ISA cards are fun, working with PCI at low level is like kicking dead whales down a beach, to use a hardware hacker phrase used for something both disgusting and requiring extreme effort for small result. Part of that extreme effort is the additional work of resource allocation and writing much more complex drivers. PCI speeds need a few thousand bucks worth of tools for troubleshooting, ISA speeds, mere tens, ergo most hobbyist shops are underequipped or on the raggedy edge for PCI testing.
Then we have the "Whoa, my favorite card is over $50, so someone should build a new one for $30, despite the rare NOS chipset costing $50 a pop by itself and the rest of it costing $40, and engineers who could do this should do it for free and not expect to recover development costs either. Because I ain't paying a real world price of $150 for a new one even if it reflects actual parts cost and labor with a small reasonable profit margin." ... and most of the ppl going "But I would pay $150 for a brand new X..." have a substantially more complicated X in mind with substantially higher part and production cost.
"Oh you can't get the part, just emulate it in FPGA" yeah, the FGPA capable of it may be somewhere in the hundreds of dollars (If available at all at the present time with shortage issues)
"But I can emulate the entire machine on a $50 Pi type board" so emulate the whole machine on a $50 Pi type board cheapskate.
Most repros are passion projects, the guy doing it was/is intensely interested in the board and did it for personal satisfaction... nothing very immense will come of "Someone should..." unless the someone is halfway there already.
Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.