The neverending waveform of the full-stack developer
There was an article on Techcrunch a couple days ago which was linked in our internal mailing list the other day, titled The Rise And Fall Of The Full Stack Developer. I read it, and I just couldn't figure out why the title is about "the fall" of the full-stack developer (and I said as much on the mailing list, after which I was encouraged to write this blog post). In this post I'll try and explain why it's not the end, but just a stage in a recurring cycle
What I read was a waveform - the article describes that first you had low-level assembly, which is specialised but still pretty straightforward since it's just one platform and language. Then things started to get more complicated, with the larger web applications involving lots of experts in various fields (Java, database management, server and JVM management, to name a few from the article).
But, as the article states, that isn't the full stack of today - added to that are things like machine learning, mobile development, more advanced and less traditional programming languages, frameworks and persistence solutions, etcetera. The conclusion of the author is that it's too much for one full-stack developer to take, that there's no way a full-stack developer can be an expert in all of those fields - and he basically renames the person that knows a bit of all of those technologies a full-stack integrator, and declares the full-stack developer dead.
But I didn't see that. What I saw was just history repeating - from simple and manageable, to complex and too much for one person to handle and know all about. From assembly and PHP, to Java enterprise software and Docker-contained AWS instances running a MongoDB and Scala REST interface to power your Android, iOS and single-page AngularJS-powered webapps (to name but a few buzzwords).
I'm sure the next 'simple' wave is already around the corner - maybe it's already here, lurking somewhere until some more influential guy goes "Hey guys, let's go back to the simpler, good old times where you wrote code and it Just Worked™!", where a wave of developer will follow - most likely a combination of younger people, new to the software development world, and older people that have been stuck in an overly complicated software development rut for far too long.
As for not being able to keep up with it all, this is why it's probably better to assemble teams out of T-shaped developers - full-stack developers with a broad knowledge set (and more importantly, broad interests and curiosity), but with at least one specialisation. This was extended within Xebia to pi-shaped developers, then made extreme to comb-shaped developers, but the latter is just a generalist like mentioned in the article - a jack of all trades, master of none. And it's important to realise, as a developer, that it's OK to not know everything, to miss some update, to not learn that fancy new programming language by people disgruntled by Java's slow development - there's simply too much information updates today, and trying to manage it all in the extremest forms can lead to serious problems. But the rapid development is a good thing, I might add, I don't think the software development world has ever moved as fast as it does today, and it's not nearly the end yet as long as more than half of the world's population doesn't have access to the internet and its boundless resources.
I think increasing complexity is just part of software development. I mean look at Facebook and Twitter - they were started using those simple, accessible tools like PHP and Ruby on Rails, which allowed them to get a flying start and adapt quickly to their huge growth (and with the former desperately clinging to their decision, despite lots of people telling them they should use a different technology), but despite that they still turned into some of the most complicated pieces of software ever. The important bit is to be able to manage all that, not so much a decision between a simple or complicated toolset - or having all of your developers know everything. Similarly, the current trend is microservices - again back to simple, one-purpose miniature applications that do one thing and do them right. But those will just end up being the next generation of huge, complicated software systems, given enough time. As a colleague stated, "Microservices are just hipster SOA".
As the saying goes, the more things change, the more they stay the same. The full-stack developer isn't dead and won't be going away any time soon, he'll just go by a different name depending on the times - J2EE Certified Software Engineer, full-stack developer, multidisciplinary expert, T-shaped developer, Xebian, chief of technology evangelisation, or whatever else the HR or marketing departments of the near future will come up with.