How do you create a 100th Monkey software development culture?
Wednesday, July 17, 2013 at 8:49AM
Todd Hoff in software

Someone reading my now ancient C++ coding standard recommendation for using doxygen to automatically generate documentation from source code, asked a great question:
I've often considered using doxygen, I always ask myself - is this really useful? Would I use it if I were new to a project? Would programmers working on the project use it?

I'll rephrase their question to more conveniently express a point I've thought a lot about: Why do companies put so little effort into automating their own development process to make development easier?

It's like the hair stylist whose own hair looks like someone cut it using a late night infomercial vacuum cleaner attachment. Or it's like the interior decorator whose own house looks like a monk's cell.

Software organizations rarely build software to make developing software easier. Why is that?

Because there are three ways changes are made in an organization:

And of course you can't get time to work on anything extra because where is the ROI for your customers? If bathroom breaks weren't physically required they would outlawed because there's no ROI in it.

Let's say you do get time to work on that extra problem. You probably won't get enough time to do a good job so you end up with one hack built on top of another hack because people can't assign value to all the infrastructure that often makes the critical difference between project success or failure. 

So someone makes a quick little hack. But the next person or group that comes along won't like it because it doesn't do X and the person who did the original hack doesn't have time to make any changes. There's no ROI in it for them. 

On learning the tool they may want to use won't be improved the new group has a choice: improve the existing tool or make something new. 

Nine times out of ten the best strategy is to make something new because that is what will solve their problem faster. The existing tool isn't supported,it's not well documented, it's a long ways from being general enough to solve new problems, so why take the hit of trying to fix it? There's no ROI in it.

An organization can end up with dozens of similar systems and none of them is ever quite good enough to win over enough hearts and minds needed to dominate. 

Even worse and more likely is that people in an organization don't even try anymore because they know any efforts to make the development process better won't work out in the end. Why bother? There's no ROI in it.

I've seen this same cycle over and over and over again. 

They say when bears are thriving in an area it means the ecosystem in that area is pretty healthy because bears are at the top of the food chain. If bears are doing well it means the food chain below them must doing pretty well too. We can say the same about many infrastructure components supporting development.

Automatically generated documentation is one of those things that usually doesn't get done because it takes too long to do right and there's no ROI in it. 

But automated documentation generation is not so much about the technology needed to make it happen, it's more about the culture needed to maintain it and make it useful. It's your culture more than anything that will make a project successful. 

Automated documentation assumes people wrote the documentation to begin with. It assumes a tool was selected and integrated into the build system and more over that there even is a build system. It assumes people maintain it. It indicates that someone thought about what they were doing enough that they could explain it coherently to someone else. It means that when someone wants to know if a capability is available then they can do a search of the documentation and hopefully find what they are looking for.

And just maybe they will feel the code is good enough to extend and use rather than rewrite. It means you can have a culture where people are building on each others work rather than replacing it. That's the secret to hyperproductivity.

You can make a similar case about many other development practices: unit tests, source code control, bug tracking, training, having a methodology, automated builds, automated regression tests, continuous deployment, monitoring, auto-provisioning, etc etc. These practices are often ignored because the ROI isn't direct enough. But if these practices are in place it indicates a pretty healthy development ecosystem in which people are probably pretty productive.

Another tool that doesn't get enough use is a company or development wide wiki (or equivalent). If they are in place the productivity gains are huge. And again the active use of a wiki indicates a healthy development ecosystem.

No matter how useful wikis are they are hard to get adopted. A while ago I created a number of wiki pages describing how I have I worked to get wikis adopted at several companies. I now think the approach described in Getting Your Wiki Adopted doesn't just apply to wikis, it applies to most any tool or process you want to add without 100% corporate buy-in. It's a sort of 100th Monkey manifesto. Give it a read and consider how it might be transmuted to apply to your project and corporate culture.

Let's take the automated documentation system as an example of how to apply the ideas. I'm not fully fleshing the idea out, I am just giving a few thoughts on each item.

So don't give up. Next time you want to make a change and nobody else seems to care then go 100th Monkey and see what happens. You may be able do the impossible: change an organization for the better.

Related Articles

Article originally appeared on (
See website for complete article licensing information.