Martin and Me and DSLs

Back in June of this year, Martin Fowler and I did a keynote at The ServerSide Symposium in Barcelona on a subject for which we both have some interest (and dare I say passion): Building Domain Specific Languages. I saw cameras at the back of the room, but it didn't actually dawn on me that they were filming the whole thing...and that they would put in on the web. But, sure enough, here it is. It took them two months to post it and I think I know why: some poor person had to go and transcribe Martin and me yammering away for an hour. I hope they got paid well!

This keynote pretty much summarizes much of the thinking that I've been doing about DSLs over the past couple of years, and it is nice to see the agreements and disagreements that Martin and I have on some nuances of this topic. Clearly, though, we both think this remains a Big Deal.

Lights, Cameras, Geeks!

At the Agile 2007 conference this year, they asked for submissions for videos featuring agile development themes call the Agile Advert contest. ThoughtWorks submitted a couple (created in ThoughtWorks UK) and won the top prizes. You can see them on YouTube under Developer Abuse and Being Agile is our favourite thing. Hilarious!

Pair Programming Spellchecking

My colleague Jay Fields has a nice post about pair programming, when it's good, and when it's not so good. I wanted to add to and amplify his remarks a bit. Basically, anytime you pair two developers of vastly different skill sets (what Jay calls a "spell checking pair"), that's not pair programming: it is mentoring. That doesn't mean that it is wasted time, but it is a different activity.

I find that mentoring is tiring for both the advanced and novice developers: tiring for the advanced developer because they have to go slower than their natural pace and tiring for the novice because they are having to peddle faster to keep up. However, mentoring is a great way to greatly increase the skill level of the novice. If you are going to mentor, it should be for just part of a day. After the mentoring is done, switch pairs with more matched levels.

A lot of companies seem to think that's what pair programming is: take a intern from college and pair them with an ace to build up their skill level. All that does is stress both.

We are generally careful when engaging with clients to pair program with their developers to determine if the project is enablement or delivery. If it is delivery focused, we tell them that the more senior developer may well leave their pair in the dust, and should be reorganized. You can't usually do both enablement and delivery at the same time except in rare circumstances.

Dependency Injection in One Sentence

Jim Weirich issued a challenge to describe dependency injection in once sentence. My friends at Relevance had a good version. Here's mine:


Dependency Injection enables a vitally important but nevertheless weak, limited, syntactically confounding, and dauntingly complex form of one of the kinds of meta-programming that should exist in the language.


This is fun -- can we do annotations(C#)/attributes(Java) next?

No Fluff, Just Stuff eXchange: London 2007

Way back in the late nineties, I was working for The DSW Group, which was a Borland partner. We did Borland training classes worldwide, and that's ultimately how I got into the business of speaking at conferences. At the time, I was all geeked up about Delphi and at the first few BorCons I spoke exclusively about it and C++ (in the form of C++Builder). However, in the late 90s, I stopped having that lovin' feeling for Delphi and started really dedicating my efforts in the rising tide of Java. But what was interesting to me was the adoption level in different countries. I was still traveling to BorCons around the world, talking about JBuilder. And some countries I would have huge audiences. And in others, you could see tumbleweeds blowing through the aisles. I remember the first few Java talks I did at the Entwickler Conference in Frankfurt, no one even knew what "Java" was.

It was interesting because the UK seemed to be one of the last holdouts for Java. Even into the early 2000s, I would go to Borland conferences in London and be in the small room, while the Delphi groups were packed. Of course, these were Borland conferences, so the crowds were self selecting, but it seemed to me like Java was a slow burn in the UK. Then, of course, overnight, it was the cool thing to do.

London is now a first-class Java city, with lots of Java work going on (Java is very popular in the financial sectors, and London sure has lots of that). The only thing missing from the London Java scene finally appears this August: No Fluff, Just Stuff. What better place to make the cross-ocean debut of the premiere Java conference. In the past, people have literally flown from the UK and mainland Europe to see No Fluff, Just Stuff shows in the US, in random cities like Columbus, OH. Now, you can ride the tube to see a No Fluff, Just Stuff show!

The traveling circus that is No Fluff, Just Stuff invades London on August 29th, for 3 days. It's the normal cast of speakers from the US (including Venkat Subramaniam, Ted Neward, Brian Sletten, David Geary, and I've managed to stow away as well) mixed with some great guys from the UK (my ThoughtWorks colleague Erik Doernenburg, Graeme Rocher of Groovy/Grails fame, and some wild cards).

To help pump up the enthusiasm for this event, No Fluff, Just Stuff and Skills Matter have a cool promotion. Every attendee who registers before August 17th with the special code of NFJS-NEF666 will receive a Nintendo Wii. Not "entered into a drawing for a Wii" -- "will receive a Wii". Wow.

Come see us in London -- it should be a blast.