Hacker News new | past | comments | ask | show | jobs | submit login

The thing is, Category Theory and the practice of programming are too far apart. To create a reasonable application of some notion of CT in software development, the notion must be first thoroughly digested by the theoretical computer science, with the usual stress on the computational efficiency and the constructive formalism; the result of this must be a set of well-understood computational schemes - data structures and algorithms; only then one can hope to arrive to a design of a universally useful library, in a given programming language, that could withstand criticisms and the test of time. This is similar to the distance between, say, knowing Analytical Mechanics on the one hand and being able to build an efficient and economically viable internal combustion engine on the other...



Tell me about it, I've been trying to develop a complete visual metaphor to category theory for years.

Edit: you're missing that to formulating the hard- and software stack we already have in a category theoretical metalanguage solves that problem also - but that requires overcoming DRM and lacking tooling for software analysis.


Dunno if that is helpful to you, but one tip from the link related it to always think of new CT constructs in terms of partially ordered sets if possible. They make for nice visualizations and are generally quite intuitive.


The thing is that the gulf between CT and other branches of math is also there. CT maybe be generally capable of specifying any given problem but the further you branch away from it the more convenient it is for individuals to use a more domain specific language.

The simple fact is that CT is not a linga franca no matter how much advocates would like it to be.


Thanks a lot for your thoughtful warning!


True, e.g. whatever happened to SRI/Kestrel Specware?




Applications are open for YC Winter 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: