Re: fsolve test failure

[Jaroslav Hajek]

On Fri, Jan 30, 2009 at 10:25 PM, Ben Abbott <address@hidden> wrote:
> On Friday, January 30, 2009, at 03:50PM, "John W. Eaton" <address@hidden> 
> wrote:
>>On 30-Jan-2009, Jaroslav Hajek wrote:
>>
>>| Technically, there is no incompatibility.
>>
>>I'd say, that technically there is, because optimset returns different
>>things.  But in practice, I agree that you are right that this
>>probably won't matter.  I seem to keep forgetting that even though
>>Octave's optimset returns an empty structure, you can add fields to
>>it.  So if someone has done this in some Matlab code, it will stil
>>work with Octave.  It only matters if someone calls optimset and
>>expects to see a get a non-empty structure and then wants to do
>>something with those fields.  I guess this won't happen often.
>
> Actually, I often make use of the structure returned by Matlab's optimset ... 
> but only because I forget the name of the fields I need to set.
>
> Not much of a compatibility issue, but I find it useful.
>
> Ben
>

Yes, this is useful, and you can query for individual functions'
options easily. The problem is Matlab always throws in all possible
options, no matter whether they make sense for the solver you want to
use.


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