The author makes some good points, but I’m in the group of comments they refer to as “hates XML” and I am forced to conclude that they really desperately underestimate how MUCH I hate XML, and the suffering it inflicts on me when I am forced to deal with the XML horrors they so affectionately craft, so I find it hard to not be excited that yet another ugly XML-based technology is dying. With shitty motivations and as part of a malicious strategy perhaps, but I am still not mad that it’s dying. It’s sad to understand that people still use it and rely on it and even like it but I’ll never be one of them and while I have empathy for them, I just can’t generate much sympathy. I have my own technologies that I’m fond of that have been replaced with ones that I am personally not fond of, so I know how it feels, but I’ve accepted that I’m probably the one who’s wrong in those situations, and… XML just really sucks, and XSLT is built entirely for and with XML.

His response that you can use it for JSON too is maybe technically true, but I really don’t need to or want to use it with JSON either. His argument there is sort of like responding to someone complaining that a boat is unsafe for a transatlantic voyage by telling you you could just put wheels on it and use it as a car instead and then you won’t have to worry about whether it’s safe to take it out to sea. Like, technically, yes that’s true, but I don’t want to do that either, that’s missing the whole point of the dislike. XSLT sucks because it is attempting to solve a problem for XML that wouldn’t be a problem if XML wasn’t an ugly, overcomplicated, verbose nightmare. Once you’re already using JSON, a versatile, flexible, universally available and equally turing complete tool that does everything XSLT does for JSON is… actually I’ll give you a hint, you can just look at the first two letters. And many, many other options are available for JSON too, XSLT simply doesn’t need to be one of them, no matter how much some people apparently defend and love it.