Subj : Re: Challenge: Multithreading & Synchronization
To   : comp.programming.threads
From : Uenal Mutlu
Date : Wed May 18 2005 07:55 pm

"Sergei Organov" <osv@javad.ru> wrote in message news:d6fqsp$qfb$1@n6.co.ru...
> "Uenal Mutlu" <520001085531-0001@t-online.de> writes:
>
> > "Sergei Organov" wrote
> > > "Uenal Mutlu" writes:
> > > > "Sergei Organov" wrote
> > > > > "Uenal Mutlu" writes:
> > > > >
> > > > > > /* Here's a real-world synchronization challenge for accessing shared
> > > > > >data from multiple threads (some call this "thread synchonization"):
> > > > > What is the aim of posting the challenge? If it's to illustrate
> > > > > how much easier it is to implement inherently broken design using
> > > > > recursive locks, then your challenge is indeed an excellent
> > > > > example.
> > > >
> > > > It is a real-world example, ie. the requirement is there. Just find a
> > > > solution.
> > >
> > > Well, you didn't answer. What is the aim? Why should I bother to fix your
> > > broken "real-world" design? What is the aim? I.e., why should I or anybody
> > > else be interested in fixing it? To teach you programming?
> >
> > If you feel it needs to be fixed then just say what part you think should be
> > fixed.
>
> I'm afraid it's so broken that it can't be fixed in parts.

What exactly do you mean is broken?

> If you indeed need something working immediately, create single recursive lock and
> lock/unlock it at entry/exit to every function, deliver the application,
> and then start to re-design your application ASAP, where ASAP in fact
> means "as soon as you actually understand where and why it is flawed."

Can you give an example for just the member function f2(), or more if possible?

> > What's unclear to you? The aim is to collect multiple solutions to the
> > problem.
>
> You see, for me your challenge sounds like this: "For undisclosed
> purpose I've created this messy broken skeleton and called it
> "real-world synchronization challenge". How do you fix my problem?".
> For me it's unclear what "solutions" you are going to collect then.
> IMHO, the best solution is to throw it away, sorry.

I don't feel so. I would say it is a well-designed classic real-world example
for shared access to a class from multiple threads.
How would you have designed or formulated it?

.