BloggEd

I originally thought to write these once per day, but there was so little time because I didn’t want to miss any sessions, interesting discussions or social opportunities whatsoever. It was awesome, really. So I’ll give a very superficial overview for each day as I saw it, including only the most interesting (to me at least) parts.

Day 0

I was registered for two Monday tutorials: a BDD tutorial and one related to Really Scaling Rails. So I left my hotel for Maritime Pro Arte so enthusiastic I was about to burst. After ending up in the wrong Maritime hotel (because the taxi driver didn’t notice me saying Pro Arte, of which there is only one), I finally got to the event area which was already full of equally excited people loitering around and registering for the conference. Happily I received my papers and the bag full of interesting stuff related to the conference.

So my first tutorial session was A Half-day of Behavior-driven Development on Rails by David Chelimsky, Aslak Hellesoy and Dan North, main developers of RSpec.

The tutorial was good; it was informative, fun and well presented except for one thing; we didn’t have time to get the software to our machines, so it practically reduced to lecture. But it was a good lecture, so I was ok with that.

Another tutorial was a lecture by Jason A. Hoffman from Joyent about Scaling a Rails application from the Bottom Up in Europe. Jason certainly seemed to know his stuff and he pretty well, but otherwise I think it was a bit off-topic – it wasn’t really spesific to Rails, but was a kind of quick overview suggesting solutions to real scaling problems (think of millions of regular users, and you get the picture; 32-processor multicore servers in clusters and other stuff along those lines). So my overall impression was that even though Rails performance isn’t that good compared to many other technologies, it really doesn’t matter when there are enough users (then again, it seems like with good caching techniques and some cheating you can run quite heavily used sites with Rails like, say, Twitter). Though a tad off-topic, it was interesting and very informative.

In the evening Dave Thomas delivered an excellent speech in his keynote about engineering and art, their interconnectedness; in software engineering, many of us agree with Dave by thinking that no, there is no art without engineering and vice versa. He also called for taking responsibility for your own work by urging us to sign our work from the very beginning.

If you have never heard Dave speak and have an opportunity to do so, I definitely encourage you to, for he’s an excellent speaker.

Day 1

Highlights of Tuesday included the keynote by DHH himself, mostly about Rails 2.0 and about throwing away the rebel jacket, meaning it is time to turn from revolution phase to mundane craftsmanship; we don’t have to fight for Rails to be acknowledged anymore. Over thirty books published and IBM, Sun and many other large companies backing up Rails development, it is time to polish Rails to even more fine product as we’ve all used to expect to be.

I just had to attend Meta Magic in Rails: Become a Master Magician by Dr Nic Williams (yes, the Dr. Nic we all love for his enthusiastic ramblings and magic stuff).

You know, I never knew he’s so fun. I don’t think I learned anything new there, but it was the funniest presentation I had at the whole conference. His clever tricks may seem useless (and to be fair, some may be so), but I think he’s one of those people from whom you’ll learn something very smart every now and then. That day wasn’t meant for me, but oh boy it was fun. Tsah dum di dum.

Another session I attended was The Rest of the REST by Dr. Roy Fielding, who wrote the original dissertation about REST and is also known for being one of the Apache HTTP project co-founders, and a long-time member in the development of HTTP itself. Not strictly Rails related, it deepened my understanding about the REST architecture itself and its usage.

In the late evening (starting from nine o’clock) we went to RejectConf, where many people gave a very short, informal presentations about…. any stuff they deemed worthwhile with a strict time limit. Australians definitely seemed to dominate the event and were cheered with enthusiastic yells, screams, whistles and even roars. Not very educational, though some of the stuff was still good to know about, but what it lacked in content, it compensated in humour.

Day 2

Our last actual conference day started with keynote Bring Ruby to Enterprise, Not the Other Way Round by Cyndi Mitchell from ThoughtWorks. It was a very good speech and informative, giving good advice on how to gain more ground for Ruby in the enterprise systems. JRuby is crucial in this respect because it allows you to use existing Java solutions from inside Ruby. JRuby itself was more closely inspected by one of its developers Ola Bini in his session JRuby on Rails, also a thoughtworker.

They were both very good, but the best presentation I had that day was Mike Koziarski’s and Marcel Molina’s keynote Best Practices, offering very concise but great ideas and tips on software development with Rails (or software development in general). Molina’s exceptional talent in identifying better ways of doing things keeps on amazing me; some of the tips are actually quite obvious and simple afterwards, yet new to me even though I’ve read dozens of books on the subject and been working professionally for say, 10 years. One of Marcelli’s afterwards obvious but oh-so excellent advice is to avoid mixing abstraction levels in your methods. This has very nice side effects; higher level code is very clear and reads almost like English and as such is easy to follow. Lower-level methods in turn tend to become highly cohesive and loosely coupled, resulting in easily testable, isolated units. The idea doesn’t seem like rocket science, but even though many professionals probably follow the idea subconsciously, Molina is the first person who managed to say put it into precise, well-formed sentence.

My last session was a very close to the best session (see above), it was given by Tom Locke, the author of Hobo, Rails’ opinionated software concept taken to the extreme and allowing users to develop web apps very quickly, somewhat depending on the nature of the application. His session was extremely well put, both technically and presentationally. He also reminded me of a very important aspect in creating successful software – you have to have a philosophy in designing the software, which you consistently follow (this was clearly present in hobo). So I guess I’ll look forward to starting my next VREAD (Very Rapid Exploratory Application Development) project with Hobo. Good work!

Leaving Berlin

Ah, the time went just too fast. O’Reilly had organized the event very well, and the average quality of sessions was at least good. I guess I’ll just have to get to RailsConf some day, maybe next year? Oh, and many thanks to my employer for sponsoring me to the conference!