XULRunner was a great idea. It allowed us to create good-looking desktop applications with a native look&feel with an option to embed HTML and SVG content with only a simple server application.
At that time, it looked like a good idea to let the hard work of making a platform-independent UI browser on Mozilla and focus on the applications. Unfortunately, it wasn't.
The question is really "why wasn't it?"... XUL, conceptually speaking, had a lot of potential, largely squandered by Mozilla-the-corporation. Somewhere between the implementation based on RDF (a literal plague over web tech) and confused commercial strategies, it died a slow death.
I call it "The Law of Conservation of Unusability". This states that whenever a technology gets to the point where it becomes usable, where you become comfortable with it and are happy with it, something will always happen that either makes it completely unusable or makes it significantly more difficult to work with. This "something" often has no reasonable justification and the reasons for it look quite artificial.
Before it was killed, I created two technologies based on it. Phobos for Pharo and Squeak (see the screenshots): https://github.com/pavel-krivanek/phobos-framework
The second one was Seaside inspired XULJet for JavaScript: https://en.wikipedia.org/wiki/XULJet
At that time, it looked like a good idea to let the hard work of making a platform-independent UI browser on Mozilla and focus on the applications. Unfortunately, it wasn't.