Terminals in the mid-80s did have sophisticated features, including multiple screen buffers (and larger screens than 80x24), but 512K seems like a lot.
The ROM code does not seem to be doing anything as simple as taking input from an ASCII terminal on an SIO, although I just started looking at it. I might be able to guess at parts of the I/O map (based on how chips are being initialized), but it does appear there might be some sort of parallel-connected keyboard or input device, without which we are probably at a larger disadvantage. There is an interesting string "NETWORK ADDRESS" in there, which might imply that this board was on some sort of network (not necessarily - and probably unlikely - Ethernet). There is also a string "PROM DIAGNOSTICS" which may indicate the whole, or partial, purpose of the ROM (I've not fully discovered the circumstances of when that string is emitted to a peripheral). There is also a "LOAD FILE ERROR" string, which may indicate the ability to "boot" or run external software (although "file" may not refer to disk - it could be Intel HEX "files" loaded over some channel).
One thought is that this board was part of a development system used for creating firmware to run on terminals (or other appliances). That might explain the 512K RAM - perhaps this even ran CP/M 3 or MP/M, or some terminal emulation software (isolated in alternate banks of RAM).
