Hacker News new | ask | show | jobs
by julianeon 952 days ago
Curiously I fall in between the two groups he describes.

I usually have about 3-4 browser windows open, each in its own workspace (I'm on i3) - basically its own screen.

In each one I have about 5-15 tabs open. That's about the limit of what I can read at a glance to quickly flip between them.

Periodically I close a few. That's how 50% of the tabs get closed.

The other 50% go when I declare 'tab bankruptcy.' If a window has been at the tab limit for a couple days and wasn't compelling enough for me to act on before, I'll close all its tabs. If I didn't take notes about it or process it in some way, too late, gone now.

I actually do take screenshots of some of my tabs, full page screenshots, which I think in practice is my 'bookmarking' system. I title the file with what it's about - say, "shadcdn.png". Now it's 'saved', in theory anyway for me to look at at some undefined future date (no hurry).

That's my system, which works well enough.

2 comments

This matches my usage pattern and I use the term "tab bankruptcy." Twinsies!
Great taste my friend! Once you get used to it, it's hard to imagine anything better.
Wouldn't PDFs be a better way to save them? Full page screenshots glitch out me about half the time, esp sites that use js to muck with scrolling
This is pretty tightly tied to my particular setup on Ubuntu, using scrot to screenshot and feh to view images and the terminal to rename them, but here's the issues I'd have w PDFs...

The biggest problem is that sounds more verbose. I'd be saving the page as a PDF, but then I'd have to hunt more to find out why I saved it, since save as PDF often extends beyond one page iirc. Really one page of screenshot is plenty for my needs.

Also PDF's are a lot "heavier" than images. I can scroll through 50 images in a minute, using feh. I don't know how I'd go through multiple PDF's actually, but in any case I imagine in the best case scenario it's still slower. The effort of loading 10 PDFs would deter me from doing it as often, I think.

Finally, using scrot to get the screenshot, it's as simple as one command. Using ctrl+d, it doesn't even require a terminal, I just enter it into the 'command bar' (basically like an always-open browser address bar but for commands on i3). I also do scrot -s to save images sometimes if I don't want the whole page, but that takes slightly longer. With scrot (no arg) I get the whole page instantly, with scrot -s I have to draw a box around the part I want to save.

About 90% of the time scrot (for screenshot) - then open terminal, rename file - is good enough and takes on the order of 15 seconds. It's already very lean, and I don't think I can abstract away the naming part, which is where most of the time goes anyway.

If I wanted to save time I could just not rename the screenshot and keep the name as basically a timestamp, but then I don't know the image's contents. I did actually try using OCR, which kinda worked, but not as well as renaming, which succinctly gets my point across better than OCR's verbosity. Bottom line, if I want to remember a browser screenshot, it's worth 10 seconds of my time to explain why through a descriptive filename. I save <10 images daily so I can spare that.

I should say I've had no issues using scrot: works 100% of the time so far, faithfully getting what's on my screen. If I did have glitch issues I might search for an alternative - but I haven't had any so far.

PDFs are ok but I prefer the SingleFile extension which has the benefit of not introducing page breaks which often split images in two. It also preserves the original source and assets in case one ever needs to extract them.