3 ms·
I have a weird hypothesis about the “worse is better” phenomenon which is what I think he’s getting at. In back office / cloud / IT type stuff I wonder if comp
by api 2y ago
I have a weird hypothesis about the “worse is better” phenomenon which is what I think he’s getting at.
In back office / cloud / IT type stuff I wonder if complex things like Kubernetes win over simpler approaches precisely because they are more expensive and labor intensive. As a result of being labor intensive they pick up more users who after investing in climbing their learning curve become champions. Simpler or more “fire and forget” systems require less labor and so win fewer converts.
- mrkeen 2y agoI've been doing mostly backend dev, and watching from the sidelines as buzzwords come and go: vmware, virtualbox, puppet, vagrant, ansible, zookeeper, mesos, docker-compose, chef, etcd, docker-swarm, terraform, helm. I don't know what half of them do. But honestly two of them stand above the others: docker and kubernetes. Docker is what your program is, and kubernetes is somewhere for it to live while it's running.
- api 2y agoMy beef with Kubernetes is how hard it is to run and how complex its configuration is. The thing it does is good but something much cleaner seems possible.
- mwcampbell 2y agoWhat are the simpler, more fire-and-forget, approaches that you have in mind as alternatives to Kubernetes?
- slow_typist 2y agoAlso, once a certain technology X is established within an industry, sticking to X can be understood as risk management . Since everyone does it, X can’t be all that wrong. And if X fails, all competing business are relying on it, too. Then, universities include X in their curricula… The result is a market where more grounded approaches to development are outcompeted by stacking abstractions on top of each other, leading to systems that are mediocre in the best case, and killing people in the worst.