I started learning Clojure because it seemed to be the logical next step for me. In the few years that I've been programming full-time, I've noticed some patterns: in general, imperative code with lots of side effects in it is difficult to reason about and hard to test, and pure functions are easy to reason about and trivial to test. The same rule seems to hold for code that uses mutability vs code that deals only with immutable things.
Over the past couple of years, I've learned Common Lisp by working through Conrad Barski's Land of Lisp[0] and dabbled in Haskell via Learn You A Haskell[1]. I had a great time learning both, but I found that Common Lisp had a decent number of warts (e.g. four different kinds of equality) that seemed to be primarily due to backwards-compatability concerns, and I had a really hard time getting up and running when trying to write a simple webapp in Haskell. In addition, I've gotten spoiled by how batteries-included Python (which I use at my day job) is; the library scene for both of those languages seemed lacking in comparison.
What I like about Clojure is that it gives me fast immutable data structures, has a really strong emphasis on functional programming, and has access to a really huge amount of useful libraries.
So - I've mainly been learning Clojure as a nice treat for myself; the amount of jobs that exist that employ programmers to code specifically in Clojure didn't even really occur to me as something worth thinking about. It's just really a lot of fun to program in. Also, with any luck, I'll be able to transfer any lessons I learn from Clojure to my daily work in Python.
(Haskell still has a very special place in my heart [I do frequently find myself wishing I had types], I'm not trying to dis Common Lisp nor Haskell, I'm just trying to honestly answer the question of why I thought it was worth learning Clojure).
I realize this is quite off the main topic, but I thought it was interesting and worth passing on.
If you really think that "four different kinds of equality" in Common Lisp is a wart, then I recommend that you read "The Best of Intentions" (http://www.nhplace.com/kent/PS/EQUAL.html) by Kent Pitman. The article explains why that situation is quite intentional, and why it is very rational, and why it is in fact ill-advised to expect a single meaning for "equal" ... or, in fact, other "obvious" concepts like "copy" ... to be sufficient.
That article was one of the many over the years that helped me refine the way I think about programming and programming languages.
What do your superiors think about implementing new, small projects in Clojure?
When you say "has a huge amount of useful libraries" but then say that Python is "batteries included", which libraries in Clojure are particularly useful? Which ones did you wish it had?
2) I used those two phrases interchangeably - I consider both Clojure and Python as "batteries-included" languages. Case in point - before I was aware that Clojure Toolbox existed, I was able to google "clojure http request" and find several libraries that let me perform HTTP requests programmatically that I could include in my project.clj and start using right away. My memory's a little fuzzy, so apologies if this is completely wrong, but I don't recall having as easy a time doing that in e.g. Common Lisp or Haskell.
As far as which libraries in Clojure are particularly useful - Ring and Compojure are fantastic, and core.logic[0] is extremely interesting; I'll write a blog post showcasing core.logic (and using it for my first time ever) at some point. Midje is great for writing unit tests. I've only been using Clojure for a few months, so I can't point to too many others that have really been extremely useful for me, but I definitely recommend checking out the Clojure Toolbox[1] for a more thorough list of the main contenders.
To me, this is the smartest part of using Clojure in your projects. If you're a company with an enormous Java stack, it's trivial to add things like Clojure or Scala and use them as scripting languages. Or break out a library and rewrite it using one of the two, if that problem would be better solved more functionally.
Basically it makes it easy to convert projects without tearing down your infrastructure, ie, easy to convince your boss.
This post is specially relevant for me, since last weekend I've decided to take a quick look at Clojure, as one of my personal goals this year is to get acquainted with a Lisp dialect.
From my _very_ short experience (I've started coding a simple game), here are my two cents:
- If your application is heavily dependent on state, which changes continuously and must be maintained during most of the execution time - like a game - it may not be a good fit. You'll waste a lot of working around the fact that each modification results on new objects.
- If your application comprises mainly of short requests that may create some state that will most certainly be discarded shortly after - like a web application - then I think it's a good fit and you'll probably enjoy it better.
