This session started as a simple curiosity: how hard can it be to count a few XML nodes? It turns out the answer depends a lot on the language, the parser, and the file size. In my pet project xml-i ("xml-eye"), I implemented the same basic XML-processing task - counting different nodes - across a range of languages including Rust, C++, Java, .NET, and PowerShell, and ran it on files from small test cases all the way up to multi-GB inputs.

This talk is a practical comparison of languages, parser models, and real-world tradeoffs - less about “which language is best” and more about how easily performance assumptions can fall apart once you start measuring.