 
Thomas

Post a reply to this message

On 15/02/2015 12:53, Orchid Win7 v1 wrote:
>
> I wonder... Does anybody have this much trouble hiring a carpenter?

Not in fundamentalist America. There you can hire one to support every 
belief that you want. ;-)

-- 

Regards
     Stephen

Post a reply to this message

On 15/02/2015 15:22, Thomas de Groot wrote:
> On 15-2-2015 13:53, Stephen wrote:

>> During my time in Aberdeen I met a lot of geologists who only studded it
>> to get a job in the Oil Industry.
>
> Oh yes, there are those. To tell the truth, back in the sixties in NL at
> least, those where mostly concentrated in another university than the
> one I was in (Amsterdam) which was considered more 'leftist' and less
> oil-minded :-)
>

Some universities can become specialised in certain fields, supplying a 
market.


>>
>> Good body awareness is important. You don't want to knock your coffee
>> over your notes.
>>
> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
> happen :-)
>

You are giving your age away, now. ;-)
Those sheets seemed to have no friction between them at all.
Or so someone told me. :-)

-- 

Regards
     Stephen

Post a reply to this message

On 15/02/2015 02:56 PM, Nekar Xenos wrote:
> I don't get it. We learned about logarithms in high school.

In which country?

For that matter, we learned about fractions in school. But I've yet to 
meet anyone from my own generation who has the slightest clue how 
fractions work... [I suppose that's another one of those "nobody cares; 
the computer does it for you" type things.]

Post a reply to this message

>> Good body awareness is important. You don't want to knock your coffee
>> over your notes.
>>
> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
> happen :-)

You must never do a tango with an Eskimo. No no no, oh dear no!

[I still don't actually know why...]

Post a reply to this message

On 15/02/2015 02:42 PM, nemesis wrote:
> they are useful idiots

Why does that sound like a band name...

> Bach (and many other bright composers from the highly specialized and complex
> instrumental music of that era) was much like as autistic as we are.  Hence, his
> many bouts of fury against his patrons, arrests, etc.  He simply didn't know how
> to harmoniously deal with other human beings as he did with an organ.  He did
> teach his children pretty well though.  His oldest 2 sons were as capable as
> himself as performer and just had a different composing style because times had
> changed...

Harmoniously.

I SEE WHAT YOU DID THERE! ;-)

Post a reply to this message

>> There is something deeply worrying about... OK, so let's try this. Come
>> on this journey with me.
>
> [snip]
>
> As I said, I am truly appalled about what you tell. It is unbelievable
> but I suppose I am naive... ;-)

I sincerely hope this is unique to the computing field.

I *think* it's because what we programmers produce is so intangible. 
Like, if you hire somebody to build an office block, and they bash a few 
bits of wood into the ground in the shape of a rectangle, you tell them 
they're fired, and then you sue them into oblivion. [Which I gather is 
some kind of MMO?] But if you commission somebody to build a payroll 
system, you can't really tell that it's just an MS Access skin - unless 
you're enough of a computer expert that you don't need to hire an expert 
in the first place!

That, I *think*, is maybe what the problem is. Although I do wonder how 
nobody gets electrocuted because the "electrician" they hired doesn't 
actually know how electricity works...

>> I wonder... Does anybody have this much trouble hiring a carpenter?
>
> No, but I am starting to worry.

It does make you wonder, doesn't it?

>> I guess that just leaves us with the problem of nobody wanting to learn
>> then. ;-)
>
> Sad, sad, sad. I believe learning is one of those rare incentives that
> make life worthwhile.

I find all this stuff fascinating - but then again, I'm weird, so...

But still, there *was* a time when the great mathematicians, physicists 
and engineers of the day were lionised as visionary idols... Now we have 
Pop Idol. *sigh*

>> I'm sure if I give a lecture while doing the Charleston to some
>> authentic music, everybody will remember me. No idea if they'll remember
>> a damned thing I said, but... ^_^
>
> Right. You can overdo it. Watching lecturers is a great school for do's
> and dont's. I have seen hilarious examples of dont's... ;-)

"The Charleston lectures" would make a hell of an Internet meme though...

...you know, I mean if I could actually *speak* while doing the 
Charleston, instead of gasping for air...

Post a reply to this message

On 15/02/2015 20:36, Orchid Win7 v1 wrote:
>>> Good body awareness is important. You don't want to knock your coffee
>>> over your notes.
>>>
>> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
>> happen :-)
>
> You must never do a tango with an Eskimo. No no no, oh dear no!
>
> [I still don't actually know why...]

If you do, you'll get the breeze up
And you'll end up with a freeze up

[And you would not want her to break it off.]

