Jeff Garland wrote:
[...] Yeah, I know you are trying to extend the original work, but as a group it doesn't seem like we have that focus. Also, I wasn't sure if you were going to look at Hugo's initial implementation or not. At the end of the day, I, like Doug, don't really care where or how the code appears -- so if people want to abandon Hugo's initial effort that's fine, but it would seem like a shorter path to build on the original work. I haven't looked at the new stuff in the sandbox or read the other thread yet...
I went through all Wiki pages about Boost.Socket and Boost.Multiplexing including the ones set up by Hugo. I didn't look however at his implementation for reasons I explained in another mail (see http://article.gmane.org/gmane.comp.lib.boost.devel/121743). I am sure we will not create anything new which has never been done before - so we might reuse Hugo's or anyone's source code. I don't care neither about who implements the library. However I would like to see decisions about the design first before anyone starts implementing. Hugo created eg. an acceptor and connector but there is a discussion going on between Michel and me in another thread what they really should be used for (see eg. http://article.gmane.org/gmane.comp.lib.boost.devel/121908). I try to summarize what is discussed in this list at http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl?BoostNet. In the end of the day we should have a long list of requirements and lots of links into this list for explanations. The summary of design decisions and explanations should help to avoid that one day we start all over again because everything is questioned. As noone knows why Hugo designed his library the way he did I don't think it should be used as a base. Boris