You guys might think I am stupid, but I think both Emmanuel and Stefano are correct. It's a matter of point of view.
The primary mission of MINA project is to provide an easy way to develop a network application. Today, many net apps are built with proprietary immature protocols because of lack of research and customizable protocol implementation. If we provide an out-of-the-box codec for HTTP, FTP, SMTP, and DNS, people will start to reuse existing standard protocol and overall interoperability in the industry will increase. (I hope!) But it doesn't mean that MINA has to host all the protocol codec. There are implementations for such protocols already. If they are mature enough, MINA team will just provide a thin integration layer for it and work closely with the authors of the protocol implementation. If there's no such implementation or it's not being maintained actively, MINA team or other folks can create a new one. From this standpoint, I think Apache Lab is the best place to start a new protocol implementation. We can discuss again about where the project should go later when it graduates the lab. Someone could think that hosting the codec under MINA umbrella brings a lot of benefit because MINA project would like to provide tooling for protocol analyze and full-featured L7 switch which needs to closely cooperate with protocol codec implementation. Other people might think such a task can be done even if the codec implementation doesn't belong to the MINA project. It's a very very controversial issue. To be honest, I think it actually depends on how big project MINA community can manage. For right now, we can't host all the protocol implementations we want to integrate with. We have too many things to fulfill to get to our ultimate mission yet, so discussing about where it should go right now is not very appropriate until a certain protocol implementation project cuts a release IMHO. Trustin On 6/13/07, Emmanuel Lecharny <[EMAIL PROTECTED]> wrote:
Hi Stefano On 6/13/07, Stefano Bagnara <[EMAIL PROTECTED]> wrote: > Emmanuel Lecharny ha scritto: > > DNS, DHCP, NTP are protocols but are also about storing data. There is > > nothing wrong in developping the protocol codec somwhere else, but I > > don't believe it should be MINA. For instance, AsyncWeb is now > > invcubating, because it was simply using MINA, and having under a MINA > > umbrella was just limitating its audience. MINA is supposed to be a > > network framework, if you start to aggregate all the protocols under > > its umbrella, I think you will soon have problems managing the > > project. > > A DNS client has nothing todo with storing data. Very true. This is something we have to think about : offering a client for each protocols. As I wrote this morning, I did had a lot of time to read all the mails, so my answer was based on what I read about writing another DNS server based on MINA. More to come in a few hours... -- Regards, Cordialement, Emmanuel Lécharny www.iktek.com
-- what we call human nature is actually human habit -- http://gleamynode.net/ -- PGP Key ID: 0x0255ECA6
