This idea has been nagging at me for a while.
So, firefox tabs behave a lot like a virtual machine, I thought wouldn't it be useful sometimes if we had the same capabilities that your average virtual machine software has ?
The ability to pause, resume and save state as a snapshot
The ability to re-route it's network interface
The ability to sniff and dump all traffic coming in and out
Duplicate the tab VM, buffer network access for replayability, create memory mosaic
The ability save it's content and sift through it
In particular, I was on a facebook group, a very long group for my work union and the person who ran this group left the company and was now going to close the page entirely.
I wished to save this group page but it was impossible. Facebook actively prevents this by loading in and out data as your scroll. The whole page simply is not in memory all at once for you to search or save it.
In the end I just screenrecorded the page while I left it scroll overnight.
But I wish there had been a way to save all the agglomerated content of the page as a single whole page.
As websites become more and more complex web programs, this might be the only way to preserve them as dynamic entities.
So that would include saving the current machine state of the tab from the start and all network transfers in and out. Including make RNG functions repeatably giving the same numbers at the same time. Saving all traffic as a virtual proxy.
And then in he future it might be possible to still "browse" a Web2.0 website even after it is gone. As more than a screenshot or video.
One way that I can do this right now, is to spool up a complete Linux virtual machine in proxmox or virtual box, load firefox on it, browse the website through a caching proxy running inside the virtual machine. And then create a snapshot of that virtual machine.
It isn't very efficient to snapshot an entire 8 gigabyte ram OS to create a virtual machine of a website but at least it's possible. But I think in the future there should be something like this built in to firefox, it should embrace it's nature as a web virtual machine and I imagine most of the functionality is already present in the code, it just need to be broken out in the user interface.
What I am most excited about this is, tab sleeping that is less destructive and, most of all, a bookmark manager that saves more than link,title and favicon. I would like the bookmark manager to save, the entire thing, re-usable offline, even in year 2525 after the internet and humans is still long gone, bookmark manager should still be able to browse some Web2.0 website, at least in a limited cached capability !
... View more