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

Re: A possible solution?




On Thu, 30 Oct 2003, Taylor Campbell wrote:

>
>On Thursday, Oct 30, 2003, at 15:53 US/Eastern, bear wrote:

>> You're just trolling here, aren't you?
>>
>> Things that don't exist aren't useful.  EVER.
>
>Let me rephrase this to mingle the context into the sentence:
>
>'Types such as dictionaries, sets, et cetera specified by SRFI 44 are
>never useful because they don't exist _yet_.  They can't possibly be
>useful, so they're not even worth considering now!'

Call/cc was not useful before it was implemented.  It was worth
considering, designing and implementing, but it could not yet be
used and therefore was not at that time useful.  Please note that
it didn't get anywhere *CLOSE* to a standardization process before
it had a concrete implementation and people could use it.  But it
was not useful until people could use it.  That's what "useful"
means.

Standards standardize useful things.  Plans are for the implementation
of _potentially_ useful things.  See the difference?  If it's not
useful, here, now and today, then we don't have a mission to standardize
it.

It might be worth implenting, in order to _find out_ whether it's
useful.  If the implementation, including concrete implentations
of every collection type specified, and portable means of doing
dispatch, and the efficiency functions that make it worthwhile for
different collection types to even exist, and all the other things
this guy is just handwaving about and saying "not my problem"
about, turns out to be useful, *THEN* we should consider standardizing
it.

Not one moment before.


				Bear