Specs, shmecs

October 28, 2005 8:12:21.178

Apparently, Dave Winer went to the Torvalds school on specs:

Linux creator Linus Torvalds began the discussion saying, "a 'spec' is close to useless. I have _never_ seen a spec that was both big enough to be useful _and_ accurate. And I have seen _lots_ of total crap work that was based on specs. It's _the_ single worst way to write software, because it by definition means that the software was written to match theory, not reality."

The crucial difference being, Torvalds is talking about an OS kernel, while Winer's "specs" are supposed to define a method for interoperability.

Comments

TCP/IP? Internet email (RFC 821, 822, etc.) ?

[Lex Spoon] October 29, 2005 5:48:23.232

I haven't read the original, but TCP/IP and Internet email are two obvious examples of spec-based software that has been phenominally successful. And yeah, they are also both examples of interoperability specs.

Also interesting is that these are examples where there were a lot of early efforts to test interoperability of the first several implementations, and the specs got updated in response to any issues that came up. It's not like Ada, where the spec got frozen before anyone had a chance to play with an actual implementation.