CodeMash Day 1 Mega-Summary

First day of CodeMash is over and it’s been pretty busy. Here’s some summarization of my day and random thoughts on it. It’s long. TL;DR version at the bottom.

First up, we had Neal Ford’s keynote, Software “Engineering” and Polyglot Programming. My initial impression based on his slides was that he really likes pictures of bridges, because he shows a lot of them. ;) That said, the main thrust of his argument is that languages suffer entropy (rot) just like software itself does, and that’s currently happening to Java and C# because of the feature bloat they’re acquiring.

As a result, he’s a big advocate of dynamic languages and he doesn’t agree that you can’t build enterprise software in a dynamic language — he maintains that you can, and the key is that you need as near to 100% test coverage as you can get. He also really dislikes the argument that you need static typing to make sure developers can’t get in trouble, and his response to that is that removing ineffectual warning signs will make everyone more careful.

I was discussing that with a friend on IM after the keynote, and his response was that that’s great if you’re Google and have infinite money and the ability to cherry-pick the AAA people — but for the rest of the world with crappy developers, that’s not realistic. I actually had the opportunity to speak to Neal later in the day and I put that question to him.

What he basically said (with my hearing problems, I may not have heard him correctly since it was a bit noisy in the room) was that ThoughtWorks deals with that all the time — they come in to a client company and the client’s developers don’t have that skillset, and what they do is work with the client to build that skillset in their own developer team as well. He also agreed with me when I said that argument is basically a cop-out that says “We don’t want to invest in making our people better, we just want code monkeys who can implement whatever spec they’re given.” I think it’s a BS argument, and it’s not like those skills in test-driven development sprang fully-formed out of nowhere. They had to learn them somehow.

After that keynote, I went to an introductory-level discussion of Silverlight. It was somewhat interesting, but disappointing… all he demonstrated was how to create a button and make it spin in a sickening fashion. I’m not really sold on Silverlight in general… it lets you do flashy stuff, but then again, I don’t like that crap in Flash, either. I really prefer sticking to HTML + CSS + Javascript; I don’t feel like those are lacking in capability and I don’t really want my web applications to try and pass themselves off as desktop apps anyway. I’m a curmudgeon about that.

Following the Silverlight discussion, I went to Jay Wren’s Introduction to Castle talk. I enjoyed it for the most part, although I had less than optimal seating owing to my need to get some power into my MacBook; the battery was getting pretty low, and I didn’t bother bringing a notepad to this conference. All my notetaking has been in TextMate on my Mac. :) My only complaint about the Castle talk was that it dwelled a bit more on the history and background of the project, and I would’ve preferred to see more time devoted to the actual code.

Lunch came and went and then it was time for Scott Hanselman’s keynote. He started with a brief introductory presentation talking about his life up to the present, and I must say… it was hilarious. Everything from reasons why you should sell out to Microsoft to a brief demonstration of LOLCODE. He had the crowd pretty much in stitches through that entire segment.

He then got into a discussion on changes in IIS7 and a demonstration of how it’s so modular, you can easily run PHP on it — while using .NET code to override various aspects of server behavior. To be honest, it was a bit too deep for me, but that’s not a knock against Scott’s presentation skills. My job just doesn’t have me diving that deep into IIS arcana.

I’d like to note, by the way, that Scott is a really cool, approachable guy, and a class act. I happened to bump into him later in the evening and told him how much I’d enjoyed his presentation. When he realized that I was having difficulty understanding him owing to being deaf, he moved effortlessly into sign language — it turns out that he’s a big advocate of baby sign language so he knows a lot of the basic signs, and it made our conversation a lot easier. We had a nice conversation about captioning video on the web. Scott, if you happen to read this, do me a favor and push Microsoft to add some kind of captioning or transcripts to their screencasts, would you? This trend toward screencasts in general on the web as a replacement for introductory tutorials has been a bit frustrating for me as they aren’t as accessible, and I would love to see some leadership from Microsoft on this issue. :)

Following Scott’s keynote, we had a break (I skipped the vendor sessions) and then I hit Keith Elder’s talk on Windows Workflow Foundation; we’re thinking of using it at my company and I wanted to get an overview of it. It was a pretty good talk overall; a bit hard for me to follow what he was saying, but the slides were all fantastically detailed, so major props for that. I have a much better handle on what it is now and it’s something that I definitely think could be useful.

None of the remaining sessions really grabbed my attention and I was dead tired (I’m actually worried that I might be coming down with something here — feeling a bit crappy this evening), so I skipped out on the last session block for the day. It was a very full, informative day and I’m looking forward to another drink from the firehose tomorrow.

I plan to post a summary of day 2 tomorrow or Saturday as well as some general thoughts and comments on what I’ve learned.

Too Long; Didn’t Read version: Neal Ford thinks Java and C# are bloated, likes dynamic languages. Silverlight presentation slightly underwhelmed. Castle is cool. Scott Hanselman is awesome. Windows Workflow Foundation is pretty powerful. I’m tired. More learning tomorrow.