Re: norm, 'fro' - warning: division by zero

[Jaroslav Hajek]

On Fri, Sep 26, 2008 at 10:05 PM, Jaroslav Hajek <address@hidden> wrote:
> On Fri, Sep 26, 2008 at 9:24 PM, Zbigniew Komarnicki <address@hidden> wrote:
>> Hello,
>>
>> Please see on this code in octave:
>> octave-3.0.1 and octave-3.0.2 in Debian lenny.
>>
>>
>> Here is an example:
>> =========================
>> octave:1> A=zeros(5)
>> A =
>>
>>    0   0   0   0   0
>>    0   0   0   0   0
>>    0   0   0   0   0
>>    0   0   0   0   0
>>    0   0   0   0   0
>>
>> octave:2> norm(A, 'fro')
>> warning: division by zero
>> ans = NaN
>> octave:3>
>>
>>
>> But when, according to the manual (help norm) write in this way:
>>
>> >From help system:
>> P = `"fro"'
>>           Frobenius norm of A, `sqrt (sum (diag (A' * A)))'.
>>
>> =========================
>> octave:5> sqrt (sum (diag (A' * A)))
>> ans = 0
>> octave:6>
>>
>> Here no warning 'division by zero' and answer is 0 not NaN.
>>
>> It is a bug in 'norm' implementation or something else?
>>
>
> It seems as a problem of the current norm implementation.
> The following patch
> http://hg.tw-math.de/highegg/rev/3c8c55dde77b
> reimplements norm adding a number of new capabilities, and does not
> suffer from this problem. I guess, however, that this particular issu
> is worth fixing separately.
>

OK, I see this is not present in the development branch, so I'll fix
release-3-0-x separately.

>
>> Thank you, for any comments.
>>
>> Best wishes,
>> Zbigniew
>>
>> _______________________________________________
>> Help-octave mailing list
>> address@hidden
>> https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
>>
>
>
>
> --
> RNDr. Jaroslav Hajek
> computing expert
> Aeronautical Research and Test Institute (VZLU)
> Prague, Czech Republic
> url: www.highegg.matfyz.cz
>



-- 
RNDr. Jaroslav Hajek
computing expert
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz