Thanks Paul, but I'm having a hard time finding the precise version I would like to archive on any ftp mirror. My scrape is actually working quite correctly now, though, since I added a sleep in there -- the source and machine-installation instructions are tidily tucked away in different directories, with names, locations, and success/failure logged into a key-value (more of a dictionary-ish, really) text file. I get that an ftp request is more civilized, but this scrape is quite convenient for me. If it is more palatable to the community, I can increase the sleep time in the loop to a couple of minutes or even a few or more, and throw it on one of my raspberry pis and forget about it for a while, since I'm not in a major hurry.
Have a great Monday folks,
John