Re: A quick draft of a preliminary proposal for a possible version of the crossfire protocol

["Eric A. Anderson" <eanders+@CMU.EDU>]

"Carl Edman" <cedman@cedman.remote.Princeton.EDU> writes:
& "Eric A. Anderson" <eanders+@CMU.EDU> writes:
[ Line oriented packets .vs. size & data oriented packets ]

> (a) that you can't any longer use telnet or any of the other standard  
> line oriented clients to just check on a server when you have no client  
> available which makes debugging harder 
This is absolutely trivial to write. It takes 20 minutes -- I checked,
I wrote it.  I also wrote a server process in that time that will
broadcast messages to all connected clients.

> [ need an arbitrary limit because otherwise people will be hostile,
>   or not enough memory]
Yea, but then the limit can be changed based on the amount of memory
available, allowing more batching of commands.  Anyway, with the given
limitations on network bandwith, a message over 4-8k wouldn't be
practical anyway.

> (c) (for the byte counters)  That adds an extra 1 or 3 bytes to every  
> single command !
I don't think this really matters -- the ascii commands will be longer
in the general case also.  Anyway, lzw would compress out this
information if it repeats very often, so the cost is low.
Furthermore, I believe it makes handling the packets much easier
because there is an additional check as to how much information should
arrive. 
          -Eric 
*********************************************************
"It seemed like a good idea at the time"
           -The Mad Hatter
"Yes, you're very smart.  Shut up."
           -In "The Princess Bride"
*********************************************************