Amit Rathore talks to RubyLearning's Clojure Course Participants

On the eve of the first free, online “Clojure 101” course, Michael Kohl of RubyLearning caught up with Amit Rathore, author of the forthcoming book – Clojure in Action. In this interview, Amit Rathore talks to the Clojure 101 course participants on Clojure.

Michael>> Welcome, Amit and thanks for taking out time for RubyLearning’s Clojure course participants. For their benefit, could you tell us something about yourself?

Amit>> I’ve been programming since I was 11, and been designing and developing software systems in a professional setting for about ten years now. The last few years have seen me transition from Java to Ruby, along with a smattering of other languages such as Python, Scheme, and Smalltalk. Since late 2008, I’ve been using Clojure full-time. I’ve made a few open-source contributions to the Clojure community – some examples are data-mappers for HBase and Redis, and another one called Swarmiji which allows you to write distributed programs (that span multiple CPUs, not just multiple cores) in Clojure. I’m currently the Chief Software Architect at a startup called Runa, in Mountain View, CA. We are a provider of SaaS solutions to online e-tailers to enable them to provide real-time, analytics-driven promotions to their shoppers. More than 90% of our backend is written in Clojure.

Michael>> How did you get involved with Clojure?

Amit>> As I said earlier, I’ve used Scheme on my personal projects from time to time. At Runa, we always knew we could benefit from using a Lisp for our back-end, what with all the analytics and machine-learning that the system needs to do. When all the scalability requirements were thrown in, we seriously considered using Erlang, thanks to it being functional and its concurrency support. It was just around then that the Clojure community really started taking off, and we decided to try it out. We’ve never looked back, and we’re extremely pleased with the outcome so far. Our analytics-powered, adaptive, conversion-marketing engine is miles ahead of any potential competition… and we can add features (and make sure things still work, and are as performant as needed) faster than any potential competition. If you have read Paul Graham’s essay called Beating The Averages, you know what I’m talking about. In Jan 2009, following our initial Clojure deployment, I started the Bay Area Clojure User Group. It’s going to host its 17th Meetup next month. It’s great to see the community growing, and its great being a part of it.

Michael>> You are currently writing “Clojure in Action”. What can you tell us about the book?

Amit>> When Manning Publications contacted me regarding a new Clojure book, they wanted something different from what was already available, and what they knew was in progress. When we discussed my Clojure experience and startup background, we came up with the concept of the Clojure in Action book. The idea basically resulted in a Clojure book that does a few things:

1. Teaches Clojure from first principles i.e. why are certain things the way they are, and how they’re better than what currently exists in popular languages.
2. Teaches a developer new to Clojure to get going after reading the book – it answers the “OK, what now?” question – by addressing issues like test-driven development, IDEs, dependency management, debugging/profiling, and so on.
3. Real-world usage for web-scale applications – from all the experience I’ve gleaned working with HBase, Redis, RabbitMQ, Amazon services, Hadoop/MapReduce, and so on.
4. And finally, advanced usage of macros to build DSLs (domain-specific languages). So readers of all levels will get something from the book – folks new to Clojure can get started quickly, while intermediate to advanced users can gain also.

Michael>> Many of RubyLearning’s Clojure course participants have a Java and or Ruby background. What, according to you, are the benefits to these participants after learning Clojure?

Amit>> The functional approach is just a better way to program. So that, in itself, is a huge learning opportunity. In the words of Eric Raymond, the famed hacker, “LISP is worth learning for a different reason — the profound enlightenment experience you will have when you finally get it. That experience will make you a better programmer for the rest of your days, even if you never actually use LISP itself a lot.” And the great news is that Clojure, the latest incarnation of LISP, is actually so usable that once you start, you will never want to program in any other language again. This influence includes meta-programming, which will especially benefit Ruby folks. On the other hand, it will also make developers from both the camps (especially Java) realize just how limiting their languages are. Actually, this is also a good thing.

Michael>> Why do you think that such a free, online Clojure course at RubyLearning would be beneficial to the Clojure community?

Amit>> Clojure will only succeed if more and more people adopt it. While it is a no-brainer once you understand what it offers in terms of productivity and code-quality, it has the initial problem of getting-the-word-out. Such a free, online course can do this quite effectively.

Michael>> How should they go about acquiring knowledge and skills in Clojure? What’s the best approach?

Amit>> There is one book published so far, and several coming out. Those are good resources. There are plenty of tutorials online, as well as a ton of open-source code which can be great resources. There are several very active Clojure user groups around the country, and indeed all over the world. Also, the IRC channel is fantastic.

Michael>> Which areas in Clojure should a would-be Clojure programmer concentrate on?

Amit>> Functional programming would be an important topic. People coming from an imperative background have to sort of re-wire their brains. Another topic is “the Lisp way”, which is this idea of bottom-up design (as opposed to the traditional top-down approach to breaking things down). Creating mini-languages (fashionably called Domain Specific Languages these days) that allow the developer to program at a much higher-level of abstraction is another important design philosophy. These I’d say are things that take time to really “get into”… so some deliberate attention should be paid them. Other things, such as Clojure’s concurrency mechanism are also important, and yet easy to learn and use.

Michael>> Do you have any suggestions for RubyLearning’s Clojure course participants? Anything you would like to share with them?

Amit>> I’d just like to say this: As developers gain experience, they learn certain things that become guiding principles. One such thing is the power of abstractions. As they gain expertise of the programming language they’re using, they discover the natural limit to the kind of abstractions that their language of choice can create. They realize that certain things just can’t be expressed in that language. At times this manifests in wishing that the language had this feature or that. Lisp frees you from such tyranny. The parenthesis are there for a reason, and that reason is to make the macro system possibly. And that makes wonderful things happen. And Clojure is an incredible Lisp. So, stay with it, give it a real go, the parenthesis disappear in a few days, and the real power becomes apparent. And then you can never go back

Michael>> After the course, most participants would like to contribute their time, skills and expertise to a Clojure project but invariably are unaware of where and how to do so. Could you suggest how this can be achieved?

Amit>> There are plenty of open-source projects that add something to the Clojure world. GitHub is a great place to start looking, and there are others.

Michael>> On a final note: What do you perceive as the future of Clojure?

Amit>> We are all craftsmen: programming languages and software requirements are our raw material. Given that complexity of software requirements is continuing to grow, it behooves us all to use better and better tools and languages to create the next software system. As a matter of fact, tomorrow’s software almost needs to be adaptive – it needs to learn and adjust its behavior based on things like the data that it sees, or patterns of usage, and so on. This kind of complexity needs really flexible and powerful tools on the implementation side. A Lisp is a natural fit for such a dynamic, demanding world. Further, Clojure’s functional and concurrency story is very strong – as software moves to clouds of multi-core CPUs, this becomes a great advantage. Finally, being able to leverage battle-tested Java libraries in such a seamless manner is another boon. I see more and more people being enlightened about these issues, and more and more adoption of Clojure. I’d be very surprised if it doesn’t become the discerning developer’s first choice within the next couple of years.

Thank you Amit. In case you have any queries and/or questions, kindly post your questions here (as comments to this blog post) and Amit would be glad to answer.