Comment by giantrobot

To this day I'm frustrated that developers (web devs mostly) tossed XML aside for JSON and the requisite JavaScript to replace relatively straight forward things like converting structured data to something a browser could display.

I bought into and still believe in the separation of data and its presentation. This a place where XML/XSLT was very awesome despite some poor ergonomics.

An RSS XML document could live at an endpoint and contain in-line comments, extra data in separate namespaces, and generally be really useful structured data for any user agent or tool to ingest. An RSS reader or web spider could process the data directly, an XSLT stylesheet could let a web browser display a nice HTML/CSS output, and any other tools could use the data as well. Even better any user agent ingesting the XML could use in-built tools to validate the document.

XSLT to convert an XML feed to pretty HTML is a great example of the utility. Browsers have fast built-in conversion engines and the resulting HTML produced has all the normal capabilities of HTML including CSS and JavaScript. To the uninitiated: the XML feed just links to an external XSL stylesheet, when a web browser fetches the XML it grabs the stylesheet and transforms the XML to an HTML (or XHTML) representation that's then fed back into the browser.

A feed reader will fetch the XML and process it directly as RSS data and ignore the stylesheet. Some other user agent could fetch the XML and ignore its linked stylesheet but provide its own to process the RSS data. Since the feed has a declared schema pretty much any stylesheet written to understand that schema will work. For instance you could turn an RSS feed into a PDF with XSLT.