[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Bug when dividing a scalar by a vector?
From: |
Sebastian Schöps |
Subject: |
Re: Bug when dividing a scalar by a vector? |
Date: |
Thu, 12 Jan 2017 17:51:18 +0100 |
> there was a similar discussion some years ago:
>
> http://lists.gnu.org/archive/html/octave-maintainers/2008-04/msg00160.html
Thanks for bringing that up! This further suggests to keep Octave's current
solution and not aiming for further Matlab compatibility. However, the user
might have no clue whatsoever that octave and matlab disagree.
> Under "doc /" I see
>
> If the system is not square, or if the coefficient matrix is
> singular, a minimum norm solution is computed.
Ok, that is pretty clear. I did not find a similar statement in the Matlab
docs. The only page is the one on mrdivide and this does not discuss the
underdetermined case,
https://de.mathworks.com/help/fixedpoint/ref/mrdivide.html,
>> I checked: no, Matlab does not warn in the case of 1/x. It does warn
>> if you solve a singular equation system ("Matrix is singular to
>> working precision.") but it does not mention that a least squares
>> solution will be returned. Older version of Octave were more explicit,
>> right? I liked that but it seems that I missed that it was changed.
>
> In Octave 3.6.4 I get the same answer with 1/[1;1;1;1] and no warning
> message.
I thought that octave was saying something like "... computing least squares
solution" in former times. Maybe I remember wrongly.
Sebastian