Language spec

> > This specific problem shows that we need to do something > > about it. We are currently in a state where the > > implementation equals the language definition. The > > incompatibility list (PHP 3->4) would not have been > > necessary, if we had a formal language definition. > > > > Once we have such a language definition, people can still use > > side-effects of a scripting engine, but they won't be able to > > claim that those side-effects are valid. And thus, other > > language implementations would not have to simulate those > > side-effects. > > I think it would be a good thing to start writing a language spec. I'm not > quite if there's a standard format but it would be nice to start working on it. > It will allow us to review all of the language aspects and only include > things which were really meant to be. > Do you think we can pull it off in a reasonable manner? Yes, I think so. While the language itself is not very complex, a commonly agreed-on language specification would prove very useful to the community and the continuity of PHP. I haven't been involved with designing a language specification before, so feel free to comment on this approach. Task 1: Defining the syntax of the language The syntax can be represented in the usual pseudo grammar style, which has been used by K&R and many ANSI/ISO language standards before. Task 2: Defining the semantics of the language This step basically comes down to defining what meaning the various pieces of the syntax have. I.e. the concept of a function, variable scopes, overloading, et cetera. Does it make sense to set up a separate workgroup for defining (a) the procedure and (b) the language spec itself? - Sascha