Sunday October 2, 2022; 11:55 AM EDT
I've been writing glue code for ages. Basically, someone provides an API for a service. Could be the IBM PC with calls to the ROM BIOS to read files, or the Apple II hardware screen memory (invented by the clever genius Steve Wozniak), or the Apple Events supported by various Mac apps in the 90s, or the XML-RPC interfaces for web apps, or the REST interfaces of GitHub and Slack or thousands of other services. They give you an API, and then after you see what it can do, and if you want to use it in your app, you write some glue, so all your calls flow through code you control, so if you learn something about their API, you only have to support it in one place. It's especially important for the few services that reserve the right to break their developers. Apple used to do this all the time, I haven't programmed to an Apple API in a very long time. Maybe they stopped? It's a way of forcing your developers to redeclare fealty to the Great Platform every so often. I like the web where no one has the power to break me (although Google might argue with that). #- Anyway -- today I have a new release of my glue code for GitHub, the code I use to read and write files from and to GitHub. I use this to keep archives of stuff, because I think GitHub has a good chance of sticking around for a few decades, longer than the S3 storage I use for everything. That'll disappear the day someone stops paying the bills. But the GitHub stuff should stay, no bills to pay (though I do pay for a developer account, only seems fair because I use it a lot). #
- I use GitHub for archiving my blog. Starting on May 16, 2017. And since then every post, every day and every month has been stored there in JSON and OPML. I've also managed to find and store OPML files for most of the stuff on my blog going back to 1994. #
- It's not much of a secret that I'm working on a product that reads feeds in RSS, Atom and RDF formats. It contains a feed reader, but it is more of a development tool for news products. It stores info about feeds, items, subscriptions, users and likes and other stuff. I want that stuff to flow to GitHub right from the start. It'll be a slice of what we're doing online in 2022 and beyond. Maybe that'll be useful to some researcher in the future? #
- The new thing in my GitHub glue is that it manages a queue of writes, so you can just tell it to write something and then go on to the next thing. GitHub has a rule that you can only have one call extant at a time. This is a terrible for for an environment like JavaScript that doesn't let you do that easily. I got tired of managing that stuff in my application code so I added the queue to davegithub and that simplifies all future applications that have to write to GitHub. Absolutely nothing revolutionary in that -- it's just nice to be able to forget about that particular feature of GitHub, let the software manage it for me. And it's MIT licensed stuff so you can use it too. #
- I also released an example app that uploads a folder full of JSON files to a location on GitHub. #
- So If you are a programmer and want to use GitHub to archive your stuff, or to create a connection to GitHub for other apps, I've made it simple for you. Enjoy! 😄#
- PS: I think Stewart Brand would find this interesting, btw. #