Part of the issue is that XSLT in the browser is stuck at version 1.0 so lacks a lot of the improvements added in 2.0 and later that make working with it a lot nicer.
Part of the issue is that XSLT in the browser is stuck at version 1.0 so lacks a lot of the improvements added in 2.0 and later that make working with it a lot nicer.
XSLT 2.0 is a failed spec, there is only one implementation by Michael Kay in the paid version of Saxon.
And he was also the spec editor, his incentive was to get lucrative contracts from BigTech, not make the world a better place.
XSLT 2.0 is not a failed spec. From [1] RaptorXML (XSLT 3.0) and xjslt (XSLT 2.0) are listed as implementing that spec. MarkLogic also provides XSLT 2.0 support.
Saxon has a free HE version [2] that has the source code available and implements XSLT 2.0 REC, 3.0 REC, and 4.0 ED at the baseline conformance. The paid version implements optional features and vendor-specific extensions [3].
Even though Michael Kay is the editor of the spec, several others are involved in the standardization of XSLT, XPath, and XQuery, including members from BaseX and eXist-db which provide XQuery implementations. And as XPath is a subset of XSLT and XQuery there's a lot of overlap there, and features come from many people, not just Michael Kay.
[1] https://en.wikipedia.org/wiki/XSLT#Processor_implementations
[2] https://www.saxonica.com/download/java.xml
[3] https://www.saxonica.com/products/products.xml