«

"" <> writes:

> If you can cope with it I'd try Knuth's Art of Programming (volume
> one). It takes you through a lot of detail about how computers
> represent things like lists and arrays internally, and all that and
> all in terms of classical computing.

I have never, ever, been able to understand how ANYONE can recommend
these to a beginner. They are not eminently readable at all. They are a
classic work used for reference. They are in NO WAY suitable to
teaching someone how to program a computer.

Technical reference, or a resource to look up some good legalize in
order to dazzle the clc regulars, maybe, but a beginners reference? Come
off it.

Richard wrote:

> "" <> writes:
>
>> If you can cope with it I'd try Knuth's Art of Programming (volume
>> one). It takes you through a lot of detail about how computers
>> represent things like lists and arrays internally, and all that and
>> all in terms of classical computing.
>
> I have never, ever, been able to understand how ANYONE can recommend
> these to a beginner. They are not eminently readable at all. They are
> a classic work used for reference. They are in NO WAY suitable to
> teaching someone how to program a computer.
>
> Technical reference, or a resource to look up some good legalize in
> order to dazzle the clc regulars, maybe, but a beginners reference?
> Come off it.

They are suitable for someone who is quite proficient with mathematics.
The concepts are not difficult at all for an undergraduate or graduate,
as long as he can understand the maths. That's the part where most give
up.

In article <fr12j6$qf$>,
Richard <devr_@gmail.com> wrote:
....
>Technical reference, or a resource to look up some good legalize in
>order to dazzle the clc regulars, maybe, but a beginners reference? Come
>off it.

But that's what's important here in CLC, so it is quite natural that
that is what they would recommend. Makes perfect sense.

santosh <> writes:

> Richard wrote:
>
>> "" <> writes:
>>
>>> If you can cope with it I'd try Knuth's Art of Programming (volume
>>> one). It takes you through a lot of detail about how computers
>>> represent things like lists and arrays internally, and all that and
>>> all in terms of classical computing.
>>
>> I have never, ever, been able to understand how ANYONE can recommend
>> these to a beginner. They are not eminently readable at all. They are
>> a classic work used for reference. They are in NO WAY suitable to
>> teaching someone how to program a computer.
>>
>> Technical reference, or a resource to look up some good legalize in
>> order to dazzle the clc regulars, maybe, but a beginners reference?
>> Come off it.
>
> They are suitable for someone who is quite proficient with mathematics.
> The concepts are not difficult at all for an undergraduate or graduate,
> as long as he can understand the maths. That's the part where most give
> up.

"Most". Thank you.

These are reference books.

The ONLY way to program is with a suitable language and a good
tutorial. Hands on. Lets say "C" here. 99% of good programmers are not mathematics
gods - they are logical thinking, ordered people.

The best introduction to programming I ever read was, and still is, the
tutorial part of K&R.

My first language was z80. I spent ages with pen and paper and my Rodney
Zaks. Yet I learnt more using a monitor on a nascom in 10 minutes than I
did in days of "paper work".

Knuth is not recommended as a beginners "programming tutorial". It is
something for all programmers to keep in their library in order to big
it up a bit IMO.

Richard wrote:

> santosh <> writes:
>
>> Richard wrote:
>>
>>> "" <> writes:
>>>
>>>> If you can cope with it I'd try Knuth's Art of Programming (volume
>>>> one). It takes you through a lot of detail about how computers
>>>> represent things like lists and arrays internally, and all that and
>>>> all in terms of classical computing.
>>>
>>> I have never, ever, been able to understand how ANYONE can recommend
>>> these to a beginner. They are not eminently readable at all. They
>>> are
>>> a classic work used for reference. They are in NO WAY suitable to
>>> teaching someone how to program a computer.
>>>
>>> Technical reference, or a resource to look up some good legalize in
>>> order to dazzle the clc regulars, maybe, but a beginners reference?
>>> Come off it.
>>
>> They are suitable for someone who is quite proficient with
>> mathematics. The concepts are not difficult at all for an
>> undergraduate or graduate, as long as he can understand the maths.
>> That's the part where most give up.
>
> "Most". Thank you.
>
> These are reference books.