-- 

Regards
     Stephen

Post a reply to this message

Orchid Win7 v1 <voi### [at] devnull> wrote:
> Still, at least he could speak English. Many of my lecturers couldn't
> (or at least not very well). Seriously, if your JOB is to teach people
> who only speak English, not being able to speak English is surely an
> instant disqualification. No?!?

are you sure people in the UK still speak english?

Post a reply to this message

On 15-2-2015 22:16, Stephen wrote:
> On 15/02/2015 20:36, Orchid Win7 v1 wrote:
>>>> Good body awareness is important. You don't want to knock your coffee
>>>> over your notes.
>>>>
>>> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
>>> happen :-)
>>
>> You must never do a tango with an Eskimo. No no no, oh dear no!
>>
>> [I still don't actually know why...]
>
> If you do, you'll get the breeze up
> And you'll end up with a freeze up
>
> [And you would not want her to break it off.]
>
ROTFL

-- 
Thomas

Post a reply to this message

On 15-2-2015 16:38, Stephen wrote:
> On 15/02/2015 15:22, Thomas de Groot wrote:
>> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
>> happen :-)
>>
>
> You are giving your age away, now. ;-)
> Those sheets seemed to have no friction between them at all.
> Or so someone told me. :-)
>

I know. Overhead sheets are only found in museums nowadays.

No friction? well yes but often to the contrary. The damn things often 
stuck together like hell with static electricity! But as you could put 
them in individual holders they became a manageable pile.

-- 
Thomas

Post a reply to this message

On 15-2-2015 21:45, Orchid Win7 v1 wrote:
>>> I wonder... Does anybody have this much trouble hiring a carpenter?
>>
>> No, but I am starting to worry.
>
> It does make you wonder, doesn't it?

It does, and with the trash sent over from China it is often a nightmare 
come true. Nowadays, I often control /where/ devices are put together, 
even the labels with a traditional reputation. :-(

[snip]

>
>>> I'm sure if I give a lecture while doing the Charleston to some
>>> authentic music, everybody will remember me. No idea if they'll remember
>>> a damned thing I said, but... ^_^
>>
>> Right. You can overdo it. Watching lecturers is a great school for do's
>> and dont's. I have seen hilarious examples of dont's... ;-)
>
> "The Charleston lectures" would make a hell of an Internet meme though...
>
> ...you know, I mean if I could actually *speak* while doing the
> Charleston, instead of gasping for air...

LOL That would be a great show indeed!

-- 
Thomas

Post a reply to this message

>> You are giving your age away, now. ;-)
>> Those sheets seemed to have no friction between them at all.
>> Or so someone told me. :-)
>
> I know. Overhead sheets are only found in museums nowadays.

Oh, I don't know. When I was at uni, every time assignments were due, 
you could *guarantee* that some moron will have put overhead sheets 
through the library laser printer - the WRONG overhead sheets. The ones 
that melt in the fuser unit. :-S

Post a reply to this message

> According to the Academia Stack Exchange portal, it seems if you have a
> PhD, everybody immediately assumes you're going to be hellishly
> expensive to hire and summarily drops you from consideration.
>
> Unless you want to work in the finance industry, which only exists in
> London.

I work for a manufacturing company about 50 miles away from London, and 
we have quite a few PhDs working in our R&D team (physics and chemistry) 
and in more senior positions elesewhere in the company. We certainly 
wouldn't reject your application just for having a PhD.

My previous employer was much more R&D oriented and at least 50% of the 
staff had PhDs. Didn't you have some at your previous job too? Maybe the 
problem you highlight is unique to people with computer science PhDs?

Post a reply to this message

>> In short, I haven't come across a lesson where people are there because
>> they actually *want* to learn something. Most of them are just seem to
>> want the bit of paper that makes employers think they're awesome.
>>
>
>
> You are too pessimistic I believe (or the world has changed too much).
> Except for the last point (the Microsoft course) with which I agree, at
> all other levels I have been with people who genuinely wanted to /know/
> and to /learn/. Not all of them of course but a substantial part, yes.

I have to agree with Andy here, my experience of school (compulsory up 
to age 16) and sixth form college (optional up to age 18) is that most 
people didn't want to learn. Obviously there were some who did, but the 
majority you felt were there because someone else had told them to be 
there but they didn't want to be. It was a place to meet people and hang 
out, having to go to lessons and do homework was an inconvenience.

It was only when I got to University that it was suddenly full of people 
who actually wanted to be there and wanted to learn. But I think there 
are likely only a few Universities like that with that atmosphere, 
speaking to friends elsewhere it sounded like other Universities were 
just like school and college where people would spend more effort trying 
to avoid doing any work that actually learning anything.

