We have some major communication and usability issues with software. I’m not referring to the lack of documentation (although quality documentation is needed), nor am I talking about UX. It appears that there are way too many choices in what software product to use, but there is very little structure into how to implement it to solve your problem. For example, there has been a bit of a fight between Swarm, Kubernetes, and Mesos. Even though they have similar usages, each of them have vague descriptions of what they actually do. When I’ve seen comparisons of the products, I tend to see complex routing strategies being discussed rather than what they have in common – or where one product succeeds over others (and fails in other ways).
What am I ranting about? Well, a lot of products describe what they do, but the descriptions rarely ever make it easy to understand to the reader. All of the projects that I talked about earlier describe what they do [they tend to make sharing physical resources transparent for containers, each have their strengths and weaknesses]. All of the projects claim that they help to build a cluster, but each tend to fail on describing how they do so on an infrastructure level which can actually give implementers a shot at making an intelligent decision for their environment/purposes. They tend to build up a picture in the user’s head, only to be let down, that it’ll make the user’s scaling problems go away. Unless the user of the project has made their application distributed (in the way they expect), these technology choices won’t help them.
Silicon Valley has latched onto a harmful mindset. This has been encouraged via the Lean Startup/MVP mentality where someone decides what they believe that you need and only that is what is built. Your actual needs from the product are dictated by someone else, and are only implemented when they decide it’s going in the product. Complex products tend to be sidestepped due to insecurity about the people using it. It makes me think that technology being produced by startups in the Valley are akin to promoting the education of math to its users, but to be in denial of operations beyond addition and subtraction. Complex products can work, and can be used to their fullest extent. The problem is due to the communication about the product. Rarely have I ever seen a complex product described in a simple-enough manner up front to make it easier for the user to explore different avenues after he/she has achieved proficiency in the basics.
One thought on “In Support of Complex Software”
Comments are closed.