Lambda the Ultimate

A Generic Component Framework for System Modeling
started 5/3/2004; 10:29:24 AM - last post 5/7/2004; 5:51:59 AM

Andris Birkmanis - A Generic Component Framework for System Modeling

5/3/2004; 10:29:24 AM (reads: 171, responses: 4)

The key concept of our framework is a generic notion of transformations of specifications, especially motivated by - but not limited to - rule based transformations in the sense of graph transformation and high-level replacement systems

It so happened that during the week I had to read a dozen of papers on algebraic specifications, and their refinement and composition. It was a nice dessert to read this one, aiming to unify them all (despite the marketing-flavored title).

As a side note, it was really refreshing to see that Petri nets and algebraic specifications can be relevant to my current project - I never hoped to use anything like that for commercial applications. I only hope the theory will endure through all the dirty little troubles on the way to release :-)

Andris Birkmanis - Re: A Generic Component Framework for System Modeling

5/3/2004; 10:38:31 AM (reads: 174, responses: 0)

It was a pity the formalism didn't make it to categories, otherwise it would be double fun.

Andris Birkmanis - Re: A Generic Component Framework for System Modeling

5/6/2004; 2:02:19 AM (reads: 77, responses: 1)

To make it clear - Petri nets as such can be explored using categorical framework, e.g., see A Taste of Categorical Petri Nets, AFAIK it's the properties of transformation morphism (mapping transitions to subnets) that precluded the author from categorical reasoning.

For some reason, I feel there are other ways to bring componentization to Petri nets, without loosing expresiveness and gaining rigor at the same time... Too bad we need to release the product very soon, so there is no time for elaborating :-( Must talk product management into dualizing them into coproduct management...

Andris Birkmanis - Re: A Generic Component Framework for System Modeling

5/6/2004; 5:30:49 AM (reads: 83, responses: 0)

I feel there are other ways to bring componentization to Petri nets

Here it is, liveness-preserving transformations of Petri nets, stemming from graph rewriting.

Andris Birkmanis - Re: A Generic Component Framework for System Modeling

5/7/2004; 5:51:59 AM (reads: 47, responses: 0)

Diagonal Compositionality of Partial Petri Nets not only proposes a nice way of treating Petri nets transformations (with cool pictures) but also demonstrates use of interesting categorical tool - partial morphisms.