Re: Re: fsolve features

[Rolf Fabian]

----------

Francesco Potortì's (researcher) address@hidden 
REPLY>

>> [x,RC]= fsolve(f='x^2 +1',randn(1)); #>V2.0.14
 >>      #NOTE: prior to V2.0.14 , you have to  define  f  explicitely:
 >>      #function y=f(x) y=x^2+1; end
 >>      #[x,RC]=fsolve('f',randn(1))

> !!!  Where is this documented?

In octave V2.0.14 documentation  (

>>   (2)   a call to fsolve  like :  [x,info]=fsolve('f',x0)   is adequate for
>>  function   (vector) Y =  f ( vector X  ) 
>>        >>> where Y and X have same length <<<<
>>         It's supposed to handle equally sized problems only!


>>   It is easy to work around this. 
>>   Maybe you are suggesting that a workaround
>>   is nt what you would like to have?

If this's SO EASY why didn't/don't YOU hack these obviously rather limited 
couple
of  short program lines together and provide it to the public ?  Why didn' 
anybody
else do such an EASY job within the last 5 years or so ...?

>>......
>>   Because of all these pitfalls of 'fsolve', I decided to write my my
>>   personal solver 'cfsolve' , returning correct complex roots for this (or
>>   similar) problem


>Where is your cfsolve?

cfsolve -algorithm development started in 1998 and  reached Version 4.1 up to 
now !
Originally I planned to provide it to  octave community. But it's not published 
yet, because
there are still some problems present associated with the operational mode of 
'over-dimensioned' problems. 
Equal-  and  under-dimensioned modes are very  well treated on the other hand.

Looking  back at more than one year's intense ( homebrew ... not full ) work  
dealing with 
this problem  while reading your statement above, I get in really great doubt, 
whether my
conceptual approach  to the problem could be smart enough  to be efficient in 
contrast
to your genious EASY solution !

Obviously, there seems to be no necessity anymore to finish this project from 
my side ....




---------------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.  To ensure
that development continues, see www.che.wisc.edu/octave/giftform.html
Instructions for unsubscribing: www.che.wisc.edu/octave/archive.html
---------------------------------------------------------------------