[Ktechlab-devel] simulator

[Celelibi]

2008/11/10 Glen <gcanaday at gmail.com>

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> I must be completely misunderstanding your point, or you mine. I've been
> of the impression that people have been suggesting that the specific
> equations for each model should be hard-coded into the software, not
> just the solvers! My thought on that was "Uh.. what?" haha... I think it
> turns out we are in agreement after all.


I may emitted this idea, but this discussion is more like a brain-storming
(as said [I-don't-remember]).
The gloal is to end up with something that sounds good to everyone.

We can't embed solvers in components.
We shouldn't put equations inside C/C++ code because user has to be able to
define his own one. (But, most part of user-defined component won't have
equations since they will be only a set of components and wires.)
We can't just write equations inside XML code and hope the magic solver will
find itself some advanced tricks to solve them.
We should be able to extend the solver easily because we can't have a
perfect solver that solves anything.
The component should be able to choose which solver it wants to use.


I think there are two types of components.
- Those with just a set of equations that define how it behave
- Those composed of other component.

But the user should be able to write both.


Celelibi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/ktechlab-devel/attachments/20081110/f2e49e55/attachment.html>