Lambda the Ultimate

...interesting project that I've watched for years. The only implementations have been commercial. The new open-source version has a hybrid license, I think (free for noncommercial use). Correct me if that's wrong.

Some new features in the very, very recent release of Mathematica 5.1 make it appealing for simulation work. The new piecewise overhaul comes to mind. Modelica has always been somewhat biased in a Mathematica direction. Hence the similar name, I suppose. Of course it is a stand-alone language in its own right. It cries out for a free-for-all-uses implementation in someone's favorite open-source language.

is there a way of bumping forum posts to the front page? maybe it's not necessary, given the sidebar? if people have cool links like this, perhaps it's worth considering the front page, given the current lack of an ehud...?

A couple of comments:

While it is true that the main implementations of Modelica are commercial, the number of implementations is slowing growing and I know several groups working on open source implementations (e.g. see the preliminary program for Modelica'2005 at http://www.modelica.org/events/Conference2005/program).

The overview that is referenced in the original posting is pretty out of date although I'm not sure (off the top of my head) if there is something more recent. The basic idea is the same, but Modelica 2.0 included some important new features and Modelica 2.2 is due to be released this spring.

I would not agree with the contention that Modelica is very heavily influenced by Mathematica. While one of the commercial implementations, MathModelica, attempts to fuse the two technologies, my understanding is that the language design itself was originally based on concepts from "A Theory of Objects" (http://www.luca.demon.co.uk/TheoryOfObjects.html). The similarity in name is mainly coincidence. The use of "model" in the name is obvious and the "ica" had the advantage that it is an actual word (in Spanish I believe...meaning "warrior" if I remember correctly) but had not been trademarked.

Finally, I could not agree more that "[i]t cries out for a free-for-all-uses implementation". I have been actively encouraging such developments for many years. There are really two main hurdles. First there is the implementation of the Modelica semantics (the elaboration process that leads to a fully specified mathematical problem). This has been undertaken by several groups, but I haven't seen something I would call "complete" (i.e. implementing the complete specification). The other piece of the puzzle is the simulation side. In order to simulate elements in the standard library some fairly sophisticated structural algorithms are necessary to reduce the models to state-space form for ODE solvers. This is another area where specific expertise is required. That being said, I am optimistic that such functionality will eventually emerge in a "free-for-all-uses" form.

Those interested in the language may like to visit the web site or attend the upcoming Modelica'2005 conference (see the website).

Finally, if anybody has any questions about Modelica, feel free to contact me.

--
Michael Tiller
Secretary, Modelica Association