Re: save function

[Mike Miller]

On Wed, 9 Mar 2005, John W. Eaton wrote:

> On  9-Mar-2005, Hamish Allan <address@hidden> wrote:
>
> | I'm not entirely convinced that anyone is actually asking for
> | Matlab-compatible behaviour. All we want is an efficient way to output
> | a matrix of numbers to a file (just as it is possible to load such a
> | matrix from a file with no header information without using fscanf). I
> | suggest we *don't* aim for compatible behaviour with the braindead
> | Matlab, and instead we have an -ascii-noheader option like Mike
> | proposes.
>
> But why introduce another incompatibility?  We might as well fix this
> so that -ascii does the compatible thing (no comment headers) and
> introduce a separate option to produce the current output format.


I like that idea, actually.  The reason for introducing another 
incompatibility would be to keep Octave "save -ascii" the way it is for 
people who like it that way and are used to it, but there are advantages 
to forcing those people to change.  I'm sure someone will lose some data 
under the new system.  Octave is still young and it is a good time to make 
a change.

Making "save -ascii" do what it does now, but without the header 
information would be a nice compromise between full MATLAB compatibility 
(with columnar alignment, string-to-number conversion and dropping of 
imaginary parts of complex numbers) and current behavior.  Maybe something 
like "-ascii-annotated" would be a good name for the current "-ascii" 
feature.

No solution will be perfect and people will always disagree somewhat about 
these choices depending on their tastes and experiences.

Mike



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------