NH> The embedded systems with which I worked had no mass storage
NH> devices. Programs were kept in EEPROMS and data was held in
NH> static memory (which was _really_ scarce).

BB> Strangely enough, that doesn't preclude it, Neil. It's 
BB> ancient history now, but it was possible to fit loads of 
BB> EPROMS and bank switch them.

Holy mackerel, none of the devices with which I ever worked even 
considered that.  Would the code meant for coordinating bank switching 
reside with the system code, or would each application that used it be 
responsible for rolling their own?

NH> There resources required for storing the overlay manager weren't
NH> there.

BB>     In bank switched systems, it's in bank 0. So is the 
BB> scratchpad RAM. However - this *is* ancient history. There 
BB> should be no need for it nowadays. Storage is now cheap and 
BB> plentiful.

Another problem is size.  The units for which I programmed were shaped 
like bars of butter, only larger (16 cm x 6 cm x 6 cm).  That included 
the input device for reading bar codes (at one end) and the output 
device (the other end) for connecting to telephone lines (we used a lot 
of cell phone connections).  There wasn't a whole lot of room for 
anything like another EEPROM.

þ CMPQwk 1.42 999

--- Maximus/2 3.01