We use the HTML Extractor a fair bit when prototyping processes internally, but it'd be very useful to be able to use XPath in addition to CSS Selectors when extracting values from HTML.
Main benefit is that XPath allows for much more granular queries for selecting tags/text and can do plenty of things that CSS selectors cannot do.
I have fiddled a bit with using the various XML transformers to extract XPath, but with standard HTML websites rarely being valid XML, it gets trickier to troubleshoot extracting values from 'bad' sources
