4 ms·
I feel like many of us are searching and probing around the same space. It definitely seems to be the case that existing programming languages are way too rigid
by breckinloggins 12y ago
I feel like many of us are searching and probing around the same space. It definitely seems to be the case that existing programming languages are way too rigid, but that the current popular example of "just make the language whatever you desire" (Lisp) is not quite right, either. It's close, but just far enough away that it seems alien.
I'll definitely be paying closer attention to this as it agrees with a lot of hunches that I have about what's wrong with programming as we currently practice it.
Thanks so much for the link. I'm taking notes :)
EDIT: Anyone who likes Avail, and particularly the points made in the History section, should probably read CTM. [1]
[1] http://www.amazon.com/Concepts-Techniques-Models-Computer-Programming/dp/0262220695 http://www.amazon.com/Concepts-Techniques-Models-Computer-Pr...
- smt88 12y agoIn the web space, we're all actually looking for languages that are more rigid (rebelling against PHP, JavaScript, Ruby, etc.)
- breckinloggins 12y agoTrue, but in a language that's designed to quickly make other languages, the rigidity can be controlled by the system architects. An example of a current language with this philosophy is actually Typescript. The gradual typing means that you, as a module designer, can choose to make rigid exactly what you wish (and choose how much rigidity you require). If the consumer of that module also uses Typescript, your rigidity requirements will be respected. The reason we don't think of programming languages as being per-project, per-module, per-domain, etc, but rather as large global things is because we don't currently have a clue how to do that (and if we did, we don't have a clue how to make the tooling work). If and when we do, these types of conversations will seem quaint. One can hope, anyway...