[boost] Re: Moving Boost.Thread forward

From: Michael Glassford (glassfordm_at_[hidden])
Date: 2003-12-17 06:50:16

Daniel Krügler (nee Spangenberg) wrote:

> 
> Thorsten Ottosen schrieb:
> 
> 
>>In this regard, I would also be interested in peoples' oppinion about Kevlin
>>Henney's threading approach as
>>he presented at the ACCU 2003 conference:
>>
>>http://www.two-sdg.demon.co.uk/curbralan/papers/accu/C++Threading.pdf
>>
>>Even though I saw the talk,  I could use much more detail explanations for
>>rationale's etc.
>>Anyway, his approach
>>is very different to normal threading libraries.
> 
> 
> I am also strongly interested in the answer to that question, which I asked
> William
> Kempf a while ago.
> 
> (Re: Boost proposals accepted by C++ committee, Fri, 25 Apr 2003 11:12:20 +0200)

Please note that I do intend to consider this, but it will be quite some 
time before I am able to because my first priority is to get more 
familiar with the library and work on bug fixes; then possibly to move 
some changes out of the thread_dev branch of CVS into the main branch; 
then to think about suggestions for library improvements.

I believe Bill Kempf's original vision for the Boost.Thread library was 
to implement low-level "plumbing" first, then to implement higher-level 
abstractions such as mentioned in the above link using the lower level 
code. I agree with this approach. I'm pretty certain that, in his mind, 
the Boost.Thread library is still the low-level plumbing stage, and that 
he wanted to get that firmly in place before spending a lot of time on 
designing or implementing higher-level layer or layers. I'll likely 
follow a similar course.

Mike