Re: Using the RM for ADTs
Date: Mon, 6 Jul 2009 05:18:26 -0700 (PDT)
On Jul 6, 4:17 pm, Roy Hann <specia..._at_processed.almost.meat> wrote:
> Brian Selzer wrote:
> > I think that this is the perfect counterexample to some of the myths that
> > are perpetuated here on cdt. It illustrates valid uses of both artificial
> > identifiers and nesting. I think one would be hard pressed to come up with
> > a solution that doesn't involve the use of artificial identifiers. The
> > nesting can, of course, be dispensed with through the introduction of
> > additional artificial identifiers.
> I think it is a perfect example of wondering why chisels don't make
> good screwdrivers.
Are you suggesting the RM is poorly suited to representing circuit values? I think this is an important question. I don't see much evidence that the RM is particularly useful here.
For example, to perform a nodal analysis one may associate a voltage with each labelled node and use Kirchoff's voltage and current laws to create a large system of simultaneous equations. This in turn may require sparse matrix techniques for efficient computation. I don't see anything in the RA that will help do numerical analysis.
In fact I'm not entirely sure why one would want to use the RA at all. Is it useful to find all the resistors connected to a 5pF capacitor? If pattern matching is useful I expect the RA isn't as appropriate as some kind of graph based unification technique.
Nevertheless, complex circuits may contain millions of components and as data it needs to be managed - i.e. there needs to be a DBMS providing persistence, multi-user support, data entry validation etc. Received on Mon Jul 06 2009 - 14:18:26 CEST