Re: Idea...

[Cyrus Roton <croton@ridgecrest.ca.us>]

On 8 Jun 1995, Dave Althoff wrote:

> 
>  (is it a coincidence that SCSI allows up to seven devices?!).
>
No, not a coincidence. Three data lines allow eight device codes 
(from zero to seven). Four data lines would have allowed sixteen codes, 
but the chain probably would have been too long and would have 
required one more wire.
 
> Why doesn't somebody develop a universal Ethernet interface to conform to
> the SCSI standard?  Obviously, I'm thinking in terms of the Apple ][
> market, where SCSI interfaces exist, but (for the most part...I wont get
> into that discussion...) Ethernet boards do not.  I'm not talking about
> connecting machines together in a SCSI chain, but rather some kind of
> adapter...Ethernet in/SCSI out.  This would allow machines which do not
> have Ethernet capability (such as the old Macs, as well as the ][ series)
> to join into the various Ethernet networks.  In fact, it would also be
> possible to use such an adapter to form the backbone of a high-speed
> Ethernet-like network, capable of higher data rates due to the awesome
> transfer rate of SCSI devices.
> 
The Apple II market is too small to make it worth while. As for speed, 
note that the SCSI buss is "parallel", while communications are "serial". 
One byte on a serial line requires ten time-spaces, whereas on a parallel 
buss, one byte requires one time period. Obviously, serial communications 
tend to one tenth as fast as parallel communications. A SCSI interface to 
a network would require a parallel-to-serial converted, and all that 
speed would be lost. You can not send parallel data over phone lines and 
radio with out at least ten channels (which is not practical nor economical).

Although you have not considered all the aspects of the topic, it is 
obvious that you are thinking. When enough people offer enough ideas, who 
knows what might come of it. Anyway it was an interesting thought.

Cyrus Roton