I got some time with the car and followed the suggestions. I'll put the results into the first post to be more useful for others with this same problem.

I also swapped in a new OVP relay and the ABS light went out. Hurray!

OK, the Edit option on my first post has timed out so I'll have to report here.

Faulty coolant temp sensor (Post #4): Ordered a new one and installed. Checked the old against the new in a water bath. This is essentially a switch. It is open until about 115-120F and then closes with a slight resistance. The resistance drops as the switch warms up so perhaps this is also used by the ECU. Here are numbers for the new sensor 60-115F open; 120F 1.9 ohms, 130F 1.7 ohms; . . . and so on . . . until 150F 1.1 ohms. The switch I pulled out had a '93 build date on it (original) and matched all the above numbers within a small margin of error. Conclusion: I don't think the temp sender is the problem.

No diagnostics from Socket 3. (Post #7): This is a Canadian car and it has the California setup with the diagnostic module (DM) under the steering column and a button on the socket cluster to read without a blink code reader. I pulled and reset the DM multiple times, cleaned the pins and nothing changed. By the way, getting the DM out is a slight puzzle. There is a dial on the front. Turning the dial from I to O moves the DM away from the contact pins. To get it out from that position, you have to lift (or depress, can't remember) lock tabs on each side near the back side of the DM. It then comes out easily. You do not need to remove the mounting bracket with the DM! Anyway, I checked Socket 3 with the DM out and there was no longer short to ground. Put the module back in and the short reappeared. The car runs the same with the DM in or out. Conclusion: Looks like the DM is bad, and not the wiring. I've sourced a used one and will swap it in next time I have access to the car. It would be nice to be able to read the codes from Socket 3, but that will have to wait. If these are indeed just for diagnostics and the DM does not feed any info back to the ECM, then probably it is not the cause of my problem.

MAF bad (Post #8) I connected and disconnected the MAF cable and there was no discernible change at idle. I fiddled mightily with the cable while plugged in to make sure there was no intermittent connection problem. I pulled the MAF and cleaned it (did this last December but why not?). No change either plugged in or disconnected. What I didn't do, and I'm kind of mad that I didn't, was to disconnect the MAF and drive around a bit. I only monitored idle. Conclusion: I may have a bad MAF module but then why does the car run so well?? Not super eager to buy a used one on the web, and new ones are a little pricy, but this seems like the next step.

I don't expect the problem has been fixed, unless by some miracle the OVP relay was messing things up. My daughter is running a mileage check and will report back. In the meantime I may visit my independent and see if they have a known good MAF module.