Sad to say these are that I don’t care for it very much.
The “Unity” interface is just ugly and it reminds me of the way Microsoft shifted their Office menus so that, three years on from first using it I still don’t know where familiar and useful things are.
Why does it have Ubuntu One and Ubuntu Software Centre icons in pride of place? I have never made much use of either and frankly it smacks of somebody forcing their wares on me – the sort of thing that is likely to go down like a cup of cold sick in the Linux community.
Maybe there are lots of other good things here, but I cannot find them thanks to Unity. I cannot even find my bookmarks in Firefox the way that has been messed with.
If you have ever downloaded any software from the internet you are quite likely to have seen a reference to a “CRC” – a cyclic redundancy check.
I have been reading about them as I plough on through Andrew Tanenbaum‘s (pictured) Computer Networks: so here is an explanation based closely on his text (though it’s not a copy except where indicated)…
CRCs rely on some properties of polynomial arithmetic, namely polynomials with the coefficients 0 or 1 only.
We say that a is a list of coefficients for a polynomial with terms ie for a polynomial ranging from to (which is said to have ).
So represents .
Polynomial arithmetic is a form of modulo-2 arithmetic “according to the rules of algebraic field theory” (that’s what it says in the book!). In practise what this means is that there are no “carry-overs” and addition and subtraction are identical and the same as exclusive or (XOR).
Hence while and so on…
To use a CRC a sender and receiver must agree a generator polynomial: in fact these are now generally standardised and IEEE 802 (network protocols) defines:
(the last two terms are of course equivalent to ).
1. Let us call this generator polynomial and let represent the polynomial (ie the sequence of 0s and 1s) we wish to transmit (with bits) – must have more bits than .
2. Let r be the degree of – so we append r zero bits to the lower end of the frame, which now has bits and is (or ).
3. Divide by using modulo 2 division.
4. Subtract the remainder (which will always be r or fewer bits (as is of degree r) from the string (again using modulo 2 subtraction). This gives the checksummed frame to be transmitted – we shall call this .
If the frame is transmitted correctly then when it is received it should be cleanly divisible by (as taking away the remainder in step 4 should leave us with a cleanly divisible frame). The additional checksum is merely the additional bits at the end.
The power of the checksum relies on some of its mathematical properties – it will always detect single bit errors if contains more than one term – as the bad frame is now where is the error but in this case where is the “bad bit” but with a two-termed , for any .
Similarly, if there are two errors then so if does not cleanly divide where can be any value from 1 to the maximum of (ie the frame length) then any two bit error would be detected.
Similarly, if there are an odd number of errors then cannot be divisible by a that contains the terms . We can show by a reductio ad absurdum proof:
1. We know that an must be capable of evaluating to one: eg must evaluate to 1 if . But if is divisible by then , but if then, modulo 2, so .
To really fool the system the error pattern would have to match which is, assuming all transmission errors are equally likely (in fact they are not, but we will ignore that for now), has a probability (again, the book says) of which is pretty small for the IEEE picked . (The GNU calculator says it is zero – so it’s too small to register there!)
There is a very interesting article on the THES website today – “So last century” – which says that universities are failing students because they teach them in such a compartmentalised way.
Whether the study is conducted by the CBI in the UK or by commercial for-profit educational providers drumming up business for their remedial post-baccalaureate job-training services, everyone seems to acknowledge that today’s students are good test-takers but lack the workplace essentials necessary for the 21st century. These include people skills (especially in diverse global contexts), communication skills, collaborative skills, analytical skills, networking skills, an ability to synthesise information across a wide range of evidence, and even the most elementary skills, such as how to write a great job application letter and curriculum vitae or represent their character and talent at a job interview. No wonder they face the career centre with such trepidation.
Well, while the article is well written and an interesting read I think it is fundamentally wrong.
My objection is not just the standard, liberal, view that actually universities are not meant to be factories turning out people fit for labour (albeit higher labour) – though it is partly that. Good employers invest in their employees in any case and do not expect them to be delivered for free as the fully formed article.
I certainly do not believe that there was some golden age where graduates were being turned out with all these skills either. One only has to look at the somewhat doolally behaviour – from Newton onwards – of some of the greatest scientists to know that academic excellence and socialisation are always the easiest of bedfellows.
But it is more fundamental – degrees have to be specialised, certainly in science, because they also have to be, at least partially, grounded in research and a preparation for research.
My first degree was in Astrophysics. “Communication skills, collaborative skills…” and all the rest of it have nothing very much to do with cosmology and every minute that would be spent teaching me about them would be a minute wasted in preparing me for being an astrophysist.
The fact I never became an astrophysist is hardly the point – there would not have been much point to an astrophysics degree if it at least did not offer that path.
Science degrees need to be specialised because science is ever-more complex and specialised. Some scientists think this may be a temporary thing – in A Brief History Of Time Stephen Hawking suggests that further scientific advance might simplify our theories – but there is no sign of that at present.
So, if universities are to produce scientists they have to focus on science and not think of themselves as pre-office work trainers.
Date is concerned not to write another SQL primer but to give those with some SQL experience a good grounding in the relational model and the set theory on which it is based. It just feels more like a proper scientific/mathematical text book and it also seems to be well written.
Very pleased with the purchase and will keep you all posted on how it goes: at the moment I am just wishing I had bought it months ago.