There’s been something in the back of my brain that’s been bothering me about talks at the big conferences lately but I just couldn’t figure out how to talk about it. Until I listed to this episode of The Hacker Mind Podcast on Self Healing Operating Systems (it’s a great podcast, like and subscribe). The episode was all about this incredibly bizarre way to store operating system state in a SQL database (yeah, you read that right). The guest made no excuses that this is a pretty wild idea and it’s not going to happen anytime soon. But we need weird research like this, it’s part of the forward march of progress.

This sort of obviously experimental research is what I’m going to call “rocket ships” in the context of this blog post. It’s work that is extremely interesting, but will take decades to turn into something that affects us directly. Almost nobody can really do it, there’s no short term money in this research, and even fewer will directly benefit for decades. It will almost certainly have benefits someday, and won’t look anything like it does today.

Now, there’s another side of this story. All that stuff we actually have to do. This is the boring every day stuff that keeps the trains moving and electricity running. Nobody is giving talks about this because it just sort of exists. I don’t think KubeCon is going to accept the talk “How I keep your family from freezing to death in the winter”. Let’s call this work “eating our vegetables” or “radishes” because I needed something that started with an R. Humans like alliteration. Rocket Ship … Radish … oh hey, Research also starts with R!

So anyway, this brings me to a lot of the big conferences. I think what I see is a lot of rocket ship research that’s dressed up as radishes. I’m not going to pick on any one company or project specifically because that’s not just bad taste, but I also can’t afford to make any more enemies. Instead, I’m going to make up an example to explain what I see. Let’s use supply chain security as the backdrop because that’s the universe I mostly live in right now. You could just as easily substitute in AI, or blockchain, or french fries.

Our hero, or maybe villain, who knows, everyone is a hero and villain to someone, they are giving a talk on how we can secure open source build systems by rewriting them all in Apple II basic. It will of course be creating reproducible binaries, and we should sign everything with a new variant of Sigstore crossed with PGP called Pigstore, the mascot is a duck. And it probably also draws pictures of clowns, or maybe cats, some sort of mammal. There’s a demo, and a GitHub website. You know it’s been donated to the OpenSSF because that’s a cool thing to do. And of course the name is an unpronounceable German word related to ship building, and the mascot is some sort of obscure animal that lives in a cave.

The room was packed, the talk went over great. The conference WiFi only let half the demos work, but that was enough to show off the internet connected Apple II building Node.js. Animated GIFs showed off the rest. On second thought none of this mattered because the demos were 75 pages of terminal text nobody could read. Whatever, everything went perfect. Oh, and did they mention we should expect all open source projects to start using this build system because if they don’t everyone will call them names! The future starts today!

In the academic days (like our operating system example from the opening), it would be well understood that this was rocket ship research. It almost certainly wouldn’t go anywhere anytime soon, but was a step as part of the larger story of progress. As the arrow of time drags us all into the future, so does the path of progress, as long as you don’t live in Florida.

But this isn’t how these talks work anymore. On Monday, there will be customers asking where on your roadmap this new build system falls so they can take advantage of all these features they didn’t even know they needed. They didn’t know they needed more clown pictures, but suddenly it’s very obvious it was the thing missing from their soul that previously a fidget spinner was filling. That senior developer who quit last week said the biggest problems were no unit tests or code reviews, but it’s probably actually this.

Your leadership will say this needs to be part of Q2 planning because all the competitors are also doing it, so they read on LinkedIn. The community meeting for the project has ballooned to 300 people, all talking about how they are busy integrating things into their production environments as we speak, just skip right over test and dev, this goes straight to prod! So you know it’s going to be a huge deal.

I want to stress, these sort of projects have immense value, but I also suspect that pretending they’re something that can be used, a radish, is actually hurting the research, or rocket ship if you’re keeping track. Now instead of researchers doing weird research things, you have community meetings filled with developer advocates trying to make sure their company gets in on the ground floor. There’s no more room for experimentation because apparently version 0.1 of the framework just got released yesterday, nobody knows who did it or how. In 3 months at the next conference something even newer will get talked about, this project will be mostly forgotten, and what could have been important research is going to fester in an abandoned GitHub repository forever.

I’m not sure how to wrap this up. Is this a problem with the companies funding this research wanting to pretend it’s something more real, or is it the large conferences looking for more “real world” focused solutions instead of research that won’t go anywhere for decades? Maybe things have always been like this and I just didn’t notice before now.