10Gen CEO Dwight Merriman Still Writes His Own Code

#cool. very impressive!

Why wouldn't someone else fix the bug?

Why not? Of course you can't have the CTO drop everything when there's an unexpected bug, but I do think there is value in coding, even for "upper" management. Even looking in on the code from time to time can give you a snapshot of where the product is at and in which direction it is evolving. If that direction is different from what you expected, you can ask the programmers why they chose to use the approach they did.

I think the most important thing is that the CTO does not place his or her code above the code of the individual maintenance programmer. The worst thing he or she can do is somehow mark his or her code as "untouchable". Untouchable code is bad code, regardless of how well it works or how well it's written. It doesn't adapt to changing requirements like the rest of the application, and the rest of the application has to go out of its way to accommodate it.

However, as long as he or she makes it perfectly clear that his or her code is just as subject to change and modification as everyone else's, I don't see why the CEO/CTO can't get their hands dirty now and again.

Agreed, if you are able to make it to any of the Mongo events, don't miss seeing Dwight talk (particularly about MongoDB internals). He's an even better lecturer than my favorite college professor.

If 10gen make it big: perhaps. And then again, perhaps not.

The fact that this guy spent 10 years as the CTO of a company and was still capable of producing a piece of quality software like MongoDB says that this guy is isn't quite like other people. Very few people who make it into the executive ranks (and stay there for a decade or more) manage to stay sharp enough to actually write a coherent piece of code. Much less a piece of code worth building a company on.

However such people do exist. Years ago I had the pleasure of working with a fellow who was nearing retirement age. We worked on a project together and he was learning a new programming language. As the weeks passed I slowly learned that he used to be an executive, but found this work boring after some years and wanted to return to the floor as an engineer. I also learned that he worked on software for the Apollo program.

I have no doubt that if 10gen become very successful he might have to dial back his level of direct involvement. But if he is the brains behind MongoDB, it would only be natural that he stayed involved in the project.

As an investor I'd see this as a plus. If nothing else then because it proves the CEO does have some brains and is less likely to be a business dullard.

I'm not sure what that anecdote was trying to say. I think the majority of the 2 hours would be getting to grips with the codebase, given that he probably hasn't had time to look at it in years. I've never seen anything to suggest that Zuckerberg is an incompetent programmer.

Similarly, if you took one of the engineers and showed them the commercial agreements and reporting that Zuck would be instantly familiar with I'm sure they'd take a long time to get grips with it.

It usually means that you're spending most of your 9-5 sending/receiving email, on the phone and in meetings. Then you go home and look at the commit log to see what's changed, check the issue tracker and push some commits yourself.

Well, 10Gen is a little smaller in scale that that Facebook company.

Also, doesn't DHH code too?

Would it bother you if he had shifted into a management role? Or just less inspiring.

I was an AVP at a large company...it had 6+ CEOs....hundreds of directors + managing directors...thousands of VPs...thousands of AVPs.

Point is, large companies throw titles around like candy. Also, I hate to burst your bubble, but programming at a large company is very different than programming at a start up (especially one on of the rarer technology-focused startups like 10gen).