Hiring developers: any idiot can ship product PDF Print E-mail
Written by Ed Willis   
Wednesday, 14 March 2012 00:00

Thank you!When I was a little baby developer, colleagues used to worry about getting a ton of product releases on their resumes and, similarly, looked for people with lots of product releases under their belts when hiring.  People spoke in respectful tones of developers with large collections of "Ship-It" awards.  You still hear this kind of talk - for example, people will say things like "at the end of the day, delivering product is all that matters".

But none of it means much, if anything in hiring.

Let's try a thought experiment.

  • Think of all the projects you've been involved with.
  • From those, pull out all the ones that released with acceptable quality in the eyes of your customers and stakeholders.
  • From that reduced set, pull out the ones that shipped reasonably close to the intended timelines and budgets.

If you're like me, what remains is not close to the majority of projects you've been involved with.   But that set is comprised of what might be termed "minimally successful projects":  acceptable quality, and more or less on time and on budget.  The set of projects you discarded in the above analysis are those that weren't even minimally successful.  Sobering.

Capers Jones, in Estimating Software Costs, presents data showing that nearly half of all 1000 KLOC projects end up getting cancelled (rising to nearly two thirds of 10000 KLOC projects).  Steve McConnell uses that kind of data as a basis in offering cautionary advice to the effect that when you're worried about your project being late you should probably be worried about it shipping at all.

I've struggled with those views - they just don't describe my experiences at all.  To be sure, I've seen a fair number of projects get canned - but in the majority of these cases, cancellation was due to erosion of the business case that motivated the project to begin with.  In particular, I have never seen a project get cancelled due to operational incompetence.  I've seen all kinds of scrutiny and oversight imposed on such projects but, as long as there is a business case, projects like these will limp along indefinitely - no matter how poorly they are planned or executed.  And they will inevitably ship.  So any idiot can ship software as long as there is a market patient enough to wait for them.

That a developer has shipped products in the past or the extremes of effort they went to to do so has little real meaning in making the call on whether or not to hire them.  How they go about planning and executing their work and what they do when things don't go according to plan is infinitely more important.  So it's not so much what a developer has done as it is how they've done it that really matters.

Last Updated on Thursday, 15 March 2012 14:45