On Wed, Mar 24, 2010 at 12:32 PM, ToonSuperLove <sruitangkavanitsuk at gmail.com> wrote: > Hello, > > My name is Sruit Angkavanitsuk, Thailand. I am in last semester on > MSc.Computer Science at Christ University, Bangalore, India. I have read > through your GSOC 2010 idea and I am interest in Transport Plugin Project. > > However, I have some doubt in this project that I need you to point me, > once, why do we need to made it as a plugin ? why do we not made it in major > release? (like can change in setting which protocol that we will using?) > second, you have write to support for TCP, HTTP, protocol in transport > layer. So this project, how many minimum protocols that we need to made? > (I'm not sure I understand this correct, I thought that you include like > VoIP, SIP, XMPP, as well ?) > > As start earlier, I am really interest in this project, if you have any > suggestion I would appreciate to discuss in more detail.
I think the major motivation behind making it a plugin is to keep the code modular and make it easy to add new options. There are lots of interesting transports; some, like TCP and HTTP, are more interesting than others. I don't think there's a hard number on how many would be needed. IMHO a very well written, well documented transport plugin architecture with one good plugin is more useful than more plugins that are harder to improve later. (But, I should say that I'm not the one to offer a definitive answer on this.) There's another potential gsoc student interested in transport plugins (zidel_ on irc); the two of you should probably talk to each other. Also, transport plugins should probably be done in parallel with changes to FNP (the current FNP has a number of flaws, some of which would complicate the transport plugin layer). I'll send another email to devl about that momentarily; I recommend you read it. Evan Daniel
