On Wed, 13 Feb 2002, David Harris wrote: > Saying "the implementation is the specification" is silly and bad > programming practice.
It may not be the best practice, but sometimes it's just the nature of developers. It's not uncommon for specifications to lag behind the implementation. Most tcp/ip implementations pay some homage to BSD UNIX. Other early tcp/ip implementors found that some things weren't spec'd and they were very glad a clean, easy to read implementation was available. > It's a waste for the community to reverse engineer the protocol so > that other implementations can be created. This smacks of what happens > with closed source software: the protocol has to be reversed > engineered. Reverse engineering a black box is nothing like reverse engineering an open implementation. Even if C, Perl, or Java isn't to your taste, you should be able to read it. The vast majority of available source is in one of those languages and they're more similar than different. 20 years ago you had to be able to read other folks Fortran so you could write decent C programs. > I'm all for Robert contributing and publishing his protocol docs. But > this does not remove the fact that OpenSRS really should be doing > this. OpenSRS is trying to do so, but it's hard to write good protocol docs. They'll never realize everything that should be in there without significant support from the developer community. > Almost two years and counting (I think) without Protocol docs: just > shameful. Given the choice of having them work on the things they've been working on (adding other registries and improving the transfer process come to mind) or writing better protocol docs, I'm really glad they prioritized their time the way they did. -- </chris> "Outside of a dog, a man's best friend is a good book. Inside of a dog, it's too dark to read." - Groucho Marx
