On Thu, 2013-07-18, Chris Angelico wrote: ... > You can certainly do your server-side programming directly in Python; > in fact, I recommend it for this task. There's no reason to use HTTP, > much less a web framework (which usually consists of a structured way > to build HTML pages, plus a bunch of routing and stuff, none of which > you need). All you need is a simple structure for separating one > message from another.
> I would recommend either going MUD/TELNET style > and ending each message with a newline, or prefixing each message with > its length in octets. Both ways work very nicely; newline-termination > allows you to use a MUD client for debugging, which I find very > convenient It's definitely the way to go. It's not just MUDs -- a lot of Internet protocols work that way. Netcat is one popular client for talking to/debugging/testing such servers. No doubt MUD clients too, but last time I did such stuff was in 1993, and I think I used telnet ... In fact, I'd design the protocol before starting to write code. Or better, steal some existing chat protocol. Like a subset of IRC. There's also another question in the original posting that bothers me: paraphrased "do I need to learn database programming to manage users"? No! Unix does fine with plain-text files. Managing credentials (understanding cryptography, setting up a support organization for resetting lost passwords ...) is non-trivial though, so try to do without such things at first. It's not obvious that you should need an account for an experimental chat, anyway. /Jorgen -- // Jorgen Grahn <grahn@ Oo o. . . \X/ snipabacken.se> O o . -- http://mail.python.org/mailman/listinfo/python-list