Post a reply to this message

On 16/02/2015 08:09, Thomas de Groot wrote:
> On 15-2-2015 16:38, Stephen wrote:
>> On 15/02/2015 15:22, Thomas de Groot wrote:
>>> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
>>> happen :-)
>>>
>>
>> You are giving your age away, now. ;-)
>> Those sheets seemed to have no friction between them at all.
>> Or so someone told me. :-)
>>
>
> I know. Overhead sheets are only found in museums nowadays.
>
> No friction? well yes but often to the contrary. The damn things often
> stuck together like hell with static electricity! But as you could put
> them in individual holders they became a manageable pile.
>

Now we are getting back to the binary nature of life and Sods Law.

-- 

Regards
     Stephen

Post a reply to this message

On 16/02/2015 08:03, Thomas de Groot wrote:
> On 15-2-2015 22:16, Stephen wrote:

>>
> ROTFL
>

The old ones are the best ones. :-)

-- 

Regards
     Stephen

Post a reply to this message

On 16-2-2015 9:52, Stephen wrote:
> On 16/02/2015 08:09, Thomas de Groot wrote:
>> On 15-2-2015 16:38, Stephen wrote:
>>> On 15/02/2015 15:22, Thomas de Groot wrote:
>>>> Oh dear, no! Nor drop all your overhead sheets on the ground as I saw
>>>> happen :-)
>>>>
>>>
>>> You are giving your age away, now. ;-)
>>> Those sheets seemed to have no friction between them at all.
>>> Or so someone told me. :-)
>>>
>>
>> I know. Overhead sheets are only found in museums nowadays.
>>
>> No friction? well yes but often to the contrary. The damn things often
>> stuck together like hell with static electricity! But as you could put
>> them in individual holders they became a manageable pile.
>>
>
> Now we are getting back to the binary nature of life and Sods Law.
>

Beware of the /binary/ nature though. I must have told this story many 
times, maybe even here...

Settings: Split, in then still existing Yugoslavia; a lecture hall 
during a conference with a huge stage in front of the audience.

Characters: (1) Somebody high up in party hierarchy going to give a 
presentation on mineral resources; (2) his secretary.

Scene 1: The session's president introduces the next presentation 
(applause); the presenter comes up on stage with his secretary; 
presenter takes place behind the lectern; the secretary sits down 
besides the overhead projector, facing us, a little nervously, the 
audience; a few murmurs in the audience.

Scene 2: Presentation begins; English a bit difficult to follow; 
presenter signals to secretary to put up the first overhead, which she 
does; for about a couple of minutes all goes well.

Scene 3: Presenter asks for next overhead; image comes on screen; 
/wrong/ overhead; presenter starts to become irritated; secretary starts 
to become (more) nervous; overheads are successively put up and put down 
again without success; nervousness increases; presenter tries to go on 
from overhead then visible; secretary start to put up another overhead 
before being told to do so; presenter turns red; secretary panics and 
drops all the sheets on the stage;

Meanwhile, the audience: slow grins appearing on faces; bodies settling 
down more comfortably in chairs; hands fold on bellies.

Scene 4: panic extending on stage; presenter, secretary, and session 
president reassemble the sheets but no logical order can be found any 
more; presentation ends in chaos. (applause)


-- 
Thomas

Post a reply to this message

> You see a job advert for an English / French translator. The advert says
> there will be a test during the interview. You don't know a single word
> of French, but you think to yourself "bah, nobody will notice. I'll just
> Google search of the words when I need to. It'll be fine."
>
> WHAT. THE. HELL.
>
> I'm sorry, IN WHICH UNIVERSE IS NOBODY GOING TO NOTICE?! What the HELL
> are you smoking?!!
>
> Nobody thinks this will work. Nobody ever tries this. Because nobody is
> that stupid. But replace being able to write French with being able to
> write C#, and suddenly everybody thinks that somehow "nobody will
> notice". Wuh?!

I assure you the sames happens for French and everything else. My gf 
works for a company that sells stuff mostly outside the UK, so they are 
always recruiting people for customer service that are fluent in German, 
French, Italian etc. People *do* apply for jobs that blatantly state 
"German speaker" that cannot speak more than the 3 words of German that 
everyone knows. Their reasoning is often along the lines of "well I can 
speak a bit of French and I thought you wouldn't notice, and if you did 
you would anyway see how uber cool I was and see I could learn German in 
a few weeks and hire me anyway". Err no, go away!

The same happened to me in Engineering in my last job too. A guy applied 
quoting 20 years experience in analogue electronic circuit design and 
test. We got a bit concerned when he couldn't even say what a simple 
2-component RC circuit was meant to do, let alone how to calculate any 
response. After a bit of drilling down we got to the question of "have 
you ever actually used an oscilloscope?". His answer was No. I mean come on!

