Re: More octave-forge functions!!!

[Søren Hauberg]

tor, 01 06 2006 kl. 16:39 -0400, skrev John W. Eaton:
> On 29-May-2006, Soren Hauberg wrote:
> 
> | Okay, I wasn't exactly sure about the current politics. I'm not a huge
> | fan on depending on autoconf, since that mean that Octave core is
> | defined at compile time. I think all installations of Octave should have
> | a garantied basic functionality. When things are determined at compile
> | time no such things can be garantied.
> 
> I'm not sure what you mean by this.  Can you explain or give an
> example?
Reading what I wrote, I can certainly understand why you didn't
understand what I said - sorry.

So the problem is that if I download the octave sources and compile it,
the functionality of my new version of octave depends on which of the
dependencies I happen to have installed at compile time. As an example,
I'm currently missing some sparse functionality when I build the 2.9.x
series.

So this was mentioned in the topic of when to introduce new dependencies
into octave. I was trying to say the if octave "core" only supported
very basic things, then all versions would have the same functionality.
Your version of "core" octave would support the same things as my
version. In this setting it might be more easy to introduce new
dependencies, because it wouldn't be dependencies of octave, but
seperate packages.

When I made this point, I refered to my thought as brainfarting. I still
think that description is fairly good, because I haven't really thought
these things through. Perhaps I'm arguing for a solution that is hard to
maintain and hard to use...


Soren