Developeronomics: a very necessary debate!

Last month, I read the Venkatesh Rao’s article about Developeronomics, a new word invented to underline the huge value that can be embedded into a single valuable software developer. Leveraging the idea that made Freakonomics popular, the author emphasizes the endless need for more and more sophisticated software and argues on the very active talent war on the market, both for large technical companies and for active startups. Then he advises anyone with some surplus capital to invest it in a brilliant developer as you would invest in stocks or in good wine. By the way, wine certainly was a good financial investment to make before the 2008 crisis, but this is yet another story.

Basically, I tend to agree with the idea developed in the article; especially regarding top developers that in my experience can produce more than a typical team of 10 average developers… provided the team has the skills to overcome all the technical challenges. So it is always a better choice to have an over-skilled developer rather than an under-skilled one. It’s much more inexpensive in the end.

When I was working at Microsoft, I had the chance to hear Bill Gates tell the Windows NT story and how he hired Dave Cutler from Digital to do the job. When Bill asked Dave what the optimal team size would be, Dave’s first answer was "1". Then he added, "It is the optimal for efficiency but if we need to deliver within a reasonable timeframe, let us build a team of about 10 very good developers".

This is totally consistent with our own observations on the field. Still, people who have never written software themselves keep reasoning in man-month logic as if the man variable in this expression were commutable. And unfortunately, many people making strategic decisions about software development have never developed themselves.

This led to the huge "off-shore" trend that no one is really able to confirm as successful. I deliberately do not qualify this as a global failure, as I am lacking reliable data to validate that off-shore as such is a failure, but I am also lacking reliable data that demonstrates that it really works. The point here is that we are lacking a key metric about what has been produced (in value, not in cost) so we could really compare results, not expenses. I will post more in the future, on both software development measurement and off-shore, because these are also key topics.

Interestingly, Jesse Jiryu Davis posted a valuable reaction to the article that extends the discussion in my view. In the process, his "It almost never works" demonstrates at least that you cannot succeed with mediocre programmers (them being off-shore or not).

His other interesting major comment is his "Each year we write software that prevents us from having to write more software". That clearly strikes a chord to me. This is the history trend that everyone has to consider, but my personal opinion is that the opportunity to get more "out of the box" solutions also creates the need for more software. This is as infinite as the knowledge of the world and I cannot really imagine the day when we say "Ok, we are done with software development, now let us just maintain the worldwide code base!".

This is why our CodeFluent Entities approach is not to replace developers, but instead to allow them to go a mile further by automating the repetitive tasks for them. The idea is to let the tool do the tedious work so we, developers, don’t have to do it.

To conclude, software developers always have to adjust their skills to ever-improving tools and technologies. This is part of the fun and the value of this job, and these are the guys that I typically want to invest in.

Daniel COHEN-ZARDI

CEO, SoftFluent