RE: Re: Re: Modify the IE Browser

> From: Nathan M. Andelin
> 
> The 5250 interface is simpler from a programming point of view.  The
state
> of the program IS the state of the user.  I think a one to one
> relationship is still valid for many applications.

The funny thing about this is that it WASN'T originally the case,
especially on the IBM midrange.  Back on the old System/3, we had
something called a NEP-MRT (Never Ending Program, Multiple Requesting
Terminals).  What happened was that you had a single program that kept
state for all users.  Terminals would "ATTACH" to the program, and the
program would wait for a request from any attached terminal.  When it
received a request, it would identify the terminal, and then set the
program state based on that terminal's previous saved state.  At the end
of the transaction, it would save the state and send the data back to
the terminal.

Does THAT sound familiar?  <grin>

Joe

P.S. I'm reasonably certain you can still do the same thing today, BTW.
I think it's occasionally done with things like barcode scanners and the
like; they use 5250 emulation middleware to communicates with a central
application that services all the devices simultaneously.