«

Hi,

Is there a book on C++11 ?

2nd edition of "Thinking in c++" was published on 2000.
Is there any new edition of this book coming in near future with C++11?

Sincerely,
Srinivas

On Saturday, June 16, 2012 6:38:58 AM UTC-5, Srinivas Nayak wrote:
> Hi,
>
> Is there a book on C++11 ?
>
> 2nd edition of "Thinking in c++" was published on 2000.
> Is there any new edition of this book coming in near future with C++11?
>
> Sincerely,
> Srinivas

I have a related question.

Should I bother learning c++ 11 if I'm learning C++ to get in to the programming
field. Will big C++ 03 projects continue to hire or is C++11 suddenly in demand?

I'm considering just picking up the books in the FAQ because they're cheap and effective, despite being outdated by the new standard.

jjtf wrote:
> On Saturday, June 16, 2012 6:38:58 AM UTC-5, Srinivas Nayak wrote:
>> Hi,
>>
>> Is there a book on C++11 ?
>>
>> 2nd edition of "Thinking in c++" was published on 2000.
>> Is there any new edition of this book coming in near future with C++11?
>>
>> Sincerely,
>> Srinivas
>
> I have a related question.
>
> Should I bother learning c++ 11 if I'm learning C++ to get in to the programming
> field.
> Will big C++ 03 projects continue to hire or is C++11 suddenly in demand?
The answer probably most depends on your timing for entering the market. Right
now, a mid-to-big size company I work for does not use C++11 or asks
C++11-specific questions at the interviews. You would not miss a point saying
you have not had a chance to work in C++11. Some awareness about useful C++11
features (just being able to legibly explain what's good about them) might be a
(really small) "plus".

I speculate this will stay so for a couple of years; then this might start
gradually change. Smaller companies may start using C++11 faster. The keyword
here is "start": tons of 1998- and 2003- compliant code are probably bound to be
written for many years and to stay around and to be maintained for decades.

> I'm considering just picking up the books in the FAQ because they're cheap and effective, despite being outdated by the new standard.
What I pay most attention to are the focus, software design, problem solving and
practical coding skills and common sense. I bet you won't waste your time
concentrating on these instead of syntactic peculiarities you have never used --
just be honest and upfront about it during your interviews.

For the books, apart from generally solid FAQ recommendations, "Thinking in..."
should be just fine; for something more C++-ish and modern (not yet 11-ish), try
"Modern C++ Design" by Andrei Alecsandrescu -- just do not jump into applying
Loki to all seemingly appropriate problems without giving it a good second (or
third) thought.

Also, IMHO reading an important piece of source code (e.g. that of a modern
version of C++ Standard Library) can be more instructive than reading a good book.

-HTH
Pavel

On Sun, 2012-06-17, Pavel wrote:
> jjtf wrote:
....
>> I have a related question.
>>
>> Should I bother learning c++ 11 if I'm learning C++ to get in to the programming
>> field. Will big C++ 03 projects continue to hire or is C++11 suddenly in demand?

> The answer probably most depends on your timing for entering the market. Right
> now, a mid-to-big size company I work for does not use C++11 or asks
> C++11-specific questions at the interviews. You would not miss a point saying
> you have not had a chance to work in C++11. Some awareness about useful C++11
> features (just being able to legibly explain what's good about them) might be a
> (really small) "plus".

I'd think it would be a real plus, but only because it shows you're
keeping up with what's happening. On the other hand lots of people are
doing that -- C++11 got lots of media attention.

> What I pay most attention to are the focus, software design, problem solving and
> practical coding skills and common sense. I bet you won't waste your time
> concentrating on these instead of syntactic peculiarities you have never used --
> just be honest and upfront about it during your interviews.

Good advice.

....
> Also, IMHO reading an important piece of source code (e.g. that of a modern
> version of C++ Standard Library) can be more instructive than reading a good book.

You need to do both. Not sure reading the STL is the best thing you
can do though -- general-purpose libraries are very unlike the code
you'll spend most of your time reading and writing.

/Jorgen

--
// Jorgen Grahn <grahn@ Oo o. . .
\X/ snipabacken.se> O o .

