The article doesn't get very concrete, and given the origins of "'pataphysics" as parody I'm not sure how seriously to take this.
On the other hand, treating software systems as realities in their own right – with only incidental mappings back to our reality – is a fascinating idea. (Obviously an entirely self-contained reality is no good; I/O are pretty fundamental.) I guess the question is, what happens when we stop requiring software model things, and instead let it be things?
Based on the prose examples the article gives (the chess/lunch story and the crucifixion/bike race story), it looks like a truly 'pataphysical system wouldn't need IO at all. Instead, you'd have a program of some sort, that was simultaneously doing 'pataphysical computation and mundane computation. In the story of Jesus and the bike race, the narrative never shifts between the story of racing up the hill and the story of Jesus being killed on a cross, the story is both at the same time throughout. By analogy, I suspect the article would like us to consider what kind of system we'd need to construct in order to, I suppose, be a story about a bike race (or something else equally removed from programmer's usual domain) and an application to serve web pages at the same time.
Of course, given 'pataphysic's status as "the science of imaginary solutions" I won't be holding my breath for a 'pataprogrammatical Rails-killer anytime soon, but it's a fun idea.
I did have the thought that if OO is clearly metaphorical, it could be argued that all abstractions within the program are 'pataphysical. That is, we can say "an agent sends a message to another agent" or "the function passes control to the current continuation", and we're actually describing concrete reality in program-land, rather than metaphors for real-world objects.
In that sense, 'pataphysical programming might be more common than we realize.
A "pataphysical program" would be an object diagram which took on a life of its own, loosing all touch with possible implementations, with considerable documentation, a CVS tree, a length license agreement, etc...
This doesn't make any sense to me. Either there's some assumption about self I don't get, or they're just making shit up. Cuz after reading that, I'm. No closer to reproducing it.
On the other hand, treating software systems as realities in their own right – with only incidental mappings back to our reality – is a fascinating idea. (Obviously an entirely self-contained reality is no good; I/O are pretty fundamental.) I guess the question is, what happens when we stop requiring software model things, and instead let it be things?