NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c-part2/15760

Internet Message Format | 1996-08-05 | 5.3 KB

Path: mayne.ugrad.cs.ubc.ca!not-for-mail
From: [e-mail] (Kazimir Kylheku)
Newsgroups: comp.lang.c
Subject: Re: Are you really my colleagues?!?
Date: 21 Apr 1996 10:26:51 -0700
Organization: Computer Science, University of B.C., Vancouver, B.C., Canada
Message-ID: <[e-mail]>
References: <[e-mail]> <4l4n4i$[e-mail]> <[e-mail]> <4lcmbf$[e-mail]>
NNTP-Posting-Host: mayne.ugrad.cs.ubc.ca

In article <4lcmbf$[e-mail]>, Jimen Ching <[e-mail]> wrote:
>Kazimir Kylheku ([e-mail]) wrote:
>[ a lot of retorical comments ignored ]
>
>>There are no ``algorithms'' in C, only implementations thereof.
>
>I thought this was always implied.
>
>>The question might be asked as ``what is a good representation of a matrix in
>>the C language''.
>
>Do you always second guess other peoples questions?  What I wanted was
>a C function which does matrix inversion.  I.e.  an *implementation* of
>an algorithm to do matrix inversion, IN C.  BTW, the *IN C* was the major
>point.  You can substitute anything you want for 'algorithms'.

Check Numerical Recipes in C.  Textbooks on numerical analysis are likely to
have good algorithms (which are trivial to code into C). There are many ways of
doing the inversion, with various merits. There are several ways to choose the
pivotal elements. There are even ways to do _exact_ gaussian elimination of
matrices that have integral coefficients where the results are precise
rational numbers. That way there is never a question whether a diagonal element
is truly zero, or just a very small epsilon.

Which method were you asking for? Do you actually have a serious application
for this or will you accept _any_ method for matrix inversion?

What about orthogonal matrices? I have gotten by with just using the transpose
of a matrix in that case, instead of naively computing an inverse.

There are people here that might not even know or remember what a matrix
inverse is. It requires some level of expertise in the field of linear
algebra that is unrelated to C programming.

It's harmless enough to ask the question, but why are you insisting that it's
canonically appropriate for comp.lang.c?

>>You are taking Usenet far too seriously and personally. How do you know whether
>>or not I can't interact with other people?
>
>If your definition of taking Usenet too seriously/personally means that
>I read c.l.c in search of interesting topics concerning the C language
>(the only system specific group I read is Linux), and when I find such
>a topic, all I see are flames to the author about the inappropriateness
>of the topic.  Then yeah, I do take it seriously and personally.  I'm
>tired of such crap.  Seems like I'm not the only one.

So how many useful answers did you get about matrix inversion? How do the
answers rate against ones found in good reference texts or in appropriate
newsgroups? Was it really worth-while to ask here rather than somewhere else?

>As for how do I know you can't interact with other people?  Well, simple.
>From the comment below and your intolerance of off topic posts.  It shows
>that you are a self rightous and obnoxious person.  Note, this is only

No. I sometimes participate in off topic posts (like this thread---prime
example). But this is not the canonical thing. We can't encourage off
topic posts because they happen! That's my stance. I will unofficialy condone
off-topic posts, but I won't go as far as the arrogance to say that they ought
to be the norm.

>I don't know why you had the need to tell me this, but thank you for
>sharing.  I guess somehow, all the years I've been reading c.l.c, I've
>missed all your posts.  I'll try to be more thorough next time.

Start with the FAQ, which somehow also passed you by in ``all the years'':

14.12:  I'm looking for some code to do:
                Fast Fourier Transforms (FFT's)
                matrix arithmetic (multiplication, inversion, etc.)
                                                   ^^^^^^^^^
                complex arithmetic

A:      Ajay Shah maintains an index of free numerical software; it is
        posted periodically, and available where this FAQ list is
        archived (see question 20.40).  See also question 18.16.


Found that with a simple substring search for the word ``matrix''.


Here is a listing of the directory /pub/usenet/comp.lang.c at rtfm.mit.edu

-rw-rw-r--   8 root     3          258633 Apr  5 01:19 C-FAQ-list
-rw-rw-r--   8 root     3           56602 Apr 16 09:44 C-FAQ-list.abridged
-rw-rw-r--   4 root     3            1302 Apr  4 04:07 C-FAQ-list.diff
-rw-rw-r--   8 root     3           29033 Apr  5 01:19 C-FAQ-list.toc
-rw-rw-r--  10 root     3           33561 Mar  7 00:16 Part_1_of_3:_Free_C,C++_for_numerical_computation
-rw-rw-r--  10 root     3           33783 Mar  7 00:16 Part_2_of_3:_Free_C,C++_for_numerical_computation
-rw-rw-r--  10 root     3           23232 Mar  7 00:17 Part_3_of_3:_Free_C,C++_for_numerical_computation
-rw-rw-r--  16 root     3           33728 Apr 20 04:40 Programming_for_Internationalization_FAQ
-rw-rw-r--   4 root     3           53075 Apr  2 01:33 comp.lang.c_Answers_to_Infrequently_Asked_Questions_(FAQ_List)
-rw-rw-r--   1 root     3             514 Apr 21 07:18 index

see the three files promising to reveal sources for obtaining numerical code,
as the FAQ says?