[ot?] matrix inversion

[Jarno A Wuolijoki]

Um what is more critical to writing decent software. Knowing math, not
being lazy, being motivated or spelling correctly in usenet... Hmm...

Well, for instance, I just /love/ software with an attitude problem..

 

[Joona I Palaste]

If you think that he cares about being killfiled by you, you have a
serious ego problem.

What makes you think I care about him caring about being killfiled by
me? Oh, I see. Right. Well I'll stand by my words anyway.

 

[Ben Pfaff]

CBFalconer <[email protected]> scribbled the following:

[regarding Tom St. Denis]

I believe him to be a pubescent teenager from Ottawa, Canada with
a severe civility disorder.

Why specifically from Ottawa, Canada?

http://iahu.ca:8080/aboutme.html

 

[Joona I Palaste]

CBFalconer <[email protected]> scribbled the following:
[regarding Tom St. Denis]

I believe him to be a pubescent teenager from Ottawa, Canada with
a severe civility disorder.

Why specifically from Ottawa, Canada?

http://iahu.ca:8080/aboutme.html

Hmm, I wasn't aware of that. Well, that's not *conclusive* evidence -
even assuming the author of that web page speaks the truth (which he
has no real reason not to, but still), it might not be the same Tom
St Denis. *I* believe it is - but I won't jump to conclusions based
on my beliefs.

--
/-- Joona Palaste ([email protected]) ------------- Finland --------\
\-- http://www.helsinki.fi/~palaste --------------------- rules! --------/
"Outside of a dog, a book is a man's best friend. Inside a dog, it's too dark
to read anyway."
- Groucho Marx

 

[E. Robert Tisdale]

where can I find ready-to-use function like

void Inverse(float **v; int n);

that will inverse given array (I need only inverting rectangular arrays).

Best if it would be open-source.

I see that you've already met our indigenous trolls.
Please ignore them.

Try the GNU Scientific Library

http://sources.redhat.com/gsl/

Matrix inversion is performed in two steps:

1. decomposition and
2. back substitution.

The decomposition is required to test whether or not
the matrix is singular or ill-conditioned --
whether or not it has a meaningful numerical inverse.
Look for an LU decomposition and the matching solver.

 

[Alan Balmer]

CBFalconer <[email protected]> scribbled the following:
[regarding Tom St. Denis]
I believe him to be a pubescent teenager from Ottawa, Canada with
a severe civility disorder.

Why specifically from Ottawa, Canada?

http://iahu.ca:8080/aboutme.html

Hmm, I wasn't aware of that. Well, that's not *conclusive* evidence -
even assuming the author of that web page speaks the truth (which he
has no real reason not to, but still), it might not be the same Tom
St Denis. *I* believe it is - but I won't jump to conclusions based
on my beliefs.

Assuming the page itself isn't a fake, the fact that the email address
given is the same is certainly suggestive ;-)

 

[Sidney Cadot]

And that would be? If you don't know high school algebra... um.... YOU
SHOULD NOT BE WRITING SOFTWARE!!!

If you think that an understanding of high school algebra suffices for
implementing a decent-quality floating point matrix inverter, you are
sadly, sadly mistaken.

Regards, Sidney

 

[CBFalconer]

If you think that he cares about being killfiled by you, you have
a serious ego problem.

No egos around here The threat is rather pointless, however I
consider PLONK announcements useful in that they encourage others
to co-PLONK. The end result is that the PLONKee is ignored, and
sometimes goes away. 'Tis often a consumation devoutly to be
wished.

 

[CBFalconer]

Why specifically from Ottawa, Canada?

Don't remember. If you want to search his posts for the past few
years you will probably find out. Maybe his parents moved in the
interim. However he is an undeserved black mark against Canadian
education and upbringing.

 

[Mark McIntyre]

Can I say that if Tom can't spell "write" right, he has no business
writing to technical newsgroups at all?

blast, you beat me to it.

 

[Mark McIntyre]

Um what is more critical to writing decent software. Knowing math, not
being lazy, being motivated or spelling correctly in usenet... Hmm...

Anyone who's too lazy to spell correctly is quite possibly also going
to have a problem turning out quality code. IMHO, YMMV.

 

[Tom St Denis]

Anyone who's too lazy to spell correctly is quite possibly also going
to have a problem turning out quality code. IMHO, YMMV.

You guys are realy, relly, rally, petty.

