|
Boost : |
From: Andy Glew (glew_at_[hidden])
Date: 2000-01-15 17:04:13
> I agree, just "Queue" would not be a good name.
>
> How did you all feel about using the name "GeneralizedQueue", where
> the "Generalized" part is meant to mean that it can be things other
> than FIFO.
(1) IHateMixedCase and prefer_underscore_separation.
Sorry, religious topic, although shared_ptr>T> and auto_ptr<T>
have established IMHO the right precedent. ;-}
(2) I'm uncomfortable with "Generalized Queue", since there
is a subfield of queueing theory whose name is roughly similar.
Your "generalized queue" has little to do with a G/1/1 queue,
where "G" stands for a "general" distribution, but it's still
confusing.
This is a minor complaint. I could live with GeneralizedQueue.
Or the slightly shorter generic_queue.
(3) Why not "buffer"?
"Buffer" has long standing use in computer science, as something
that holds things for a while, but not permanently.
"Buffer" has no ordering connotations for me.
Historically, IBM in the 1960s used "buffer" for a "FIFO queue",
but the Berkeley terminology seems to have taken over.
Moreover, some of the earliest data structure references I have
seen say "pushdown buffer" or "stack buffer" for what we now
call a stack, and "FIFO buffer" or "queue buffer" for what we now
call a plain old queue.
I.e.
Queue = FIFO buffer
Stack = LIFO buffer
Priority Queue = buffer with some priority rule...
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk