I don’t know which change did it, but I think I detect an improvement.
Yes, I only changed one parameter at a time.
Yes, nothing changed when tested.
Yes, it got better on its own… which suggests to me that one of the changes I made worked but that cacheing somewhere along the line hid that improvement from me at the time.
So now I’ll wait to see if Gill detects an improvement and then slowly add stuff back in one at a time if she agrees that things have speeded up. When it slows again, I shall have identified my culprit. On the other hand, there may be no detectable improvement for the reader.
I hate computers.