Re: A86: Radio Link


[Prev][Next][Index][Thread]

Re: A86: Radio Link




As soon as the Radio Link is complete (heh, I talk like someone out there is
working on it right now =), then we can start on all the protocol issues.
Actually, we could do the protocol stuff now, but it would be kind of a
waste if the Radio Link never showed up in the end.  Even though it may be
tedious to type in unix commands on the calc (I got pretty good at Telnet83
though.  I could do about 20wpm complete with shifts), a unix shell is the
easiest to implement since we already have VT-100 terminals.  As soon as the
Radio Link is finished, we will already have 'net access via unix (Term86,
Telnet83).  If some people want to go that extra step further to make other
clients besides straight VT-100 text, then that would be cool too.

There are many protocol issues.  One, of course, is the crappy linkport.  It
can't transfer two directions at once.  There's always that slim possibility
that when you send out a byte, one may come in the port and you lose both
bytes.  With plain VT-100, losing a char here and there isn't the end of the
world, but any kind of packet system would totally get messed up.  With
large packets of data getting thrown in and out of the linkport, we'll
encounter many more collisions.  We, of course, could have a retry system
but there are some better ways that I'll mention later (like a sync-request
system).

Anyways, the Radio Link should not worry about this linkport problem or any
other protocol issue since we can resolve it later with software.  We just
need the hardware.  I really want to stray from the topic and dive into
client programs (like browsers, chats, ftp) but we need the Radio Link first
otherwise I'd be going off on that for no reason.

-Justin Karneges [Infiniti]

>It seems to me that a good radio link that has a long range would be very
>useful. I also think that TI-UX while somewhat intresting would not be very
>useful. I not trying to discourage any one from trying it, but it does seem
>like the calculators' built in functions can manage quite nicely. The one
>thing that it might be useful for is have a server calc that could be used
as
>like a communication center. It would be wired to the net, or a LINUX or
UNIX
>server, and have like a little chat area. Users could log in with a small
>client program, and then acess the net or whatever... Then the main aspects
of
>UNIX would actually be useful. (Multi user support, security etc...) I
would
>say that a radio link is the main reason for having a OS that is
specialized
>in these tasks. It would be pretty tedious to type in all the commands on
the
>tiny calc keybord, especialy when the nice people at TI put in all those
>menus...
>