Re: DHARMI project

[Telford Tendys <telford@triangle.triode.net.au>]

>   The aim of this project is to create a language that is well suited to
> writing scientific models,  and hooking them up like tinker-toys.  The
> hooking up can done via command line (included in every programs'
> interface) or  visually via drag and drop.  I do not intend for this
> project to rely on the GUI,  but  I still want it to have a rather nifty
> one.
> 
> It is called DHARMI (Data Habitat And Rapid Modelling Infrastructure)
> and can be found at http://www.erols.com/panda/

Sounds a bit like ``Lab View'', although AFAIK, Lab View is totally
dependent on its GUI, possibly there is some underlying linkage language
that users don't get to see. Although Lab View is nothing to do with free
software, it is becoming very popular and it does run well, beating it
would be a tough job.

I'm putting together some numerical analysis stuff using guile as
the high level scripting language and C for anything that requires
speed. So far I've got enough matrix engine to solve about 20000 linear
equations in a P150 in a few minutes (roughly a power-law of 1.4
for more equations). I've also knocked together some database B-tree
stuff to try and organise data by `entities' and `attributes'.

When you say `scientific models', what do you mean exactly?
Most of my models are material equations, some of which have some
memory associated with them. Each is related to a particular context
and `hooking them up' usually requires quite a bit of thought
(but not too much code) to get sensible results.

One thing that I have noticed is that using `cons' is very expensive
with regards to speed. Firstly, cons is quite slow and secondly,
the garbage collector turns up more often if you use cons. The trouble
is that almost everything uses cons somewhere or other. Does anyone
know how to get around this?

	- Tel