The short-lived browser

Time for another brief update on my "offline first" computing efforts, following on from an earlier post:

Gemlog post "Progress toward offline first" (2020-08-16)

I was surprised to find in my early experiments with leading an offline-first life that being offline wasn't actually in and of itself sufficient to avoid distraction and fragmented use of my time. Even without an active internet connection, whatever I was doing I would every now and then switch to the workspace with my browser open, and just cycle through the open tabs sitting there. The behaviour was entirely instinctive. This kind of habit can, of course, be broken over time, and staying offline could help with this: if the content sitting around in open browser tabs is not updated, it will eventually be entirely consumed and then flipping through it will be unrewarding. But it's also very possible to go online once a day and open up enough new stuff to keep one busy until the next day's connection. Perhaps the eternally open, multi-tabbed browser is actually a bigger problem than the permanent internet connection?

Now I am thinking instead of trying to cultivate a workflow where I open a non-tabbed browser only when I have an immediate need to do something web-based, complete that task and the close it. This obviously has the potential to be an annoying and inconvenient way to operate, perhaps even more so than being offline most of the time. It ideally requires a browser which starts up very quickly, as well as a supporting ecosystem of tools for launching said browser with a specific target URL already specified. In fact, maybe even something as extreme as a browser which accepts a single URL as a startup argument and then actually refuses to visit any others. This browser could then be launched from within, say, an RSS client, or a Gemini client, only to visit things I specifically want to visit at that moment. After reading it, the browser is closed.

I have started experimenting along these lines with two tools. One is `surfraw`, the whimsically-named "Shell Users Revolutionary Front's Rage Against the Web", a command-line tool written by Julian Assange in 2000 (i.e. six years before the launch of Wikileaks) which lets one run a command like:

> sr wikipedia cobalt

to launch a browser pointed at the Wikipedia article on cobalt, for example. There is support not only for Wikipedia, but for all the main search engines, Project Gutenberg, and a whole lot more. Support for websites is provided by so-called "elvi", and naturally you can write your own to add support for whatever sites you like.

The other tool is (the coincidentally similarly named) `surf`, a simple web browser by the Suckless project. Unsurprisingly for the Suckless crew it has a ruthlessly minimal interface which is purely keyboard-driven, which generally makes it uncomfortable for lounging around in, which is exactly the point here. It does not support tabs at all.

So, let's see how this combo works out...

Surfraw at Gitlab

Surf at suckless.org

-- CC-BY-4.0 2020-11-06 Solderpunk <solderpunk@posteo.net>