Former colleague Nathan Yerlger has a series of posts on technical debt (1, 2, 3). I’m responsible for some of the debt described in the third posting:

We had the “questions” used for selecting a license modeled as an XSLT transformation (why? don’t remember; wish I knew what we were thinking when we did that)

In 2004, I was thinking:

The idea is to encapsulate the “choose license” process (see in a file or a few files that can be reused in different environments (e.g., standalone apps) without having those apps reproduce the core language surrounding the process or the rules for translating user answers into a license choice and associated metadata.

Making the “questions” available as XML (questions.xml) and “rules” as XSL (chooselicense.xsl) attempts to maximize accessibility and minimize reimplementation of logic across multiple implementations.

I also thought about XSLT as an interesting mechanism for distributing untrusted code. Probably too complex, or just too unconventional and ill-supported, and driven by bad requirements. I’ll probably say more about the last in a future refutation post.

Anyway, I’m sorry for that bit. I recommend Nathan’s well written series.