💾 Archived View for republic.circumlunar.space › users › dbane › gemlog › 2021-03-26.gmi captured on 2024-08-31 at 12:46:29. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2022-03-01)
-=-=-=-=-=-=-
I've probably mentioned some of these before. By "modern languages" below I don't mean chronological but instead satisfying my personal criteria. Currently ISLisp is the favorite.
I think this is a perfectly feasible way to run software projects using modern languages. It was originally for SmallTalk, but I've seen it not work so well for lower-level languages like C or Java. For better or worse though, remote work is here to stay so it needs significant adaptation to actually *write things down* for communication.
"Extreme programming" in Wikipedia
This is a great overview of how to exhaustively write software information down. However, it may be unnecessarily detailed if you're using a modern programming language. I think if the most detailed mathematical notation was replaced with "Basic English" while maintaining the large-scale structure you get most of the benefit. And also, this should be done after a working prototype.
This is the latest entry. From the outside it could look new, a way to "grow" prototypes into products without rewriting. However, it's really more like a very useful compendium of how Lisp has been developed for 50 years, with aggressive use of DSLs. NB: I'd recommend reading "Structure and Interpretation of Computer Programs" first if you haven't already.
Software Design for Flexibility
Structure and Interpretation of Computer Programs, free E-book
Although I like it, ML may not be quite "agile" enough if you don't know the requirements ahead of time (which I expect). Nevertheless, this book (and the scsh project) is a great example of how to apply modern PL technology to real problems. It may be a long-term goal, but I'd like if EISL supported UNIX at least partially following the scsh docs. I think there is still a lot of value in the UNIX programming environment.
From these, you can see the general thrust of my argument. Start with a prototype following SDF, then document following adapted SF & refine. With overarching project management adapted from EPE. I thought of very little myself :-)