Wednesday, September 26, 2007

Introduction to Existential Programming Blog

This first entry for "Existential Programming" is meant to act as an introduction to the topic and an explanation for the backdated entries to be added over time. Around May, 2006 I had a series of brain farts (ahem, epiphanies) about computer software engineering that led to a theory I decided should be called "Existential Programming". Because I have now collected about a year and half of unedited notebook entries, and I now have enough of an idea of what I mean by Existential Programming to write it up, but because I don't yet have the time to polish a Manifesto into a magazine article, much less an academic paper, I have decided to start this blog and back fill it with my notebook contents, as well as, putting future entries here. The goal is to plant my flag on the topic and its ideas now, even if I can't write "Existential Programming, the Book" yet. [BTW, see my std blog disclaimers.]
A meta-idea above Existential Programming itself, is the more general notion that there should be much more explicit cross fertilization between ideas from Philosophy (with a capital "P") and ideas from software engineering (as practiced in industry). I call this "Philosophical Programming". Early on during my epiphanies, I had the intuition that Philosophy probably had something to say about my topic (even though I'd never taken a philosophy class). So, at age 50, I started reading Philosophy 101 books. It quickly became obvious that Philosophy has SO MUCH to say about data/class modeling topics that it is criminal how little explicit reference to it there is in the software practice literature. I distinguish between industry practice (and their books, blogs, magazines oriented towards tools and "best practices") versus academia. After I learned enough terminology to search for papers covering ideas similar to mine, I found that there is a whole subculture writing academic conference papers that don't really bleed over into industry conferences for things like Java or AJAX or SOA. So, my general "project" these days is to try to come up with practical application techniques based on otherwise esoteric topics.
What is "Existential Programming" and what are the central ideas associated with it? In a nutshell, it means to embrace the notion that "existence precedes essence". I.E. develop data models, object models, programming frameworks, etc. without imposing a single E/R model, OO class hierarchy, ontology, etc. By using techniques where "objects" and "entities" exist independently of a single "strong type", they can integrate multiple "strongly typed" data models by letting objects simultaneously reflect all those models. This differs from weakly typed programming, or completely type-less programming, as can be found in scripting languages like JavaScript. Ironically, it takes a "type-less" foundation to really seriously do strong types in a "multi-cultural" world. [And actually, JavaScript is not a bad platform to implement these ideas precisely because of its class-less orientation.]

The ZEN thought here is that until one can create a type-less object, one can not create an object which can be all types simultaneously.

The ideas that flow from this general notion (or more correctly, caused me to refactor the general notion out of these specific ideas), fall into general categories like:

  • mixins
  • data integration using semantic mapping
  • code integration of uncooperative APIs and Frameworks
  • promoting "roles" over "is-a subclassing"
  • open classes and external methods
  • ontology mediation
  • object evolution and class/API versioning

1 comment:

  1. The "pretty" version of this is at

    [ Note the single "m" in ...programing... ]