Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

My Increasing Frustration with Clojure

ashtonkemerling.com

Post reply on HN
Print view View original

21–30 of 240 posts

Re: My Increasing Frustration with Clojure

#21

Post by catnaroek »

sethev wrote: post #18
catnaroek wrote: post #11

There are basically four possible attitudes to abstraction: (0) We don't need no stinkin' abstractions! Obviously not befitting a high-level language like Clojure, move along. (1) We enforce abstractions in our heads. Running into an undocumented scenario is undefined behavior. (Yes, in the C sense.) (2) We enforce abstractions via compile-time checks. Clojure won't have this anytime soon, move along. (3) We enforce…

It seems disingenuous (in the blog post) to act as if Rich was saying that the results of applying union to a list were correct when he was clearly saying that the fact that it produced any value at all was an implementation detail.

I think you misread the author's complaint. He's complaining that the `union` abstraction is leaky - it exposes its own ad-hoc implementation choices.

Re: My Increasing Frustration with Clojure

#22

Post by eudox »

rurban wrote: post #12

The underlying issue to me seems to be that Clojure is still not developed by a team, but by a single developer, who cannot think out of the box. union and intersect not doing the same thing as in common lisp (accepting lists and vectors), and returning buggy values instead is just a bug, even if the developer didn't have that in mind originally, and doesn't like to support that.

>union and intersect not doing the same thing as in common lisp

Clojure is a different language.

Re: My Increasing Frustration with Clojure

#23

Post by _pmf_ »

dustingetz wrote: post #20

I think this article is spot on. But. I think the root cause here is that many of the tools are build by the community. Great tools cost a lot of money to build, and you need to build the right thing at the right time, the cost needs to be spread across the comminuty. The community is growing and tools are improving. The first generation tools with all their rough edges are being replaced by simpler tools that have i…

I think the author is right in his assessment that it's a community problem: the Clojure core team does not want outsiders to participate in the future of the language, but does not clearly communicate this up front. I absolutely support the right for them to do this (I'm a happy user of the non-open C# language, after all), but they should clearly state that Clojure is defined by Cognitect, not the community.

Even one of the most prolific Clojure users with a tremendous track record of extremely well engineered and practically useful libraries, Zach Tellman, has been bitten by Rich Hickey's / the core team's NIH syndrome[0]. That's no way to treat the people who love your language.

[0] http://dev.clojure.org/jira/browse/CLJ-1517

Re: My Increasing Frustration with Clojure

#27

Post by eric_h »

> The thing I am trying to build is not possible in any other ecosystem today

Umm... what? I'm afraid I don't understand how the "grand vision" behind Clojure adds pixie dust that makes it capable of something another Turing complete language isn't. Care to elaborate?

Re: My Increasing Frustration with Clojure

#28

Post by sethev »

catnaroek wrote: post #21
sethev wrote: post #18

Earlier quoted context omitted.

It seems disingenuous (in the blog post) to act as if Rich was saying that the results of applying union to a list were correct when he was clearly saying that the fact that it produced any value at all was an implementation detail.

I think you misread the author's complaint. He's complaining that the `union` abstraction is leaky - it exposes its own ad-hoc implementation choices.

I think you may have misread it - he's pretty clear that he considers these bugs that clojure core chooses to ignore because they either a) don't understand them b) want to do cool new stuff

Re: My Increasing Frustration with Clojure

#29

Post by twoodfin »

_pmf_ wrote: post #23
dustingetz wrote: post #20

I think this article is spot on. But. I think the root cause here is that many of the tools are build by the community. Great tools cost a lot of money to build, and you need to build the right thing at the right time, the cost needs to be spread across the comminuty. The community is growing and tools are improving. The first generation tools with all their rough edges are being replaced by simpler tools that have i…

I think the author is right in his assessment that it's a community problem: the Clojure core team does not want outsiders to participate in the future of the language, but does not clearly communicate this up front. I absolutely support the right for them to do this (I'm a happy user of the non-open C# language, after all), but they should clearly state that Clojure is defined by Cognitect, not the community. Even o…

What's the problem with [0]? I am not a Clojure expert, nor have I looked in detail at the patch, but Rich's concerns about megamorphic call sites in library code sound plausible.

Re: My Increasing Frustration with Clojure

#30

Post by catnaroek »

eric_h wrote: post #27

> The thing I am trying to build is not possible in any other ecosystem today Umm... what? I'm afraid I don't understand how the "grand vision" behind Clojure adds pixie dust that makes it capable of something another Turing complete language isn't. Care to elaborate?

Turing-completeness is a statement solely about a system's ability to perform computation, in the mathematical sense of the term (take inputs, do something for a while, return outputs, strictly in that order). It says nothing about other desirable properties like usability, extensibility, comprehensibility, etc.
Post reply on HN