[concurrency-interest] RFC -- Java7 java.util.concurrent plans

Bob Lee crazybob at crazybob.org
Thu Dec 11 12:52:28 EST 2008


I'd argue that "ReferenceMap" *is* consistent with other types. For example,
we don't call DelayQueue "ConcurrentDelayedPriorityQueue".

Bob

On Thu, Dec 11, 2008 at 9:37 AM, Endre Stølsvik <Online at stolsvik.com> wrote:

> On Wed, Dec 10, 2008 at 17:02, Joe Bowbeer <joe.bowbeer at gmail.com> wrote:
> > On Wed, Dec 10, 2008 at 7:43 AM, Bob Lee wrote:
> >>
> >> On Tue, Dec 9, 2008 at 4:35 PM, Doug Lea wrote:
> >>>
> >>> 3. ConcurrentReferenceHashMap
> >>>   A concurrent hash map allowing weak, soft, etc keys and values.
> >>
> >> BTW, I like the name "ReferenceMap" better.
> >>   - It's obviously concurrent because it's in the "concurrent" package.
> >>   - It doesn't makes sense to use anything but a hash-based
> implementation
> >> with references.
> >>   - We won't want a non-concurrent reference map because it will retain
> >> garbage for too long.
> >>   - I like shorter names.
> >> Bob
> >
> > I like shorter names, but I also like consistency.
> >
> > ConcurrentReferenceHashMap is more consistent with the naming of its
> sibling
> > classes, such as ConcurrentHashMap, whereas ReferenceMap is more
> consistent
> > with the naming of interfaces.
>
> Agree - consistency and clarity. IMO, this is way more important than
> the 'spoken name argument' that Bob gives - it is really not that
> often I tell Joe to use a particular class - it is much more often I
> have to understand a type of a variable, remember a name, read the
> code. I find it easier to remember a clear, consistent name, than a
> "catchy" one - like in what aspects of its function was it they had
> left out in that name? The "Concurrent" or "Hash" aspect? Ah, both,
> actually. I find it annoying to have to check the package name to
> understand a class's function. On the point of "brain parsing time"
> that can be raised in such naming discussion, it is shown (!) that we
> humans tend to parse entire words as an image, rather than reading
> each letter (which actually gives an important naming-argument: Don't
> let names be the same length and visually look the same). On writing
> time, one code in IDE's, not notepad. Lastly, taking this shortening
> logic further, one will end up with name clashes and hence
> misunderstandings, like in "Awh, damn, it is a NORMAL DingDong, not
> the package.blue version - no wonder everything crashes!".
>
> Endre.
> _______________________________________________
> Concurrency-interest mailing list
> Concurrency-interest at cs.oswego.edu
> http://cs.oswego.edu/mailman/listinfo/concurrency-interest
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.oswego.edu/pipermail/concurrency-interest/attachments/20081211/804c0ba3/attachment.html>


More information about the Concurrency-interest mailing list