[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Please criticise my Octave review

From: Zdeněk Hurák
Subject: Re: Please criticise my Octave review
Date: Fri, 06 Apr 2007 16:36:10 +0200
User-agent: KNode/0.10.4

David Bateman wrote:

> A Simulink like package, that uses octave and R as the compute engine is
> available at, so your statement about absence of
> anything like simulink is not strictly correct.. When you talk about
> Octave extensions you might want to also say something like

Thanks for an interesting link, David. I was not aware of this project.
Looks interesting. Unfortunately, I must disagree with you. This has
absolutely nothing to do with Simulink. Definitely not now. But browsing
through the documentation and examples, it does not even appear that the
developers are heading towards a SIMULATION tool for dynamic systems like
Simulink. Instead, it gives me an impression that they would like to
develop some GUI for data vizualization and analysis.

Honestly, I feel that at least in some branches of engineering (signals,
controls, filters, systems,...), it is exactly Simulink that makes the The
Mathworks software suite (Matlab+Simulink+...) totally unreplaceable. It
gives TMW such a momentum, that even other proprietary sw like Mathematica
and Maple cannot compete. Simulink offers not only GUI but especially
various interfaces to real-world in the form of A/D I/O cards and real-time
extensions, code generation and lots of other engineering stuff. 

If I need to compute eigenvalues of a huge matrix, or FFT of some
time-series of simulate some nonlinear differential equation (given by some
formula), I can find lots of replacement for Matlab (Octave, R, Scilab,
Numpy+Scipy, Freemat, ...). But if I need to simulate some dynamic system
for which I cannot even write its differential equations (instead, I can
build the model of dynamics "graphically" by putting together basic blocks
representing some subsystems), I must desperately admit that there is
really not even a modest competitor to Simulink even in the proprietary
world (well, there is Vissim and perpaps you can
google a few more but I have never ever met a single person using them).
Well, there is a Simulink clone in Scilab, called Scicos, but I think it
will take developers some time to bring it to a mature state.

Sorry if this sounds negative, but I think that it is constructive to be
fair in this. I did not want to lower the value of work invested into
Octave and I keep my fingers crossed for the project (I don't use Octave
often), but at the moment, there is really no reasonable alternative for
Simulink for Octave users.

Zdenek Hurak 

reply via email to

[Prev in Thread] Current Thread [Next in Thread]