I wouldn't call TAoCP as mere reference books. They broke new ground
when they were first published. They are a classic work in the field.

> The ONLY way to program is with a suitable language and a good
> tutorial. Hands on. Lets say "C" here. 99% of good programmers are not
> mathematics gods - they are logical thinking, ordered people.

I did not say otherwise. I was just explaining that *some* programmers
might enjoy TAoCP. But even they are no substitute for practise. BTW, a
programmer who is weak with mathematics is ultimately doomed, no matter
what code he manages to churn out. IMHO of course, YMMV.

> The best introduction to programming I ever read was, and still is,
> the tutorial part of K&R.

Among the best yes.

> My first language was z80. I spent ages with pen and paper and my
> Rodney Zaks. Yet I learnt more using a monitor on a nascom in 10
> minutes than I did in days of "paper work".
>
> Knuth is not recommended as a beginners "programming tutorial". It is
> something for all programmers to keep in their library in order to big
> it up a bit IMO.

I agree. Recommending TAoCP and the C Standard is probably inappropriate
for the OP, but *some* people are likely to enjoy reading the former,
and even fewer number might enjoy reading the latter.

If the OP wants to know the machine specific details of C, a good option
is to get an architecture specific book and an assembly language text,
for the appropriate system of course. A general book on computers might
also help.

santosh <> writes:

> Richard wrote:
>
>> santosh <> writes:
>>
>>> Richard wrote:
>>>
>>>> "" <> writes:
>>>>
>>>>> If you can cope with it I'd try Knuth's Art of Programming (volume
>>>>> one). It takes you through a lot of detail about how computers
>>>>> represent things like lists and arrays internally, and all that and
>>>>> all in terms of classical computing.
>>>>
>>>> I have never, ever, been able to understand how ANYONE can recommend
>>>> these to a beginner. They are not eminently readable at all. They
>>>> are
>>>> a classic work used for reference. They are in NO WAY suitable to
>>>> teaching someone how to program a computer.
>>>>
>>>> Technical reference, or a resource to look up some good legalize in
>>>> order to dazzle the clc regulars, maybe, but a beginners reference?
>>>> Come off it.
>>>
>>> They are suitable for someone who is quite proficient with
>>> mathematics. The concepts are not difficult at all for an
>>> undergraduate or graduate, as long as he can understand the maths.
>>> That's the part where most give up.
>>
>> "Most". Thank you.
>>
>> These are reference books.
>
> I wouldn't call TAoCP as mere reference books. They broke new ground
> when they were first published. They are a classic work in the field.

When they were first published. They are reference books. Tutorials do
not "break new ground" in CS.

I know they are classic works. I have them. I also have a BSc and other
qualifications in CS and EE (sorry, but here I think its justified to
mention qualifications a little in order to dispel any idea I am
clueless about technical references), but I never used these books and
know no others who used these books as part of their courses.

>
>> The ONLY way to program is with a suitable language and a good
>> tutorial. Hands on. Lets say "C" here. 99% of good programmers are not
>> mathematics gods - they are logical thinking, ordered people.
>
> I did not say otherwise. I was just explaining that *some* programmers
> might enjoy TAoCP. But even they are no substitute for practise. BTW, a
> programmer who is weak with mathematics is ultimately doomed, no matter
> what code he manages to churn out. IMHO of course, YMMV.
>
>> The best introduction to programming I ever read was, and still is,
>> the tutorial part of K&R.
>
> Among the best yes.

Gracious of you. You will note I said "I ever read".

>
>> My first language was z80. I spent ages with pen and paper and my
>> Rodney Zaks. Yet I learnt more using a monitor on a nascom in 10
>> minutes than I did in days of "paper work".
>>
>> Knuth is not recommended as a beginners "programming tutorial". It is
>> something for all programmers to keep in their library in order to big
>> it up a bit IMO.
>
> I agree. Recommending TAoCP and the C Standard is probably inappropriate
> for the OP, but *some* people are likely to enjoy reading the former,
> and even fewer number might enjoy reading the latter.

So we are agreed. Its "generally" a stupid idea to recommend either to a
nOOb looking to learn how to program.

