Since my commute now involves going from Seattle to the east side everyday, I've desperately been trying to find something to do while stuck on SR-520 during rush hour. I started looking through podcasts and hit gold with Stack Overflow (SO). For the uninitiated, this is a 60-minute podcast with Joel Spolsky and Jeff Atwood where they talk about their site, StackOverflow.com and general content interesting to programmers.
Now, I typically won't post about a podcast but I just came across a post where someone spoke about how much they hated the SO podcast. "That isn't right!", I told myself and decided to write something up. As a tribute to Jeff, I decided to steal images for this post from his previous posts and write it in the 'coding horror' style.
Here, in one sentence, is why I love the SO podcast so much.
It is like watching a microcosm of several of the discussions I see inside Microsoft at many levels.
And after many episodes, I feel like I can predict what they're going to say before they say it.
First, you have Jeff Atwood. I'm not a huge fan of Jeff's blog but I have a lot of respect fo
r him after listening to his podcast and seeing his work on SO. He has a nuanced understanding of how a lot of stuff works, especially when it comes to the web and developers. I also like Jeff because he seems to be (at least on the show) the stereotypical developer. He argues for everything to be open. He generally avoids talking about and seems disinterested the business side of things. He gets worked up when someone suggests something about making money which doesn't seem the 'open' and 'right' thing to do (ref: the episode with Jason Calacanis). In general, I get the feeling that he'd be happier coding away and making changes to SO.
On the other hand, you have Joel Spolsky (the image on the left is stolen from an old post Jeff made about Joel). Joel is going to hate me for this but Joel could walk into Microsoft today and fit right in (and I mean that in a good way). The arguments he makes and the language he uses is straight out of hundreds of internal Microsoft discussions I've been on (Microsofties reading this - think Litebulb). He is the classic Microsoft PM - a solid technical background but has spent enough time on the dark side to understand things like commoditization, pricing and various other nuances from the 'business' side of things.
Listening to these two fight it out is awesome fun. Very predictable at times but very fun. Its almost like watching a mini version of the never ending battle between the hackers and the suits. Except that Joel is actually technical, competent, smart and knows what he's talking about and thus making it hard for anyone to shoot him down as a bean counter.
If you want to get a sample of such an exchange, I suggest the last 15 minutes of their latest podcast, #60. The transcript is incomplete when I write this so you need to grab the audio. It starts off innocuously enough with Jeff taking about open-sourcing SO.
Atwood: So, the topic I wanted to talk about was, what came up was, open-sourcing Stack Overflow code. Now, to be clear, this is not something that's happening tomorrow or even next week, or even this year or even maybe next year. But eventually, I am very much for it, because I feel like that is how code... if you want code to survive in the larger world, eventually I believe in the current climate you have to open source it. The only path is open.
Spolsky: So Windows ... is dead.
Atwood: Essentially, yes, I would agree with that.
Spolsky: The iPhone operating system ... dead!
Atwood: Mm, well, the iPhone is a little bit different because it's such a closed eco-system. But in an eco-system where anybody can build it and where anybody can jump in and participate, there's essentially no cost to entry — a cell phone market is [not even?] a no-cost-to-entry market, it's pretty much a close eco-system, but a PC is definitely a anybody-can-play, there's-no-charge eco-system. I think in that eco-system, open source is kind of winning.
At that point, I was pedaling away furiously away on my exercise bike thinking "Gee, Joel got Jeff there. That was a bit of a weak response". Of course, there's a ton of open source baked into the iPhone so Joel's argument isn't exactly valid either. But the topic soon moves to a much more fun topic.
Some context first. Joel's company is working on StackExchange, which is a hosted version of StackOverflow. The idea is that you pay Fog Creek software anything between $129 to $2500 a month and you can get your own customized version of StackOverflow to run anything, even a StackOverflow for cooking.
The conversation heats up when Jeff asserts (around 49:40 in) that there are two audiences for SO - one that would pay only $0 as opposed to people who would pay something. This segways into a little discussion where Joel and Jeff start talking about whether hosting companies could take a hypothetical open source SO and do it a very low cost. At this point, I'm thinking to myself "Gee, Joel is (or should be) nervous about StackExchange. Open sourcing SO can't be good for his hosted version". I was waiting for him to say it and tt takes him a few minutes to come and say it openly.
Joel's argument (which I agree with) is that an open source SO would commoditize the product, drive down the price anyone (including Joel) could charge to the costs of running the site. A commodities market is one where there's almost no differentiation and it all comes down to price (think soap at the local supermarket). 
And that war would be won by the big hosters rather than Fog Creek. Jeff mentions MovableType/SixApart and Wordpress as examples of companies which make revenue of a popular open source project. Though Joel doesn't say it openly, my distinct impression was that he didn't think that they generated significant revenue. I would also imagine that he would want the exclusivity of the SO 'brand' rather than be yet another StackOverflow provider.
Now, it is very probable that Joel is wrong and that open sourcing SO is the right thing to do. One of Jeff arguments that resonated with me was that for code to survive for long period of times in this day and age, it needs to be open source. But I doubt that Fog Creek could build a credible business on top of an open source SO.
This was a fascinating little snippet since this is similar to several of the discussions I see internally at Microsoft on business and technical strategy (both on Microsoft and non-Microsoft products). There are tons of variables to take into account for things like pricing, whether to make something open source, how 'open' access should be and so on.
So what should they do? Here's my chance to act as armchair business strategist. Jeff and Joel need to walk this fine line where they need to stay as open as possible since their community and target audience values 'openness' as a core value. The right way to go about this might be to have features that are not open sourced and show up only in the 'enterprise' version. A good example of such a feature might be integration with Active Directory - I don't see a hosted version inside the intranet work using the current OpenID model.
In short, a fascinating little discussion. A highly recommended podcast.
