Re: [boost] What's a Mutex ? (Was: Re: Threads)

From: Kevlin Henney (kevlin_at_[hidden])
Date: 2000-08-08 08:21:12

In message <398F5F76.2E4B_at_[hidden]>, Valentin Bonnard
<Bonnard.V_at_[hidden]> writes
>Andy Sawyer wrote:
>
>>  I concur. "P" and "V" are only interesting from the archaeological
>> viewpoint.
>
>I prefer these names because they are so strange and specific that 
>they show very clearly what we are talking about.

If this logic (?) works, perhaps we should be considering a thread
abstraction simply named T, which is fairly strange and specific.
Alternatively, what is the Dutch for "thread" or "unit of concurrent
execution"? T seems a little obvious...

>lock() can lock a file, lock an exclusive lock, enter some 
>other sort of critical section.

But it locks, right? [Hint: generic programming, resource acquisition is
initialization, etc]

>I like the names P and V because they don't mean anything to most 
>people, 

Boost code and guidelines could look "interesting" if this approach were
adopted...

>but I dislike semaphores because I have never understoud 
>how they were meant to be used. (I don't see the specific problem 
>they solve.)

They solve a problem of mutual exclusion. As pointed out in another
posting, they can be considered degenerate form of ownership-based
mutexes.
____________________________________________________________

  Kevlin Henney                   phone:  [phone]
  Curbralan Ltd                   mobile: [phone]
  kevlin_at_[hidden]            fax:    [phone]
____________________________________________________________