Hacker News Reader

Comment by utx00 | Hacker News Reader

could lisp be blub for something else? if yes, what is that something. if you think not, why not?

X is blub if there is a language which is more powerful than X and you can't understand that the "extra stuff" in that language is actually power, as opposed to meaningless stuff.

For example, an argument could be made that Lisp is a blub, because of the libraries problem. I'm not a lisp user myself, though I'm learning Scheme, so I don't know for sure if the libraries problem is an actual problem. But if it is, you could say that Lisp is a blub for Python, because while Lisp users are spending their time inventing a DSL that's incompatible with everything else, Python users type 'import awesomelibrary' and get the job done.

So, really, it depends on your notion of power. In pg's original Blub essay, power is Lisp-y-ness (or macros, perhaps), so by definition Lisp can't be a blub. But if your notion of power is something else, maybe it can be.

Specifically, the Blub idea applies to language power. You could build python's libraries in Lisp, but you can't build a macro in Python without changing the language.

On the other hand, Lisp isn't at the top of the theoretical Blub hierarchy. Users of erlang can to some degree look at lisp as a Blub - what do you mean there's no lightweight concurrency, and you can't start millions of threads and have the messaging just managed? Lisp is broken! A language with Lisp's (lack of) syntax and erlang's concurrency might be able to look down on both.

"You could build python's libraries in Lisp"

Then why doesn't someone do so?

Just because one could do something, doesn't mean one will do something. I'm not arguing you should use Lisp over Python, merely that the strength of libraries is not a language feature per se. Clearly, it's one of the most important things to consider when picking a language, but it's not a feature of that language any more than how helpful the people in that language's IRC chat room is, or how good the documentation.

Have you ever been to common-lisp.net? It's no CPAN, but same idea. In practice you've got everything you need to work with, no need to reinvent the wheel.

I think (today's) Lisp could be blub, but only if Lisps were to become the standard category of programming languages. Then some other more advanced meta-lisp that is a few layers of abstraction up, and applying some yet-undiscovered principles of computer science in its high-level semantics would be at the cutting edge, and the vanilla lisps would be mundane and practical-seeming.

For some some Lisp-developers Lisp is blub relative to Haskell.

how come? could you give an example?

I've met Lisp-programmers which simple didn't understand the point of Haskell - all those arbitrary constraints (no IO, no mutation) which just required elaborate workaround (monads) to get back to square one. This makes Lisp "blub" (for them).

Reply on news.ycombinator.com