Hi, Once in a while I try to clean up my tabs. First thing I do is use “merge all windows” to put all tabs into one window.
This often causes a memory clog and firefox get stuck in this state for 10-20 minutes
I have recorded one such instance.
I have tried using the “discard all tabs” addon, unfortunately, it is also getting frozen by the memory clog.
Sometimes I will just reboot my PC as that is faster.
Unfortunately, killing firefox this way, does not save the new tab order, so when I start firefox again, it will have 20+ windows open, which I again, merge all pages and then it clogs again !
So far the only solution I have found is just wait the 20 minutes.
Once the “memory clog” is passed, it runs just fine.
I would like better control over tab discard. and maybe some way of limitting bloat. For instance, I would rather keep a lower number of undiscarded youtube that as they seem to be insanely bloated.
In other cases, for most website I would like to never discard the contents.
In my ideal world, I would like the tabs to get frozen and saved to disk permanently, rather than assuming discard tabs can be reloaded. As if the websites were going to exist forever and discarding a tab is like cleaning a cache.
Say these problems are fixed for now. How many tabs is enough? How do you see this tab hoarding progression being sustainable at all?
I would put the full text, image and video of every tab I have ever opened into the context memory of an open source LLM if I could. I would only consciously delete stuff that needs to stop existing immediately, like doxxing data or illegal data or malicious code.
This is like asking, how many email should you keep.
Well at work we auto delete all emails after 60 days.
But my personal email has every email going back to 2006, the last storage failure before backup, and it’s all quickly searchable.
The other limit would be storage space, but my cluster has still 180 terabyte empty space, I don’t see that getting filled up from plain browser data any time soon.
Of course, I would like better automated data catalogging tools. I would like to ask my local open source LLM to “pull up all tabs regarding 7 megahertz maser project” and it should should open a browser window that contains every tab I have ever come accross on that topic. Including now-dead websites. It should all be sorted by date, it should know to put the more basic tabs to the left and the cutting edge stuff on the right. All this without me tagging a single thing, without wasting a minute of my time doing sorting busywork.
It is the job of the computer to organize my data, in an offline, private, reliable, open source-based, enshittification-proof manner with infinite memory and perfect recall. So that I can get on with doing the stuff that I want to do and not fiddle with browser settings.
Mozilla foundation has revenues in the 500 million range and a 7 million a year CEO, I expect nothing less.
I applaud their initiative with llamafile, however I hope that was just an appetizer.
yeah mate - you need a knowledge management software, not a browser.
tabs were always ephemeral and that’s unlikely to change because they’re much more than text and images.
that’s simply an unreasonable expectation for a browser.
I’m honestly surprised Firefox even handles more than a few hundred open tabs.
That’s fair, maybe you’re using the wrong tool though, something like an internet archive sounds more like what you need.
Take every tab you open and save a PDF, all the text, and all the images, then put a timestamp on them before deleting the tab. That’s not the point of a browser though, that’s an entirely different product.
You’re welcome to build it though, or ask Microsoft if they can make Recall work for tabs.
I was going to also say that OP might be wanting something like Recall (which might be one of the few instances where it constantly saving shit would be perfect). But they would need like the most extreme version that isn’t just saving searchable screenshots.
I also think that one major issue for OP is more about how the actual sites are coded these days. As even if a single tab is being used, the shit can just decide to force it to update the contents at any time (like how just having Gmail open you will see new messages just show up even without refreshing your browser).
It seems like the perfect situation for OP would be if the web still worked like it did pre-web 2.0, but with using the current version of FF. Outside of that, it really seems like they need to just start having sites be auto-completely downloaded for full offline use.
I am still shocked that the main issues being had seems to be that it taking 10s of mins to allow FF to process that much stuff is the frustration. Which does seem to mean FF is holding up pretty well given the situation. Their complaint about tab isolation being too much overhead seems odd though. As it would seem that going back to not having that would mean a much higher chance of just everything just being yeet-ed out of nowhere.
I am not sure how their headspace of using virtual machines approach would be much better as shit would still have the issues of sites still self-updating and loading up in the first place. Though given they seem to have dramatically more coding experience, I am much more ignorant of this shit.