Functional programming is a ghetto | Hacker News Reader

Functional programming is a ghetto

blog.jelastic.com

35 pointsjjohns20 comments

I wonder if the original author knows his content is being used like this. I found the post very misleading in the way it grossly misrepresents the authorship of the piece. You have to scroll all the way to the bottom to know that Michael O. Church [http://news.ycombinator.com/user?id=michaelochurch] wrote this and not Judah Johns.

It might be appropriate to change the HN link to http://michaelochurch.wordpress.com/2012/04/24/functional-pr...

I guess I should have made that more clear. I'll change it so that it is. I just really liked the article.

Sorry OP, but your defense of ghettos as voluntary enclaves is not accurate. Please check: http://en.wikipedia.org/wiki/Ghettos_in_Nazi-occupied_Europe

They were also not even a tiny bit prosperous.

Absolutely: "The term was originally used in Venice derived from the word Borghetto, meaning Little Borgo, a cluster of homes and buildings often outside Italian city walls, to describe the area where Jews, tradespeople or agricultural workers were compelled to live."

As you said, the author was probably mixing up ghettos (which are by definition involuntary) with the concept of an ethnic enclave ( http://en.wikipedia.org/wiki/Ethnic_enclave ). It's easy to do because many ethnic enclaves were historically ghettos, but the involuntary factor was removed and the region developed a unique culture that remains.

To make it even more confusing, I've heard that at least some ghettos in German-occupied Poland were ethnic enclaves before the war. OTOH, never cared to find actual sources to support/disprove this claim.

The OP made the crucial mistake of not researching holocaust related material prior to posting. Even if the Jewish citizens of Europe segregated themselves for their own interest rather than being forced to, today’s widely accepted definition of Ghetto comes from either poor urban areas or in reference to the Jewish ghettos set up by the Nazis.

That being said, the OP does make a good point that the programming industry is a ghetto in the sense that culturally it has a particular way of thinking and approaching the world. What really isn’t interesting about it, is that almost every profession has the same thing. HN is an example of an internet ghetto in the form of a place that allows its users to only associate with each other if they choose even if they are geographically separated.

In 2011 the SBS talked about how broadband could lead to internet ghettos and I believe they have a point, but the question is: Is an internet ghetto really a bad thing? http://www.news.com.au/technology/broadband-could-lead-to-di...

the idea is to capture your real intent and formulate it as an expression which can then be evaluated (on demand or lazy) and chained with others ...

this requires you to focus on how to solve the problem succinctly rather than blindly wiring up factories that eventually get you to the 2m line codebase mentioned below....

if that looks like a ghetto to you i wonder what other crazy ideas you have...

Clickbaity title obviously meant to invoke the infamous "Rails is a ghetto" rant.

This reeks of ignorance of the industry, gross generalizations, and a ham-fisted call to action.

People don't use Java because they don't care about programming, or because they don't care about maintenance. People use Java (or PHP, or C#, or Perl, or C/++) because:

1. They get paid to do it. Historically, someone has realized that Java development was easier/faster/more stable than Lisp development.

2. Java has hella libraries. I mean hella. You can find a library for almost anything. Lisp...not so much.

3. Java is a language for doing things. Lisp is not a language for doing things. I'll change my mind when I see a 2m+ line codebase that's been in production for ten years.

> What we call functional programming is somewhat of a shibboleth for good-taste programming.

Your blog post is littered with drivel like this. I could quote almost every other line for the same reason.

1. They get paid to do it. Historically, someone has realized that Java development was easier/faster/more stable than Lisp development.

???

2. Java has hella libraries. I mean hella. You can find a library for almost anything. Lisp...not so much.

Clojure?

3. Java is a language for doing things. Lisp is not a language for doing things. I'll change my mind when I see a 2m+ line codebase that's been in production for ten years.

Emacs?

???

Are you familiar with the concept of worse is better?

A 2m+ line Java codebase is not desirable or in any way positive. With a Lisp (or Haskell or some other nice functional language) you wouldn't need half the amount of code to achieve the same functionality. Java is especially notorious for its boilerplate. Lisp is very much a language for doing things, at least as much as Java. A big codebase that has been in production for ten years? How about 40+ years? http://en.wikipedia.org/wiki/Macsyma

It's your post that reeks of ignorance and gross generalization.

You're trying to give me a 15 year old application and hold it up as a shining example of Lisp engineering, when the most prominent application in that arena (Mathematica) is written in C and Java?

Reddit, too, was once written in Lisp. It was rewritten for a reason.

> 1. They get paid to do it. Historically, someone has realized that Java development was easier/faster/more stable than Lisp development.

http://paulgraham.com/avg.html

> 2. Java has hella libraries. I mean hella. You can find a library for almost anything. Lisp...not so much.

http://www.quicklisp.org/beta/releases.html

And that's just Quicklisp. To say Lisp doesn't have libraries for almost anything is pure ignorance.

> 3. Java is a language for doing things. Lisp is not a language for doing things. I'll change my mind when I see a 2m+ line codebase that's been in production for ten years.

Emacs has been in "production" since at least 1985, and is 650,000 lines of elisp and 318,000 lines of C (as of version 21.4, according to `wc`). The vast majority of its functionality is written in elisp, and Emacs is arguably more powerful than any editor or IDE written in Java. It's also more stable, easier to maintain, and new functionality can be added with only a handful of lines of code.

To be blunt, it does more in 600k lines than any 2m+ Java project ever will.

1) anecdotal self praise. That a startup (one!) that got acquired used Lisp doesn't prove anything. Many many more succesful startups did not. And its not like it was the most intriguing technologically or acquisition-value wise.

2) "A library for almost anything" (and lot of them abandoned or half done) != multiple libraries for everything and then some.

3) I wouldn't put forward elisp as an example of a good functional language. Actually it's crap. And Emacs is not more powerful that Eclipse or IntelliJ, except if you mean "has lots of hohum plugins and ome good ones".

> 1) anecdotal self praise. That a startup (one!) that got acquired used Lisp doesn't prove anything. Many many more succesful startups did not. And its not like it was the most intriguing technologically or acquisition-value wise.

That's not the point. The point was that they had a product that they could improve more rapidly than their competitors; a fact that they readily attribute to it being written in Lisp. Just about any other working Lisp programmer probably has a similar story.

And I do not put forward elisp as a "good" language. It is a Lisp, and the success and longevity of Emacs is a testament to its Lisp underpinnings.

Now submitted in its own right: http://news.ycombinator.com/item?id=4285956

Thanks, Colin. Should have done that to start with. I really liked Michael's post and thought it could use some new visibility--which is why I wanted my blogs readers to see it (mostly Java guys).

Reply on news.ycombinator.com