laarctags | new | comments | ask | show | place | submitlogin
Tar-pit thinking (
9 points by rain1 to plt proglangdesign programming on Feb 3, 2019 | 2 comments

Nice write-up. I agree it's a pointless argument style. Since I'm seeing his name, I threw in a submission of the Cyclone language he helped build here:

It was a C-like, safe, systems language some of whose ideas made it into Rust.


I'm glad that I almost never see this pattern.

Proving the equivalence of particular parts of different languages is really useful: it lets us look at those parts in a different way, & sometimes gives us a new, convenient way to implement useful features.

I would expect programming theorists to be smart enough to recognize that proving equivalence doesn't by itself demonstrate value. Language design is a matter of taking advantage of reframing to make previously difficult-to-imagine tasks easier to reason about. There are no 'cheap tricks' here -- or rather, cheapness is a good thing rather than a bad one. So long as a technique is effective, it's valid.

Now, there's something to be said for criticizing languages for being an internally inconsistent bundling of unrelated ideas. A language that's like this tends to be a combination of copies of different complete language systems. The problem with this kind of language is not that it's unoriginal or trivial but that, because it doesn't have a general philosophy that unifies its components, it is harder to reason about: one must reason about components in whatever chunks have internal consistency, & a grab-bag language has many small universes of consistent behavior that interact in difficult-to-predict and difficult-to-remember ways.


Welcome | Guidelines | Bookmarklet | Feature Requests | Source | API | Contact | Twitter | Lists

RSS (stories) | RSS (comments)