Re: "Mission Statement"

At 15:17 21-11-00, Sascha Schumann wrote:

Well, that's true, but to the average user, saying what's going to work on a 32-bit platform, what's going to work on a 64-bit platform, and what he can always rely on is the best thing.

    Saying what works on 32/64-bit platforms would make the
    specification implementation-dependent. That is something I
    want to especially avoid.

Why? The spec is not for academic purposes, but for people to use in everyday life. That's something that is useful in everyday life, and therefore it should be in.

    Well, and if Microsoft now chooses to add PHP to its .NET
    platform, there is basically nothing to prevent them from
    writing a replacement for the Zend engine. If we had a
    language spec, at least the outcome would be clear and useful
    to the PHP community. PHP needs continuity which we cannot
    reach with the current status quo of a pure reference
    implementation.

I don't think that a compiler spec is useful, helping Microsoft's .NET (or other forking bodies, like GPL fanatics) team isn't on my agenda either and I don't think that it would be beneficial for the PHP project at all (and I'm sure most people here feel the same). There's theory, and there's reality, and the spec you're talking about is a purely theoretical, useful-for-forking, fairly-useless-for-practical-use type of spec. Zeev -- Zeev Suraski <zeev@zend.com> CTO, Zend Technologies Ltd. http://www.zend.com/