Re: Concurrent MVars for Guile

[Mark H Weaver <mhw@netris.org>]

Hi!  Reviving an old thread...

Andy Wingo <wingo@pobox.com> writes:

> On Sat 14 Sep 2013 15:59, ludo@gnu.org (Ludovic Courtès) writes:
>
>> It looks like a useful tool to me.
>
> FWIW I totally agree!
>
>> I don’t like the name “mvar” (“monadic variable”, I guess).  Perhaps
>> “synchronized box”, or “transactional variable”, or...?  The downside of
>> choosing another name is that people familiar with the concept won’t
>> recognize it first-hand.
>
> I guess I can see this.  I prefer "box" to "variable" fwiw.  How about
> "tbox"?  (The T for threadsafe or transactional or something; perhaps
> this is a bad idea.)

My preference would be to keep the "mvars" name, however imperfect,
simply because many people in the neighboring Haskell community already
know them as "mvars".

There are many different ways that one can design threadsafe and/or
transactional boxes.  In the future, we might very well support some
slightly different mechanism that could just as easily be called tboxes,
and then we will have to invent yet another name.  Given that these
mvars have identical semantics to the ones in Haskell, I think it's
probably best to just reuse the name that already exists.

However, I don't feel strongly.  If you and Ludovic can agree on a name,
I'll rename them.  Mainly I just want to get this module, or something
like it, into stable-2.0 relatively soon.

What do you think?

      Mark