Standards

I estimate that 90-95% of the work in Plan 9 was directly or indirectly to honor externally imposed standards. — Rob Pike

Like many things standards are basically a good idea. If everyone agrees on a single protocol, much work can be saved on every side. However a bad standard can massively increase work. A complex standard is one of the worst things which could ever happen to something in computing.

  1. Everyone will have to implement that standard. Fun.
  2. Everyone will understand the standard differently. It is complex, after all.
  3. Everyone will come up with slight differences in implementation.
  4. Everyone will be deluded into believing that it is the best thing which ever happened to them, since now everything is uniform and consistent.
  5. Fuck.

You have now created a massive clusterfuck. If someone comes along and complains, he'll be lynched for questioning The Holy Authority of The Standard. What's even more fun, eventually someone will come up with an even more complex standard to fix the issues with the previous one. It will seem like a good idea to everyone, but it turns out like extinguishing fire with gasoline.

There is a trend in standards: A simple, yet flawed, standard will be followed by an outright horrible and complex one. Yet it is considered an improvement. Examples include the PC BIOS → EFI, IRC → XMPP, APM → ACPI, MBR partition table → GUID partition table, ... In each case there are some real improvements, but the overall result is much worse and ridiculously complex.