From: joel de guzman (joel_at_[hidden])
Date: 2001-05-24 21:43:10
----- Original Message -----
From: "Vladimir Prus" :
> Theoretical study is not likely do give us much. Backtraing parser have
> exponetial complexity in worst case, and "high priests" would find their
Practically speaking yes, but a more thorough investigation will be
I am doing some research now but I guess that will take time. I'll have to
prioritize the practical aspects. I'll do benchmarks and stuff as you said.
> Therefore, what I would like in the end, is the statement in documentation
> the form: "Spirit has been used in practice for such-and-such tasks, has
> shown such-and-such performance and is considered suitable for those
> Hope that once it works with g++/bcc somebody will try it for his real
> >> 2. Syntax
> >> Proposed syntax has prefix "*" and "+". This is likely to confuse
> >> knows what EBNF is. Reuben Fries and later David Abrahams have already
> >> proposed different syntax. Is it better? I tend to say "yes".
> >I disagree.
> Well, our opinions differ. So, we need, as Greg Colvin put it, "argument
> and experimentation towards finding a nice, clean syntax.". I don't
> any argument can be usefull there -- too much is based upon personal
> preferences, so once again we need experiment and (a lot of) people
> on syntax. (And probably poll? :-))
Yes, very subjective. I still have a strong opinion towards retaining the +
and the * kleene star. These are used very often and I don't want to type
repeat<n>(a) many times. I just want to do *a.
Look at it this way. C uses the * for multiplication (infix) and pointer
(prefix). If they (the original authors) though that this will be confusing,
would have used something like deref(ptr) instead. Would you have wanted
I'd like to propose changing >> to +, keeping the prefix +, * and adding
Joel de Guzman
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk