I saw a thread somewhere about why OPML was used as the export format for feed readers. Not sure I've ever written about this. At the time we were finishing up Radio UserLand, in either 2001 or 2002 (there were two big releases). I wanted a way to export the user's feed list so people could use the same list in another reader. This one decision is why there are so many feed readers imho. Because we offered no lock-in as a key feature, everyone else including Google had to do it with their products. Users had the expectation of data portability, so it's in the DNA of feed readers. Maybe we would have dominated the market with Radio if we didn't export the feed list, it was for a while the only feed reader, but more likely Google would have entered and clobbered us, users wouldn't have been able to use both products, or easily move on after Reader shut down. Anyway, why OPML? First, we had it around, it was new, I wanted people to use it, and I also wanted to use the outliner to edit my subscription lists. JSON didn't exist as an option at the time. And using RSS for that seemed confusing and not right. RSS is a syndication format, representing a flow of information, where OPML represents a list that evolves, but each item has permanence. It's static where RSS is dynamic. It's like the difference between a podcast and an album of music. You listen to a podcast once and it's gone, you may listen to an album many times over decades. Fundamentally different kinds of data, with different needs in how they evolve, and thus the format you use to represent it. Yes we could have shoehorned feed lists into RSS, but 20 years later I'm glad we didn't. #

I asked ChatGPT how I used OPML to share lists of feeds. The story it came back with is amazing. I asked for the list in OPML. I opened it in Drummer.#

Noted that WordPress blogs can now peer with Mastodon and other ActivityPub nodes. Suppose I have a WordPress site and my username is scripting. What address would you follow in Mastodon? #

One of the cool things about this blog is that I got to write about these things when they were new, and now that they've been around quite a long time, in tech industry terms, I get to write about them again. #

I have a few virtual servers at Digital Ocean, and two of them are running Caddy, so the sites hosted on these machines are HTTPS not HTTP. Caddy is the best. You install it, configure it and that's it. It takes care of all the michegas with the EFF, makes them completely invisible. Also makes it possible to switch thank goodness, I didn't miss that the EFF built themselves into the new system, not by force exactly, but by default. Heh. They get a lot of free marketing. Anyway, the fact that Caddy exists means that Amazon S3 could offer HTTPS access to everything I store there without me having to do anything but check a box in a dialog somewhere. So why don't they do it, make HTTPS zero cost to implement and maintain. They could even charge extra. It's kind of perfect, they know who I am better than anyone. We could make HTTPS disappear, which imho would be a good thing. #