Re: Re: Approaching the First Draft [of a language spec]

On Mon, 20 Nov 2000, Rasmus Lerdorf wrote: > The way I see this actually working is that someone sits down and writes a > first draft of this thing. Probably quite incomplete, but at least it > will lay out the framework. From there we complete it as a group. > > Historically not too many things have been accomplished through concensus > and voting and whatever other structure you might put in place. You need > to start with some sort of base usually hammered out by either a single > person or a very small group. > > At this point I would prefer to see someone like Sascha just sit down and > write this base document. It should be a complete and unambiguous > language spec using some sort of established language description syntax. > However, it should also include mere-mortal interpretation hints. > Something like the following. I am using BNF since that is what I am > familiar with, but there might be some better alternatives. I agree with everything above. > I think the important part is to have the combination of the > rigid definition of something like BNF combined with the > user-friendly explanation and example text: I don't think though that the immediate target here should be the end-user. Let's define the semantics clearly and unambigously. We can add user-level descriptions later. - Sascha