I completely concur. Some of the weird stuff I have stumbled upon in my internet explorations have defied description; many of those places have evaporated into the ether since then. I think when that happens, humanity has lost a bit of itself.
A useful feature of recording everything would especially be for TOS/EULA getting stuff that's hidden in scroll areas. I'd love to be able to revisit things.
It should be feasible. I've got multiple bookmarklets that do this for downloading blogs and webcomics for later offline reading (some are so heavily scripted it's impossible to just fetch them).
Add something to the above so that it inlines referenced files in base64, and you're basically done (I tend not to bother as I'm only after the text content). For actually interactive files, it gets more complicated; you'll need to save the current JavaScript state as well as the page resources.
