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

Re: [sc-dev] [APPROVE] asBoolean

On Oct 28, 2004, at 5:31 PM, ccos wrote:

the only thing really Cish in here is Number returning false if it is zero, which is also the thing i am most ambivalent about and which i'd be glad to drop.

equating Object with true and nil with false is also C semantics where NULL is false and non NULL is true.

i don't really see how the rest is Cish in that you have to explicitly ask for the behaviour. the main point of the idea is precisely to have object return true. you use it in circumstances when you don't want to have to write a whole lot of support code to test whether something is false or nil or an object and you want to write a simple if statement to check it. in javascript which is actually a pretty neat little language imho you can just go "if( someVariable )" and nil is treated like false and objects like true.

Not requiring a boolean type can lead to hidden bugs. If you were to write if (a = b) there would be no error in javascript even though you probably meant if (a == b). A number of questions get posted to the list where the vm complains about a non-boolean type in an if statement -- which is a good thing. It caught an error which otherwise would have silently and incorrectly worked.

if (a.isNil) is aready inlined specially by the compiler and does not require a message send.

If this semantics of asBoolean were added to the language I think it would beg the question of why it is not the default semantics.

--- james mccartney   james@xxxxxxxxxxxxxx   <http://www.audiosynth.com>
SuperCollider - a real time audio synthesis programming language