Unofficial wish list status.(Jul 2008)

superdan super at dan.org
Thu Jul 24 06:59:05 PDT 2008


Lars Ivar Igesund Wrote:

> Simen Kjaeraas wrote:
> 
> > Fawzi Mohamed <fmohamed at mac.com> wrote:
> > 
> >> On 2008-07-24 07:43:55 +0200, Sean Kelly <sean at invisibleduck.org> said:
> >>
> >>> Walter Bright wrote:
> >>>> Sean Kelly wrote:
> >>>>> To be fair, you know the features that the test cover work :-)  But
> >>>>> the point remains.  With this in mind, why does contract checking
> >>>>> still not support inheritance as described in the spec?  Contracts
> >>>>> are one feature that originally drew me to D and it's a shame that
> >>>>> they're only half implemented.
> >>>>  Because the contracts haven't drawn the interest that the other
> >>>> features of D did that also needed work.
> >>>  Seems I'm always in the minority :-)
> >>>   Sean
> >>
> >> For what is worth, I also like contracts :)
> >>
> >> Fawzi
> >>
> > 
> > Thirded.
> > 
> > -- Simen
> 
> Fourthed.

contracts are like gym memberships. most people agree they're good. many actively want them. few actually use them.

i wasted a couple of hours of my miserable life reading "design by contract, explained". what a heap of shit that book is. the moron defines a stack. not even generic. a stack holding int. by the time he was done adding contracts to it, the poor stack looked like an old hooker after a gangbang.

that being said, shit walter. didn't they teach you in school you either do shit well or not at all. either dump contracts or make them specifiable in interfaces. if that were the case by holy guacamole i'd use them. interfaces with contracts are about the only sensible, palatable use of contracts. why? because the only interesting shit happens when contract is separated from body. you specify some shit in a contract then force someone *else* to abide from it. contracts with body are a bit like writing a promise to yourself. (save for inheritance but i bored myself talking already.)



More information about the Digitalmars-d mailing list