I haven’t made a lot of progress on my projects. I did create a Scaleway VM and shoved an OAI harvester on there that’s happily downloading the arXiv’s backlog of metadata. I can also parse the XML it fetches, and have some ideas about how I’m going to store it.
This lack of progress mostly comes from me being nerd-sniped into thinking about bounded queues under load. My old work project wanted to use a FIFO queue to hold its requests. That is a bad idea, because FIFO queues perform very poorly under load, as I went a little overboard in demonstrating.
Funnily enough, that very simple thing is one of my most popular Github projects ever.