> We started doing phone screenings rather than waste our time with
> hopeless people. We find a website where you can watch the other person
> type, in real-time, as they type it. Almost everybody who reaches this
> stage doesn't get to have a face-to-face interview. A few people got
> creative though:
>
> * While talking to one guy, we could *clearly* hear somebody else
> sitting next to him telling him what to type. Er, REJECT!
>
> * Another guy, we asked him to write some code. We can hear typey typey
> typey. Er, can you type into the window please? Typey, typey, typey,
> BAM! Suddenly a huge block of text appears all at once. In other
> words... you just Googled the answer and copy-pasted it, didn't you?
> REJECT!

LOL

> Seriously, how the **** do people think we're not gonna notice this
> stuff? Let us suppose, for argument's sake, that you somehow trick us
> into interviewing you... Do you REALLY THINK it won't instantly become
> EXTREMELY OBVIOUS that you're bluffing? How do you actually see this
> playing out? Are you mental?!

The scary thing is these people will probably be able to get a job 
somewhere and somehow blag there way along. If they're lucky they may 
even get to interview people at some point :-)

> I wonder... Does anybody have this much trouble hiring a carpenter?

Have you ever seen shows like "cowboy builders" or "rogue traders"...

Post a reply to this message

> But not as disturbing as me getting paid to do C++ when I know nothing
> about it! o_O

Because of your desire to learn though, I'm pretty sure you're going to 
be better at C++ than them already, and if not you will be soon. And 
undoubtedly you've got a wider area of expertise than someone who has 
just spent their entire career coding C++ and nothign else.

Post a reply to this message

On 16/02/2015 09:26, Thomas de Groot wrote:
> Beware of the /binary/ nature though. I must have told this story many
> times, maybe even here...

Not to my knowledge. I have seen the same, a few times with slide 
projectors. When the carousel has been loaded in the wrong order and the 
ensuing panic ends up with the slides on the floor.
Ah! Well. It was the best that we had at the time.

-- 

Regards
     Stephen

Post a reply to this message

On 16/02/2015 09:34, scott wrote:
> The scary thing is these people will probably be able to get a job
> somewhere and somehow blag there way along.

Oi! That's me your talking about. Have a care.

-- 

Regards
     Stephen

Post a reply to this message

On 16-2-2015 9:40, scott wrote:
>>> In short, I haven't come across a lesson where people are there because
>>> they actually *want* to learn something. Most of them are just seem to
>>> want the bit of paper that makes employers think they're awesome.
>>>
>>
>>
>> You are too pessimistic I believe (or the world has changed too much).
>> Except for the last point (the Microsoft course) with which I agree, at
>> all other levels I have been with people who genuinely wanted to /know/
>> and to /learn/. Not all of them of course but a substantial part, yes.
>
> I have to agree with Andy here, my experience of school (compulsory up
> to age 16) and sixth form college (optional up to age 18) is that most
> people didn't want to learn. Obviously there were some who did, but the
> majority you felt were there because someone else had told them to be
> there but they didn't want to be. It was a place to meet people and hang
> out, having to go to lessons and do homework was an inconvenience.
>
> It was only when I got to University that it was suddenly full of people
> who actually wanted to be there and wanted to learn. But I think there
> are likely only a few Universities like that with that atmosphere,
> speaking to friends elsewhere it sounded like other Universities were
> just like school and college where people would spend more effort trying
> to avoid doing any work that actually learning anything.
>
My school experience was quite the opposite; there was even a fair 
competition between pupils to get the highest scores. Good teachers were 
venerated; the bad treated with contempt. It was hard work too, from 
8:00 till 18:00 on average, six days a week, with only the thursday 
afternoon free. I don't know. Time (50s and 60s) and place (Paris) may 
play a role here...

-- 
Thomas

Post a reply to this message

On 16-2-2015 11:07, Stephen wrote:
> On 16/02/2015 09:26, Thomas de Groot wrote:
>> Beware of the /binary/ nature though. I must have told this story many
>> times, maybe even here...
>
> Not to my knowledge. I have seen the same, a few times with slide
> projectors. When the carousel has been loaded in the wrong order and the
> ensuing panic ends up with the slides on the floor.
> Ah! Well. It was the best that we had at the time.
>
A man is entitled to his laugh, ain't he?

-- 
Thomas

Post a reply to this message

On 16/02/15 10:07, Stephen wrote:
> On 16/02/2015 09:26, Thomas de Groot wrote:
>> Beware of the /binary/ nature though. I must have told this story many
>> times, maybe even here...
> 
> Not to my knowledge. I have seen the same, a few times with slide
> projectors. When the carousel has been loaded in the wrong order and the
> ensuing panic ends up with the slides on the floor.
> Ah! Well. It was the best that we had at the time.
> 

Guilty!
Been there, done that. All I need is the t-shirt.

John
-- 
Protect the Earth
It was not given to you by your parents
You hold it in trust for your children

Post a reply to this message