Jorgen Grahn wrote:
> On Sun, 2012-06-17, Pavel wrote:
>> jjtf wrote:
> ...
>>> I have a related question.
>>>
>>> Should I bother learning c++ 11 if I'm learning C++ to get in to the programming
>>> field. Will big C++ 03 projects continue to hire or is C++11 suddenly in demand?
>
>> The answer probably most depends on your timing for entering the market. Right
>> now, a mid-to-big size company I work for does not use C++11 or asks
>> C++11-specific questions at the interviews. You would not miss a point saying
>> you have not had a chance to work in C++11. Some awareness about useful C++11
>> features (just being able to legibly explain what's good about them) might be a
>> (really small) "plus".
>
> I'd think it would be a real plus, but only because it shows you're
> keeping up with what's happening. On the other hand lots of people are
> doing that -- C++11 got lots of media attention.
>
>> What I pay most attention to are the focus, software design, problem solving and
>> practical coding skills and common sense. I bet you won't waste your time
>> concentrating on these instead of syntactic peculiarities you have never used --
>> just be honest and upfront about it during your interviews.
>
> Good advice.
>
> ...
>> Also, IMHO reading an important piece of source code (e.g. that of a modern
>> version of C++ Standard Library) can be more instructive than reading a good book.
>
> You need to do both. Not sure reading the STL is the best thing you
> can do though -- general-purpose libraries are very unlike the code
> you'll spend most of your time reading and writing.
Yes, the domain one will work for will probably be different but this does not
make this code less instructive. In my experience, the source code of STL in
particular is one of the best exposures of non-obvious issues one would face to
efficiently implement seemingly straightforward specifications in C++.

STL API is relatively simple but an efficient implementation uses many more
standard C++ features and implementation-specific extensions than typical
small-scale STL client code (let alone a collection of examples). Carefully
studying why this or that feature was used helps understand why those more
complex features were defined as they were and where they were either
insufficient or inconvenient enough for the library author to use extensions. A
modern version may also give an idea how new C++11 features can be useful (e.g.
see usage of GNU __alignof__ extension in GNU STL that can be replaced by the
standard alignof operator in C++11).

The added advantage of reading STL code is deeper understanding STL specs. The
reader would probably have some initial idea and, as s/he oceeds, s/he will
learn the details s/he will apply when using STL later and get an idea why the
specs are what they are.

>
> /Jorgen
>


-Pavel

On Sun, 2012-06-17, Pavel wrote:
> Jorgen Grahn wrote:
>> On Sun, 2012-06-17, Pavel wrote:
>>> Also, IMHO reading an important piece of source code (e.g. that of a modern
>>> version of C++ Standard Library) can be more instructive than reading a good book.
>>
>> You need to do both. Not sure reading the STL is the best thing you
>> can do though -- general-purpose libraries are very unlike the code
>> you'll spend most of your time reading and writing.

> Yes, the domain one will work for will probably be different but this does not
> make this code less instructive. In my experience, the source code of STL in
> particular is one of the best exposures of non-obvious issues one would face to
> efficiently implement seemingly straightforward specifications in C++.

I agree reading it is useful. But you don't get this from it: the
mapping from a concrete typical real-world problem into a (probably
object-oriented) design, and the mapping of that into C++ code.

STL *does* solve a real-world problem, but an unusual one: to provide
efficient general-purpose containers and algorithms which will suit
almost all needs.

/Jorgen

--
// Jorgen Grahn <grahn@ Oo o. . .
\X/ snipabacken.se> O o .

Jorgen Grahn wrote:
> On Sun, 2012-06-17, Pavel wrote:
>> Jorgen Grahn wrote:
>>> On Sun, 2012-06-17, Pavel wrote:
>>>> Also, IMHO reading an important piece of source code (e.g. that of a modern
>>>> version of C++ Standard Library) can be more instructive than reading a good book.
>>>
>>> You need to do both. Not sure reading the STL is the best thing you
>>> can do though -- general-purpose libraries are very unlike the code
>>> you'll spend most of your time reading and writing.
>
>> Yes, the domain one will work for will probably be different but this does not
>> make this code less instructive. In my experience, the source code of STL in
>> particular is one of the best exposures of non-obvious issues one would face to
>> efficiently implement seemingly straightforward specifications in C++.
>
> I agree reading it is useful. But you don't get this from it: the
> mapping from a concrete typical real-world problem into a (probably
> object-oriented) design, and the mapping of that into C++ code.
Oh, yes, I most certainly agree STL does not provide much insights to
OOA&D-in-C++. Some other Standard C++ Library facilities, like streambuf
hierarchy, do more of it -- but I am far from calling streambuf a good OOA&D
example.

IMO the main benefit from Standard Library code reading is learning the language
features and their application to low-level problems rather than developing
one's design-at-large skills. STL, however, provides an example of good design
as an additional benefit even though it does not use many OO techniques. I have
to admit I don't really know whether STL is technically OO or not: on one hand,
it has classes and achieves main goals of a good OO design, such as loose
coupling and high cohesion; on the other, it does not use virtual functions,
is-a type of inheritance etc. I am not crazy about pigeon-holing so I won't lose
my sleep about this question.

>
> STL *does* solve a real-world problem, but an unusual one: to provide
> efficient general-purpose containers and algorithms which will suit
> almost all needs.
>
> /Jorgen
>


-Pavel

jjtf wrote:

> I have a related question.
>
> Should I bother learning c++ 11 if I'm learning C++ to get in to the
> programming field. Will big C++ 03 projects continue to hire or is C++11
> suddenly in demand?

It should be pointed out that C++11 is essentially a superset of C++03.
This means that learning C++11 is essentially the same thing as learning
C++03 along with a number of additional features. So, there is no reason to
place yourself in a spot where you believe you should pick one or the other.

In practical terms, it's probably a good idea to focus on C++03, simply
because it's something that you need to do anyway (being a subset of C++11),
and as it is expected that it will take some time before C++11-compliant
compilers have become widely adopted.

In addition, there is obviously more software out there that doesn't use
C++11 features than the one that does.

Nevertheless, some C++11 features are terribly easy to pick up, and some
were even already available through non-standard components. This means
that there is essentially no excuse to avoid learning them. For example,
range-based for loops, nullptr and using >> in nested templates definitions
are picked up without any problem, as is using override and final in member
function declarations. Even if you don't use C++11's std::tuple and
std::array, there is always boost's version.

So, in short, although it might be preferable to focus for now on C++11's
C++03 subset, you are doing yourself a disservice if you intentionally avoid
getting acquainted with the C++11 goodies.


Rui Maciel

Pavel wrote:

> Also, IMHO reading an important piece of source code (e.g. that of a
> modern version of C++ Standard Library) can be more instructive than
> reading a good book.

That really depends on the source code and the reader's background. For
example, if a reader isn't acquainted with design patterns then, when
reading source code where design patterns are implemented, some details will
either appear terribly convoluted or even won't even be noticed at all. And
let's not even talk about template metaprogramming. Some of the stuff out
there, even the basic stuff, tends to be a bit mind-blowing if a reader
doesn't know what he is looking at.


Rui Maciel

Rui Maciel wrote:
> Pavel wrote:
>
>> Also, IMHO reading an important piece of source code (e.g. that of a
>> modern version of C++ Standard Library) can be more instructive than
>> reading a good book.
>
> That really depends on the source code and the reader's background. For
> example, if a reader isn't acquainted with design patterns then, when
> reading source code where design patterns are implemented, some details will
> either appear terribly convoluted or even won't even be noticed at all. And
> let's not even talk about template metaprogramming. Some of the stuff out
> there, even the basic stuff, tends to be a bit mind-blowing if a reader
> doesn't know what he is looking at.
True. One reason I believe the C++ Standard Library code is worth reading is
exactly that the reader would now what he is looking at because the purpose of
every exposed API feature is well-documented and it has to be well understood by
the learner anyway. That is, even if he does not know some of it, it's never
waste of his time to read more of the Standard Library specs and learn.

>
>
> Rui Maciel
>


-Pavel