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

Re: [f-cpu] Re: Display Terminal Clusters

On Thu, 6 Sep 2001, Richard E. Hartny wrote:

> I couldn't let it pass!!!!!!!!!
> Juergen Goeritz stated----
> forget it!  in the time of networking it's useless to focus
> on display terminals at all.  you just need a single display!
> Only if you go for real-time cluster 3D simulation in some
> flight simulators or weather simulators will be more than a
> single display involved'
> And please stop your announcements on this mailing list now,
> Thank you
> JG
> And further;
> Look at Richard, he thinks the display is the center of all.
> He is misleaded too.  He just sees it from his application.
> And my response is --------bullshit to both of your uneducated responses.

Seufz. Sorry to have scratched your pride.

> The number of terminals in any system is guided by the Processor performance,

I am more thinking in distributed systems with a lot of computation
on a lot of nodes using pipelined or parallel processing. Most of
the problems I work on are flow through problems for real-time or
optimization stuff. The user interface is only for administrative 
and management purposes in my work.

But I worked with 32 terminals on a normal UN*X machine in the 1980's.
It worked well for the ttys without a lot of processer performance
compared to todays performance.

> if you don't want the user (people) to wait for data responses.  Most are presently unhappy with Novel networking because of WAIT time with peak user requests.
> I have addressed the unhappy ones with the design of the ERIN32 -- an M2M architecture.  That is exactly where the f-cpu was at its inception quite some time ago.  And the f-cpu is targeted toward Game Machines (single user) and Flight Simulators,  and Weather simulators - very high performance stuff.  I am targeting the World of Business where more than one terminal is required.  
>     I stated all of the above before you entered the scene - Yann Guidon knows this as well as others.
>     Since all of my designs use SCHEMATIC input and f-cpu is VHDL, the only help I can be is to provide what I have - if requested.

How about a bottleneck analysis of the applications instead?
I never said that using a window system is the best. I still
prefer to read and write ASCII on a tty and I know that the
LPT/COM/Keyboard I/O handling in a PC is long outdated. ;-)

> I will continue to provide tidbits of information to f-cpu whether you like it or not.

SOS (sound of silence).


To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/