Dudes looking for homework answers and you pick on my spelling. Seems to me
I'm not the one with messed up priorities. And as to your earlier post, if
I can't churn out quality code explain the hundreds of users of my software?

Tom

 

[Tom St Denis]

Don't remember. If you want to search his posts for the past few
years you will probably find out. Maybe his parents moved in the
interim. However he is an undeserved black mark against Canadian
education and upbringing.

You could just ask me. Since well I post to usenet with my real email
address. In fact I also post my full contact info in various things I do.

Anyways, I do live in Ottawa and I thank my public education everyday that I
was given a huge dosage of critical thinking. Otherwise I wouldn't have
taught myself all the stuff I have so far.

Tom

 

[Alan Balmer]

Dudes looking for homework answers and you pick on my spelling. Seems to me
I'm not the one with messed up priorities. And as to your earlier post, if
I can't churn out quality code explain the hundreds of users of my software?

Shucks, Microsoft has *millions* of users. Explain that.

 

[Sidney Cadot]

You guys are realy, relly, rally, petty.

Are you positively sure you didn't mean "pretty" ?

Dudes looking for homework answers and you pick on my spelling. Seems to me
I'm not the one with messed up priorities. And as to your earlier post, if
I can't churn out quality code explain the hundreds of users of my software?

You first! Present a plausible case for a correlation between number of
users and code quality.

Anyway, most here can probably stomach a spelling mistake or two, but
the faul language was certainly uncalled for. If I want any of that I'll
switch on MTV after 10pm, thank you very much.

Best regards,

Sidney

 

[CBFalconer]

I see that you've already met our indigenous trolls.

For once Trollsdale seems to have quoted somebody without making
alterations. Never trust anything in a Trollsdale message. His
quotes are edited to amuse him, and his advice is probably wrong,
as are his "facts". Unfortunately you can't just PLONK him
because he will then be able to misquote you without correction.

 

[Dik T. Winter]

>

>
> Um what is more critical to writing decent software. Knowing math, not
> being lazy, being motivated or spelling correctly in usenet... Hmm...

For most applications math is not needed at all. Spelling correctly is
critical to writing decent software.

> Maybe the question smacks of homeworkism? Cuz the bitch is probably off
> from winter break and now has an assignment due. Instead of actually paying
> attention in class and doing the lab work he probably skips class then last
> minute looks for a fix.

If it is a homework question, it is for a course badly given. Matrix
inversion is almost never done as it leads to results that are much less
accurate than needed. Also the OP's original words (rectangular matrix)
leads to questions like: "what is the purpose?" (Rectangular matrices do
not have an inverse, unless they are square, and even in that case it is
possible that they do not have an inverse.) Even simple Gauss elimination
is far from trivial.

> That and the OP ought to learn what the **** google is. E.g. a quick google
> turns up
>
> http://www.library.cornell.edu/nr/bookcpdf.html

Can you tell me where in the Numerical Recipes book information is given
about matrix inversion? I do not think it is in there, only discussion
about the complexity. Nor does the book give detailed information about
the implementation. Going from the book to an actual implementation is
an, eh, large step.

For the OP, this newsgroup is probably not the best. Sci.math.num-analysis
might be a better platform.

 

[Dik T. Winter]

> Rafal 'Raf256' Maj wrote: ....
>
> I see that you've already met our indigenous trolls.
> Please ignore them.

Eh? But:

> Try the GNU Scientific Library
> http://sources.redhat.com/gsl/
> Matrix inversion is performed in two steps:
> 1. decomposition and
> 2. back substitution.
>
> The decomposition is required to test whether or not
> the matrix is singular or ill-conditioned --
> whether or not it has a meaningful numerical inverse.
> Look for an LU decomposition and the matching solver.

In most cases it is not necessary to calculate the inverse. When you
want to solve a linear system, going the way with an inverse gives
worse results than doing the back substitution directly.

 

[CBFalconer]

Shucks, Microsoft has *millions* of users. Explain that.

Myrmidons. Both sets.

 

[Tom St Denis]

software?

You first! Present a plausible case for a correlation between number of
users and code quality.

My code is free and meant for developers [hint: I write cryptographic and
mathematic libraries].

So if the people didn't like my code they'd just move on and ignore it since
they have nothing vested in.

Before this gets blown out of proportion I'd like to say I'm the first to
admit my code isn't always perfect but I fix what bugs I find in a timely
fashion and that's kept me some big users like Sony and BitMover [makers of
BitKeeper, the thingy that keeps the Linux kernel....].

Tom