I could have used only the opaque identifiers as folder names but I think it will be more approachable for the intended recipients to be able to read the titles rather than always cross-reference a separate file.

"just use the item title as the name of the folder, how hard could the character normalization to adhere to OS naming rules be?" - I said to myself 30 minutes and 100 bash history lines ago

I turned on follow requests when I took a week off of social media and decided to keep them on when I came back. Wasn't sure if I'd develop a policy for approving new follows, but I've now settled on:

- no product accounts (people who work on products, ok)
- no accounts that lack posts or a bio

Not sure I posted an intro before but since I periodically archive and delete, here's one again:

I do digital repository things (sometimes broadly, sometimes narrowly construed) at an academic library. But my background is in archives and history with a very brief side path (i.e. internships) through investigative journalism and government transparency advocacy in the late oughts when I was between types of grad schooling.

TFW you're trying to figure out why your wget/WARC capture is missing items and Googling an error message turns up a solution from @inkdroid https://lists.gnu.org/archive/html/bug-wget/2015-06/msg00021.html

I worked out a way to compare -300,000 downloaded web pages but it turns out that probably wasn't necessary.