[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: new matlab features
From: |
Paul Thomas |
Subject: |
Re: new matlab features |
Date: |
Thu, 11 Mar 2004 23:44:09 +0100 |
User-agent: |
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 |
John,
I am sure that you are right about the calls for features that octave
lacks, relative to R14. Oddly, though, I sense a desperation in this
pushing of features that very few, if any, of Mathworks customers will
be interested in. To me, as a heavy Matlab user, its strengths are
epitomised in:
Development Environment
This is a very slick part of Matlab. One can live without it but life,
sure as heck, is easier with it.
Performance and Platform Support
Here the improvements obtained with the "JIT" compiler are truly
impressive in some intellectual sense. I am not sure that I understand
its extension to all types in R14, in that the extension must be
marginal in terms of usage. In fact, Matlab R11 is already a factor 10
times faster than octave-2.1.56 for F77-type code. From the point of
view, of every day use, this was already enough; I do not detect any
significant difference in the jobs that I am doing. If we could get to
the bottom of where this difference, pre-JIT, comes from, I think that
octave would be immeasurably improved.
Graphics and 3-D Visualization
This, from my point of view and that of all my colleagues, is the true
strength of Matlab, since version 11-ish. Not only are the graphics
good but the GUI is excellent.
The conclusions that I would draw from the appearance of R14 are:
1] The small bunch of volunteers who are developing octave are doing
fine relative to the hundreds doing R14.
2] The small bunch of volunteers who are developing octave should be
damned proud of themselves.
3] There is perhaps something symptomatic in the octave graphics list;
an anouncement that "a handle graphics project has started" stopped the
list in its tracks.....
4] The will of small bunch of volunteers who are developing octave is
such that if they worked together more, they could give Mathworks a run
for their money in the above three areas (in reverse order of priority,
I think).
Perhaps the call that John made to improve basic tests could be a
clarion call for more collective effort to develop octave? As for the
gripers... stuff-em! As my forebearers from the north of England would
say, "You don't get owt for nowt!"
Paul T