Fork me on GitHub

I had a great time at erubycon this year. The speakers were great and the Columbus Ruby Brigade had an awesome group of Ruby enthusiasts.

The most compelling talks were what I would consider the talking-head talks given by Stuart Halloway and Neal Ford. Both of these speakers had a way of identifying with the audience and bringing out things we probably already knew but maybe didn't know how to express. What I took away from one of Neal Ford's talks was that we already know we are doing it the right way. We know that agile with a little 'a' is the right thing to do, but we don't always practice it and worse yet we refuse to use and/or accept historical evidence proving us wrong. We all know about "the bad things" like NNPPs, adding resources, commoditization of software developers, etc., but somehow all of these persist in our (agile) enterprise organizations.

Stuart Halloway talked about blub and how Ruby isn't blub just yet but is on its way. While talking with a few people back at the hotel Randall from Engine Yard mentioned the Bathtub Curve which, when applied to developers, can give one an idea of when something has become blub. It all boils down to supply and demand, which brings the essential problem of commodization of developers back to the forefront of bad software.

Stuart's second talk on How to Fail with 100% Code Coverage was good. My reactions to his points ranged from "Yeah, duh..." to "Yikes! I know I am doing that currently." I really enjoy talks like his that don't dance around issues. He did an excellent job of encouraging the audience to be better developers through introspection and critical analysis. Though I know some who disagree, I don't want to waste time to pat myself on the back. I'd rather spend my time figuring out how to be better. Praise isn't supposed to come from oneself.

Here's the list of talks I found interesting from a mostly technical standpoint:

Jim talked about Erlang and Clojure. Erlang is definitely the Ruby developer's New Hotness of choice, and there is good reason. We are living in the beginnings of Ubiquitous Computing and we need languages that deal with that world correctly. Though my eyes disagree with Erlang's syntax I cannot disagree that highly scalable software is not only necessary in the enterprise it will be the competitive advantage for small companies. Unfortunately for enterprise developers, the necessity for a simpler scalability won't become apparent to the business until their laziness is challenged by leaner and more performant companies.

Giles always has interesting things to say. He was a good way of presenting, as well. I told him I appreciated his ranting nature, though I'm not sure if he thought that was a good thing. So it only makes sense that I horribly misunderstood his talk's message as well. What I took away from his presentation was that you can write something really leet and everyone will hate you for it.

Randall's talk was great. He has had a lot more experience with a lot more systems than I can even name. The only reason I bring up his talk is because he mentioned doing things right the first time. I think this message is a bit of a conundrum for me. I heard the same thing at the Silicon Valley Ruby Conference in San Jose back in April from Alex Le who works on Friends For Sale. The reason this is confusing is because that isn't exactly agile and DHH gave a somewhat conflicting viewpoint when he mentioned doing something now and worrying about scaling when you actually have customers. Wanstrath mentions something similar in his keynote at Ruby Hoedown. I understand both viewpoints, but I think there is more credence for doing something now because doing it right, in my experience, usually means never really doing much of anything.

The Dialogue presentation was great simply because of the format. Essentially it was a play... about programming... it was great! I would love to see more presentations done in this format. We do live-coding, but it is always just one guy making the same assumptions that he is solving the problem the right way and that everyone agrees with him. The dialogue let the audience sympathize with more than viewpoint, but still had an obvious message about the right way.

And finally, Josh Holmes's talk on Silverlight was interesting because we got to see some really interesting Silverlight apps. He is also an excellent speaker. I am not really excited about Silverlight, but I do believe HTML is waaaaay dead. We have it on life support with all the crazy things we're making Javascript do. The only real way forward is through service based websites and presentation layer clients that can truly take advantage of image and animation libraries. Two examples of what I mean are GitNub and Lighthouse Keeper.

0 comments: