[MUD-Dev] Re: clients anyone?...

Adam J. Thornton adam at phoenix.Princeton.EDU
Sat Aug 15 00:58:30 CEST 1998


On Sat, Aug 15, 1998 at 02:11:55AM +0100, Andrew Wilson wrote:
> 296 bytes?  That's useful information, and perhaps has some bearing
> if you're working towards developing a highly interractive site.
> I use an MTU of about 512b at home myself, chosen for me by my
> modem.  Where did you get 296 from, does in include the TCP header
> or is that additional?

The 296 should include the TCP header, and OS/2 and oldish versions of
Linux pppd picked 296 as the default value.  However, that was in the days
when modems were 14.4Kbps and occasionally 28.8, so it's probably safe to
use something larger.  If Win95/98/NT uses 512, that's worth knowing.  Me,
I have a cable modem; MTU = 1536 bytes, it's just an Ethernet as far as my
machine is concerned.

> Have you considered using the SSLeay libraries?  It should be possible
> now to teach both server and client about SSL, acting as a drop in
> replacement for the normal wide-open socket connection.

Y'know, I've got them lying around here, ever since I built Lynx with them.
Good idea.

> And then what?  Develop a totally-assed client?  Forgive me.
> Clients are not trivial pieces of code. Ten thousand lines will
> buy you something that can just about tie its own shoe-laces and
> promise not to crap on the carpet. ;)

Granted.  But getting something minimal will allow me to get the server to
something I can live with to develop a real client.  But to get the server
working, I'll need *something* a little more sophisticated than an
interactive telnet session.  Once I have a client that will support my bare
minimum tokenized protocol, then I can get the server working reasonably
solidly.  For two thousand lines I can get something that will only crap on
the carpet occasionally, and might be able to fasten Velcro, which will do
for my own testing.

Adam
--
adam at princeton.edu 
"There's a border to somewhere waiting, and a tank full of time." - J. Steinman




More information about the mud-dev-archive mailing list