[TI-H] Re: Calc Net
[Prev][Next][Index][Thread]
[TI-H] Re: Calc Net
At 19:42 2001-05-02, you wrote:
>So to recap:
>
>We can wire all the calcs strate together (red to red, black to black, and
>white to white) And hope that one; we don't fry the link port geting a
>input of
>12 insted of 1 and 0. And two let the calc handle all of the protocals.
frying the linkport shouldn't be a problem if the ports are open-collector,
wich they seem to be, and that all places I have looked have claimed
(again, havn't tested it myself)
this kind of output is made for having more then one output on the same
bus. ofcourse, if more then one try to send at the same time, the data will
be distorted, but it shouldn't be any hardware problem.
>We can use a centeral router/hub, and let the router do most of the
>protocals,
>and let the calc's do the remaining front end stuff. A 386 or a laptop with a
>parallel port could handle between 4 and 8 calc's and act as a file/simple
>web
>sever.
>
>Or lastly we can use a percalc adpter (kind of a nic card), and let it handle
>voltage/curent and clock cycles over head issues we have with option one, and
>the portabilty of option two. Its down side would be the amount of money we
>would have to put in to it. Thoe this also might be an easyer way to allow
>cross platform connection because each link would speak the same protocal on
>the network end and whatever on the host end. So we could put memory
>expanders
>pc's, printers, etc. on the network.
nice recap.
All these three methods might be possible to hide under another layer, and
then you create a network-lib, with a uniform application interface. maybe
something similar to sockets?
I think that would increase usage.
But on the other hand, it might add to much overhead.
Anyway, the actual implementation will ofcourse be up to the guy
implementing it...
Be first, and get to decide! :)
///Olle
References: