The Beauty of Small Teams
I was a solo programmer for about 20 years, starting with Sinclair ZX Spectrum in 1983 and going through several jobs including the company I cocreated in 1992, which grew up to around 300 people ten years later. This solo experience has spoiled me. I was the product manager, the ux designer, the architect and the coder. And that was extremely efficient. Of course the solo approach has limits to the scalability, as there are only 2 hands and 24 hours in a day, no matter how much money you throw in.
So at some point I had to give up and cede the tasks to the team. And that was so immensely frustrating. Despite the team being many developers (I still tried to keep the architect role), things started going super slow. I was spending countless hours teaching, explaining, then explaining again and so on, and never could get away from the impression that I could do in a month what was taking them a year.
That feeling has unfortunately stayed with me forever. And has often been the reason of my frustration. First, people today do not want to be multidisciplinary. Or only a very few of them can. They assign themselves as front-end, or back-end, or embedded, or mobile, or something else and cling tightly to that categorization. I remember myself coding a very high level abstract classes and at the same time cranking some assembler routines for the ultra high performance pieces. Very few can do this today. Well of course I know some that can (and value them extremely high), but most seem have that categorization hard wired in their brains.
So DHH makes a number of very valid observations about small teams. Less meetings. Sprints versus marathons (that resonated with me as I definitely like long runs). No retirement: dealing with problems is the way to happiness (I love that!). And finally the network-cost effect of communication between humans that simply does not scale. Which comes down to the concept of "one programmer, one designer, one feature". The magic of small teams: no advanced methodologies, no multiple layers of management, just sit down, make something and see if it's good. So close to my heart.
Comments
Post a Comment