A Question of Reasonableness

[William la Forge]

A Question of Reasonableness

How can we expect to succeed at things which others have tried and failed? To say that a grounding in Rolonics Theory makes a difference is not good enough, as Rolonics Theory contains nothing new. It is a composition of ideas, some well known, some obscure. I do not doubt that Norman Kashdan is a genius. But still there are no new ideas here. Indeed, many of the ideas in Rolonic Theory can be found in other software. Journal-centric programming is not entirely unknown. And Wagn has come pretty close to the idea of a Rolon.

Rolonics really is central to everything we are working on. Among other things it addresses systems of systems (a rolon is both a system of parts and a part of systems), which Systems Theory fails to do. The real question then is, why is there no main-stream equivalent? My best guess is that the lessons learned from software development have had little influence on established philosophical thinking. Let me explain.

Software provides us a powerful tool for testing theories by modeling them. Take a theory of astro-physics, develop a computer model and then model the life of a star. The more correct the theory and the model, the greater the correspondence between the behavior of the model and our observations about the behavior of stars. This is pretty clear cut. No one confuses the model for real stars.

But software can also be used to create new things in the world, a word processor being a good example. This is not covered by established philosophical thinking--bringing things into existence by describing them. But note that people also work a bit like computers in this regard. Memes (ideas) are spread like viruses of the mind and are the basis for social evolution. A thought can be expressed and by this means conveyed to others, and this is a creative process. Now there are schools of thought which deal with creating things through expression, but historically they have not been accepted as "good" philosophy. Cabala is one example, ontological invocation being the creation of something by naming (describing) it. G. Spencer-Brown’s Laws of Form, which was popular in the 60's (Whole Earth Catalog) being another. Unfortunately the idea of creation through expression (software, mathematics and literature being examples) is too close to traditional magic ("So mote it be"; "Abracadabra"--as I say it, so shall it be). And mainstream philosophy avoids anything with the slightest taint of magic.

So we are left with Norman Kashdan's unpublished Rolonics Theory, which I understand in small part and have used as the basis for Rolonic Software Engineering. It has taken me years just to work out a reasonable model for rolons (which Norm calls proto-rolons, as they do not have intention). I will not say that I have done a good job at this--I am still learning. But some interesting things have indeed come out of my efforts.

I see two major benefits of Rolonic Software Engineering. The first comes from having a better model of things. OO provides a very limited model for things, so every application must implement nearly every capability from scratch. The consequence of this is incompatibilities between applications at a very fundamental level. With the rich model provided by Rolonics Theory, most capabilities can be implemented in the basic model rather than being specific to an application. This gives rise to something wonderful: Everything Just Works Together.

The second benefit of Rolonic Software Engineering is more anticipated than realized just yet. Rolonics gives us a much better model for things, especially living things, in the real world. Given an appropriate UI (perhaps a desktop?), users should quickly grasp the nature of the rolons modeled in the display, as they behave more reasonably and predictably (from a user's perspective) than most other software. Since most capabilities are handled by the common model which serves as the basis for all rolonic applications, the user will not constantly be confronted with incompatibilities and capabilities which exist for one type of rolon and which do not (but logically should) exist for all other types of rolons. 

The biggest problem to date has been in developing a reasonable implementation for rolons. A rolon can easily be expressed as an XML document or even in JSON. But since a rolon carries its own history and may also serve as a very large index for other rolons, rolons can be huge. This is no problem for todays disk drives, but it does complicate the development of the equivalent to an XML DOM.