[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: dynamic vs. static typing




Last time I looked and programmed Cobol had types.
It had a very elaborate description language for putting
together lines and records and files and what have you.
(I doubt it had a soundness theorem, but I bet it was as
(un)sound as Algol's.) I don't think the problem was types
but that people used the wrong type.

Robin went wrong when he said typed programs can't
go wrong (I told him so) and he's wrong again.

-- Matthias




On Sunday, November 30, 2003, at 03:34 AM, Geoffrey Knauth wrote:

> Saw this statement by Robin Milner, and it reminded me of the LL1 
> debate:
>
>> [...] type structures could have avoided a lot of the Year 2000 
>> problem. We know that nothing serious went wrong, but in the two or 
>> three years before that people didn't know that it wasn't going to go 
>> wrong. If they had used types, which had been around for two decades, 
>> they would have been able to know that certain things couldn't go 
>> wrong. The theories were around, which would help people to know that 
>> certain things couldn't happen, barring physical accidents or bad 
>> implementations of course. We are used to the idea that theories can 
>> sit around and not be used. We just have to accept it.
>
> from:  http://nick.dcs.qmul.ac.uk/~martinb/interviews/milner/
> scroll down 80%
>
> Geoffrey
> --
> Geoffrey S. Knauth | http://knauth.org/gsk