Numerical methods with C++: Any book recommendations

[Steven T. Hatton]

Can anyone recommend a good book on numerical methods that presents the
subject using C++ in an ideal manner?
--
"If our hypothesis is about anything and not about some one or more
particular things, then our deductions constitute mathematics. Thus
mathematics may be defined as the subject in which we never know what we
are talking about, nor whether what we are saying is true." - Bertrand
Russell

 

[Victor Bazarov]

Can anyone recommend a good book on numerical methods that presents the
subject using C++ in an ideal manner?

You need to define "ideal manner".

I used "Scientific and Engineering C++" in some of my work. Liked it.

V

 

[Steven T. Hatton]

You need to define "ideal manner".

I used "Scientific and Engineering C++" in some of my work. Liked it.

V

I mean there are few if any macros, and the following are used effectively:
compositors, gslices, valarrays, overloaded operator[] for indexing
multidimensional structures (tensors). The code should be clear concise and
efficient. The code should be object oriented, but should probably use
dynamic typing sparingly.

One area in particular I would like a good discussion of is indexing, and
particularly the indexing of tensors using gslices.
--
"If our hypothesis is about anything and not about some one or more
particular things, then our deductions constitute mathematics. Thus
mathematics may be defined as the subject in which we never know what we
are talking about, nor whether what we are saying is true." - Bertrand
Russell

 

[E. Robert Tisdale]

Can anyone recommend a good book on numerical methods
that presents the subject using C++ in an ideal manner?

Have you tried
The Object-Oriented Numerics Page?

http://www.oonumerics.org/oon/

 

[Victor Bazarov]

Victor Bazarov wrote:

You need to define "ideal manner".

I used "Scientific and Engineering C++" in some of my work. Liked it.

V

I mean there are few if any macros, and the following are used effectively:
compositors, gslices, valarrays, overloaded operator[] for indexing
multidimensional structures (tensors). The code should be clear concise and
efficient. The code should be object oriented, but should probably use
dynamic typing sparingly.

One area in particular I would like a good discussion of is indexing, and
particularly the indexing of tensors using gslices.

I think you have a good idea what such a book would look like and what it
should contain. Now, all you have to do is write it and find a publisher.
Or find a publisher and write it. Whichever.

Remember this, though: good, clean, efficient, code is hard to come by,
and usually costs a lot [to produce]. So, if, say, I have a clean,
efficient way of working with tensors using gslices, why would I simply
put it in a book when I can put it in a library and sell it?

It's your choice. You can keep dreaming of finding all you need in some
book that presents it in an ideal manner or start working with what you've
got.

V

 

[Jeff Greenberg]

Can anyone recommend a good book on numerical methods that presents the
subject using C++ in an ideal manner?

Can't say if it uses C++ in the "ideal manner" or not, but for getting
real work done (and done correctly) this:

http://www.numerical-recipes.com/cpp-blurb.html

is the canonical reference.

- Jeff

 

[Ron Natalie]

Can't say if it uses C++ in the "ideal manner" or not, but for getting
real work done (and done correctly) this:

http://www.numerical-recipes.com/cpp-blurb.html

is the canonical reference.

I can't vouch for the CPP version of this book, but if it's not
vastly improved over the C version we have, it's not worth dealing
with. While the recipes are "correct" in that they will give
the right answers, the algoritms are frequently far from optimal
and the C coding is frequently horrid.

 

[Richard Herring]

I can't vouch for the CPP version of this book, but if it's not
vastly improved over the C version we have, it's not worth dealing
with. While the recipes are "correct" in that they will give
the right answers, the algoritms are frequently far from optimal
and the C coding is frequently horrid.

Also you're recommended to read the licensing terms very carefully
before incorporating anything (even an idea!) from this book in
deliverable code.

 

[JKop]

Also you're recommended to read the licensing terms very carefully
before incorporating anything (even an idea!) from this book in
deliverable code.

;-D


-JKop

 

[Ron Natalie]

Also you're recommended to read the licensing terms very carefully
before incorporating anything (even an idea!) from this book in
deliverable code.

You can't copyright ideas.

 

[Steven T. Hatton]

You can't copyright ideas.

No, but I intend to apply for a patten the numeric value Pi.
--
"If our hypothesis is about anything and not about some one or more
particular things, then our deductions constitute mathematics. Thus
mathematics may be defined as the subject in which we never know what we
are talking about, nor whether what we are saying is true." - Bertrand
Russell

 

[Richard Herring]

You can't copyright ideas.

Who said anything about copyright?

 

[Ron Natalie]

Who said anything about copyright?

The introduction to the book (I assumed that's what we're talking about).

 

[Richard Herring]

The introduction to the book (I assumed that's what we're talking about).

I was thinking of this:

http://www.numerical-recipes.com/infotop.html#licenseinfo

which is about purchase of licenses to "use" the "software", whatever
those terms mean in that context. You'd need a lawyer to find out.

(I was also being slightly facetious, so let's not drag this any further
off topic.)

 

[Ron Natalie]

I was thinking of this:

http://www.numerical-recipes.com/infotop.html#licenseinfo

which is about purchase of licenses to "use" the "software", whatever
those terms mean in that context. You'd need a lawyer to find out.

(I was also being slightly facetious, so let's not drag this any further
off topic.)

Which still is copyright issues. The license gives you the
authorization to make the copies. No intellectual property
rights other than simple copyright are asserted.