About SpT graph stuff -

The generic stuff is in the classes that reference graphs, but not code.

Some of the connected components algorithms are transcribe from
pseudocode into Smalltalk, so -
A. They're ugly Smalltalk.
B. Using hashed instead of directly indexed collections doesn't strictly
maintain the desired algorithmic speed.

But, they work and and they're plenty fast enough for my needs.

So, use them if you want, if not, and a common graph code base emerges,
I'll probably port SpT + what I need to it in the future.

Daniel

Luciano Notarfrancesco <lnotarfrancesco at yahoo.com> wrote:
> Daniel Vainsencher wrote:
> 
> >Also, in the Spaghetti Tracer, I implemented a couple of algorithms
> >(transcribed into Smalltalk from material on the web) that are
> >interesting - linear algorithms for detecting weakly connected and
> >strongly connected components. Note that Aglo graph layout (also on SM)
> >is a (quite alpha) tool for laying out graphs - I got good results for
> >laying out not-too-big DAGs.
> >
> It seems that there are many projects doing something with graphs.
> Perhaps we could work in a common codebase, or at least reuse some code 
> or ideas.
> I'll see if I can use the algorithms from the Spaghetti Tracer. I should 
> try to use Connectors and Aglo.
> 
> Joshua 'Schwa' Gargus wrote:
> 
> >I'm planning to do the minimum amount of work required to support my
> >research, but if others are interested on working on a common codebase
> >(not necessarily based on mine, which is pre-alpha), then I'm game.
> >
> It would be great to have a common codebase for graphs. Is you code 
> published somewhere?
> 
> Luciano.-