- origination date of the website
- number of theme changes
- whether ads are present
- estimated data usage / load time
- general website size
- type of website - personal, company, etc.
- last update time
- presence of javascript
- browser compatibility (well, feature usage detection & browser compatibility inference from that).
Mostly this is just me missing the websites of the early 2000's, and trying to figure out a way to rediscover them.
And I'd probably want content on top of this. (Edit: e.g. search by topics)
Lastly, it'd be nice to restrict things to sub-genres, but I'm not sure. E.g. when I'm doing a search I'd love to reference things related to micro-controllers, and so maybe I'd put in Arduino to get into the realm. Sort of what like google does for you without telling you. (tailoring your searches by some magic context).
A man can dream...
Edit 2: Search engines these days seem to be answer engines, I want a research engine.