Ask HN: What non-work task have you automated?

I have semi-automated the process of creating XML files for podcast episodes and tagging the MP3 files for upload.

I'm using BBEdit and bash on MacOS X to do both; I have a "worksheet" which is like the old Macintosh Programmer's Worksheet, if anyone remembers those - it's a text document in which you can select text and execute it as a shell command, and the results are piped back into the document.

It's "semi-automated" because I have a block of boilerplate text for each episode, and I type out the text for a short and long title, filename, and info fields. These go into bash variables. Then I just select the lines one by one and execute them. This allows me to sanity-check the results. Then I can just copy and paste the block of generated XML for each episode, paste it into the staging feed file, sync it via FTP, test the staging feed to verify that I can download the episode and that the tags all look correct, then, when I'm satisfied with it, put the XML in the production feed file.

This saves me a considerable amount of hand-editing of the XML and MP3 file metadata, which is quite error-prone. "Fully automated" was not really going to help in this case since very likely that would just automatically publish metadata and XML with errors.

This is great! Could you please provide any instructions on how you achieved it?