I learned HTML some twenty years ago and had a good few years of fun with the web, but recoiled from Javascript in horror and CSS in frustration. I eventually gave up on the server side as well, for political reasons: the strength of the Internet was in its gift of decentralized communication, but the web is all about big central servers controlled by singular institutions. I came to feel that investing time in such projects was actually counterproductive, in terms of helping to create the kind of world I want to live in.

That was a long time ago now, over a decade at least, and I am periodically shocked by glimpses into a world that has continued developing broadly and quickly, and which no longer much resembles any of the stuff I used to work with. I suppose the old mainframe hackers must have felt like this, as they watched the microcomputers take over.

The first of today’s jolts was a thread on Hacker News about a new standard for virtualization containers. I understand what virtual machines are and some of the reasons why people use them, and I know a fair bit about the low-level mechanics that make them work, but it’s clear that web people have taken the whole thing far beyond all that because I just can’t wrap my head around containers. I am ignorant of the problem they are designed to solve, and so I can’t really grasp – from the descriptions – what it is they are intended to do, or why that would be useful.

The second was a presentation about a piece of security analysis software, which started with a series of extremely startling claims about the product’s capabilities. I was running ahead with what I know about debuggers and low-level machine operations trying to figure out how they had accomplished these things… but of course the reason they can detect these things is that they’re not analyzing what I would call “applications” at all, but rather web services, and web services written in Java or .NET at that. And suddenly the whole thing seemed trivial, because of course you can analyze anything you want when you can play god with the virtual machine! Which is not to diminish the engineering work they did to make it happen, just to reduce it from the domain of magic. It seemed clear, at that moment, that I must be thinking about software from a sufficiently different perspective to their intended audience that they could reasonably expect people to understand the implied limits on their description as they apply to web programming.

I’m not really unhappy about this state of affairs, since I’m still not interested in working on web software, and I’m still not having trouble finding work in the field of what I still, with increasing quaintness, think of as “normal software development”. But it is clear that the world around me is changing, and I’m not seeing anything like a return to the kind of robust, resilient, democratic distributed architectures I want for the future of the Internet. It makes me wonder how long I can keep on holding out, and how long it will take me to catch up if the day comes that I have to hold my nose and jump in.