Why I have a perpetual browser history

April 30, 2015

I've mentioned in passing that I keep my browser's history database basically forever, and I've also kind of mentioned that it drives me up the wall when web sites make visited links and unvisited links look the same. These two things are closely related.

Put simply, the visited versus unvisited distinction between links is a visible, visual representation of your current state of dealing with a (good) site. A visited link tells you 'yep, I've been there, no need to visit again'; an unvisited link tells you that you might want to go follow it. This representation of state is very important because otherwise we must fall back on our fallible, limited, and easily fooled human memories to try to keep track of what we've read and haven't read. This fallback is both error-prone and a cognitive load; mental effort you're spending to keep track of what you've read is mental effort you can't use on reading.

Of course this doesn't work on all sites (and doesn't work all the time even on 'good' sites). I'm sure you can come up with any number of sites and any number of ways that this breaks down, and so the visited versus unvisited state of a page is not important or useful information. But it works well enough on enough sites to be extremely useful in practice, at least for me.

And this is why I want my browser history to last forever. My browser history is the collected state representation of what I have and haven't read. It tracks things not just now, in my currently active browsing session as I work through something, but also back through time, because I don't necessarily forget things I've read long ago (but at the same time I don't necessarily remember them well enough to be absolutely confident that I've already read them). For that matter, I don't always get through big or deep sites in one go, so again the visited link history is a history of how far I've gotten in archives or reference articles or the like.

There is nothing else on the web that can give me this state recall, nothing else that serves to keep track of 'how far have I gotten' and 'have I already seen this'. The web without it is a much more spastic and hyperactive place. It's a relatively more hyperactive place if I only have a short-term state recall; I really do want mine to last basically forever.

(In fact for me anything without a read versus unread state indicator is an irritatingly spastic and hyperactive place. All sorts of things are vastly improved by having it, and lack of it causes me annoyance (and that example is on the web).)

Comments on this page:

Along a similar vein, I keep my bookmarks forever, and Chrome has serious performance issues because of this. Keystrokes have a 4-second delay (presumably while it searches my bookmarks for matching sites). :-/


By cks at 2015-04-30 17:09:01:

Oh yes, that's an excellent point. Long ago I discovered that allowing Firefox to search my history when I entered things in the URL bar caused what could politely be called 'catastrophic performance problems', so I disabled that. Life might be better there these days, but for various reasons I basically never enter URLs that way (I have other, generally faster mechanisms).

It's possible that this means I'm missing out on something convenient. It's also possible that Firefox has optimized this since those days, although I suspect not; massive history databases have always been an outlier case.

(Firefox does optimize some aspects of this these days; for example, history is looked up asynchronously. Usually the lookups finish fast enough that you don't notice, but if I start Firefox for the first time on a sufficiently loaded system my initial page will first show up with unvisited links and then they'll progressively go visited over the next tens of seconds or so. Since I keep my Firefox session running almost all the time, this rarely affects me.)

By Polygone at 2015-05-02 00:52:47:

At some point Firefox chose to limit the number of pages retained in the history, based on the system configuration iirc.



Which probably fixed the performances issue, but also disregarded the user configuration at the time.

By cks at 2015-05-02 01:14:09:

I have my values for that preference (and the non-transient version) set to very large numbers. I forget how I found them; I may have just been poking around at some point, stumbled over them, and immediately yanked them up to safely high values to approximate 'infinity'.

Written on 30 April 2015.
« The 'EHLO ylmf-pc' plague of SMTP authentication guessers
I'm considering ways to mass-add URLs to Firefox's history database »

Page tools: View Source, View Normal, Add Comment.
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Thu Apr 30 00:14:42 2015
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.