>
> If the OP wants to know the machine specific details of C, a good option
> is to get an architecture specific book and an assembly language text,
> for the appropriate system of course. A general book on computers might
> also help.

You are losing it Santosh. The guy is looking for a C tutorial. Not a
dissertation on assembly optimizations in the C framework on platform
dependant setups.

Richard wrote:

> santosh <> writes:

<snip>

>> If the OP wants to know the machine specific details of C, a good
>> option is to get an architecture specific book and an assembly
>> language text, for the appropriate system of course. A general book
>> on computers might also help.
>
> You are losing it Santosh. The guy is looking for a C tutorial. Not a
> dissertation on assembly optimizations in the C framework on platform
> dependant setups.

Well he said he wanted a book that would go into system specific details
of C omitted from things like K&R and the Standard, like twos
complement, sign and magnitude, floating point formats, memory
alignment, etc.

I still say that an a book like say Write Great Code might help him.

santosh <> writes:

> Richard wrote:
>
>> santosh <> writes:
>
> <snip>
>
>>> If the OP wants to know the machine specific details of C, a good
>>> option is to get an architecture specific book and an assembly
>>> language text, for the appropriate system of course. A general book
>>> on computers might also help.
>>
>> You are losing it Santosh. The guy is looking for a C tutorial. Not a
>> dissertation on assembly optimizations in the C framework on platform
>> dependant setups.
>
> Well he said he wanted a book that would go into system specific details
> of C omitted from things like K&R and the Standard, like twos
> complement, sign and magnitude, floating point formats, memory
> alignment, etc.

Fair enough - I did lose track of that a little in the thread. Frankly I
think he's being too hopeful. Pick up K&R and learn the rest from online
resources as you go. None of it is rocket science. Writing clean
maintainable code is the best skill he could ever have over nitpicking
over whether a pointer to an array is ..."insert quibble here".

Francois wrote:
> Thank a lot.
>
> I have to possibility :
>
> [1] Computer Systems: A Programmer's Perspective
>
> or
>
> [2] Memory As a Programming Concept in C and C++
>
>
> If I have understood, [1] doesn't talk about C. Is it true ?

As I said in my earlier post, the two topics are to an extent mutually
exclusive. C is designed to be platform-neutral and deliberately hides
the mechanics from its users, whereas accessing hardware is by
definition hardware-specific, while many languages will use the same
method on that hardware.

The best you can hope for is to pick a platform, and ask in newsgroups
specialising in that platform.

> I'll wished make the good choice, because the prices are a little
> expensive.

Most libraries will obtain books for you on request.

<OT>
I quite like the Bodleian in Oxford, as it claims to have a copy of
almost every book in existence. I once tested that by ordering up
obscure science fiction books and a copy of a hand-bound work by Gibbon.
They all appeared. Lord knows what the librarians thought of my request...
</OT>

santosh <> writes:

> Richard wrote:
>
>> "" <> writes:
>>
>>> If you can cope with it I'd try Knuth's Art of Programming (volume
>>> one). It takes you through a lot of detail about how computers
>>> represent things like lists and arrays internally, and all that and
>>> all in terms of classical computing.
>>
>> I have never, ever, been able to understand how ANYONE can recommend
>> these to a beginner. They are not eminently readable at all. They are
>> a classic work used for reference. They are in NO WAY suitable to
>> teaching someone how to program a computer.
>>
>> Technical reference, or a resource to look up some good legalize in
>> order to dazzle the clc regulars, maybe, but a beginners reference?
>> Come off it.
>
> They are suitable for someone who is quite proficient with mathematics.
> The concepts are not difficult at all for an undergraduate or graduate,
> as long as he can understand the maths. That's the part where most give
> up.

The maths, the (obsolete) MIX assembly language (with its C-ish
ambiguities wrt word sizes), and the mathematicians' terseness of
language.

I have spent hours on a paragraph or two (of English) in some portions
of TAOCP, trying to delve the full meaning of what Knuth is saying.

I love Knuth, and I love TAOCP, but I've absolutely got to agree that
it is in no way suitable for beginners.

--
(But just my 2¢, of course)
Micah J. Cowan
Programmer, musician, typesetting enthusiast, gamer...
http://micah.cowan.name/