Once again, I've just got started with it, so take this with a grain of salt :)
I'd like to contrast this: if your application is heavily dependent on state, especially state which is shared between threads, Clojure offers a concise, safe, and uniform way to understand and organize state transformations--at the cost of performance. Where fast mutability is required, it's easy to drop down to explicit locks, atomics, java.util.concurrent collections, et al.
I find the majority of my time in writing stateful programs goes towards reasoning about safety, and having access to STM lets me write correct code with less thinking. Then I focus on optimizing the parts where performance is critical.
At ILC 2009, Guy Steele asked Rich Hickey how Clojure's STM interacted with I/O. To which, Rich replied that it doesn't. No further discussion continued on the subject.
I have found that this rarely enters discussions regarding Clojure, although the STM is frequently lauded.
GHC's STM also does not work with the IOMonad, which is clear because of the types, and is articulated in books such as such as Real World Haskell.
The critical thing, as you observed, is to separate IO from retryable STM transactions with a barrier; e.g. by receiving data, doing an idempotent computation inside dosync or swap!, and then emitting results. Haskell has an advantage in that this separation is provable by the type system, whereas in Clojure you need to remember.
A contrived example:
(let [pizza (get-from-fridge)
meal (dosync
(alter eaten-foods conj pizza)
(deref eaten-foods))]
(tell-friend "So far I ate" meal))
where get-from-fridge and tell-friend are IO operations, and our list of eaten foods is mutated by pure functions.
In practice I don't find this particularly limiting: dosync and swap! are almost always short operations for performance reasons anyway. Then you return a consistent snapshot of the updated state from dosync--where multiple pieces of state are involved, you can use vectors or maps along with destructuring bind. Sometimes it's a tad unwieldy, but typically much shorter than the equivalent mutex dance.
There are other aspects of Clojure's concurrency libraries, like agents, futures, and promises, which are useful in IO. Specifically, agents give you asynchronous serializability, futures give you asynchronous concurrency, and promises allow for synchronous handoff of delayed values. Those are quite useful when working with IO, though for heavy lifting you may be better off using explicit queues and worker pools from java.util.concurrent.
Sometimes I think of Clojure (as an imperative language) as the dual of Haskell's lazy evaluation model. Clojure uses futures, promises, and lazy sequences to provide explicit laziness, where Haskell uses the IO monad to provide explicit ordering. Both have an STM for serializable, atomic mutability between threads.
That's odd, that seems a little inaccurate to me. Clojure's agents are useful for encapsulating I/O-related resources, and are integrated with the STM system. Per http://clojure.org/agents -
Agents are integrated with the STM - any dispatches made in a transaction
are held until it commits, and are discarded if it is retried or aborted.
Refs and atoms shouldn't have side effects, but agents are great for performing side effects. From page 214 of Clojure Programming:
Unlike refs and atoms, it is perfectly safe to use agents to coordinate
I/O and perform other blocking operations. This makes them a vital
piece of any complete application that use refs and Clojure’s STM
to maintain program state over time. Further, thanks to their semantics,
agents are often an ideal construct for simplifying asynchronous
processing involving I/O even if refs are not involved at all.
Further:
Agents are integrated into Clojure’s STM implementation such that
actions dispatched using send and send-off from within the scope
of a transaction will be held in reserve until that transaction
is successfully committed. This means that, even if a transaction
retries 100 times, a sent action is only dispatched once, and that
all of the actions sent during the course of a transaction’s
runtime will be queued at once after the transaction commits.
For games, Chris Granger and his team have coded up a prototype game in Clojurescript using the Entity-Component-System architecture, which seems to be a good fit for functional languages. It might be worth checking out: http://www.chris-granger.com/2012/12/11/anatomy-of-a-knockou...
Actually, in my experience the practises in the high-end AAA game industry are now much closer to FP than the rest of the industry. The current-gen consoles have lots of really weak threads, and to get any decent performance you have to use a lot of them. So, modern games do things like double-buffer the entire world state to allow efficient multithreading.
Learning a powerful, less main stream language like Clojure can also signal to smart, innovative software companies you might be the kind of person they should hire.
For those of you in the industry, why would I want to use Clojure? What would be a good project to try it on? A Web app? Why is it worth learning it?