# history: i don't keep shell and browser history

tracking every minute detail of my daily failings is pointless. i am talking about collecting the history of every command i execute in bash, every url i visit in a browser, or every gps coordinate i ever visited. this is a lot of data. most of it is just noise. whenever i wanted to look up some data in history, it actually took very long time browsing through all the data and sometimes i did not find what i was seeking. i found it is actually more efficient to reconstruct the command or url from scratch. it also helps me exercise my memory so it dwindles away slower. i often relied on bash history to find the right parameterization for a command i remembered using. without history i cannot do this. but i am better off this way. i am no longer relying on an external, private, somewhat volatile database to interact with the computers. it forces me to familiarize with the documentation. if that is lacking, it motivates me to improve said documentation. better documentation makes the environment nicer for others too. since i have to type out the arguments myself, i get more acquainted with the commands themselves from which i often pick up clever, out of the box ideas. in the urls' case it forces to learn the url structure of the websites, and to come up with succinct urls for my own work. sharing urls is very easy and natural when urls are sane.

maintaining this history database is also quite the pain. i am not comfortable keeping my bash, browsing or gps history in some non-personal device or an external device not encrypted by me. when i use multiple computers, shells, browsers then maintaining this history becomes quite complicated. i value my sanity so i would need elaborate bash hacks to keep the history in sync across computers. perhaps i would consider even more elaborate hacks if i wanted to inject new history into an already running bash process. or when i use multiple browsers then they have separate histories and i usually do not remember what history is in which browser. i could probably set up some sort of elaborate syncing scheme between them too if i would care enough.

the history database might not contain everything. it will not contain entries from my incognito browsing. i might be remembering that i visited a website, then i would try to look it up in the history but i will not find there. i would spend a lot of time before i would realize i used incognito so the history will not have it. the bash process can also suddenly die and lose precious history in it which i would then feel bad about.

the history does not have good availability guarantees. i might accidentally truncate all my bash history. i would then feel very sad. or i might decide that i want to use a different computer and now the history is gone and i have hard time using it. or if i store my history on a network service, my workflow depends on that network being available. it is just another unnecessary dependency.

keeping all the history slows down the application. bash now needs to read in all history before it shows its prompt. not only i have a slower startup, but now i also have less memory available for other work too. bash and the browsers need to keep some history in the memory too to keep lookups fast. and in bash's case it probably does not even share history between the processes so i would be paying this cost multiple times.

the history might have chance of leaking. suppose somebody sits down to my computer and starts browsing. now they could easily see all my naughty sites i visit or the naughty data i have. if it is on a networked service, then there is a chance that there will be a global leak at some point and then everybody will have access to everybody else's history. that might contain some things i would not be comfortable with others seeing.

and do not even get me started on the idea of tracking gps location history. that is wrong on so many levels that i will not even bother trying to explain it. i do not use gps nor smartphones so i do not have this problem. but i suspect smartphones have this feature enabled by default. it is just something to be aware of. i understand that my carrier probably tracks my brick phone's whereabouts but at least i do not have to worry that 7 independent services also tracking it via their preinstalled spyware.

so i stopped caring about my tracking my history. i find that this leads to much healthier practices, better memory, less work and less worrying about losing or leaking said history. therefore i disabled history where i could: bash and the browsers. note that i still keep per session history in bash because otherwise correcting typos in commands would be super annoying. but i never write out the history to disk. i actually use a raspberry pi as my primary hobby computer so i get less wear on my sd disks this way which is a nice plus. besides, forgetting is a feature. if i forget something, then 99% of the cases it is a good riddance. no need to carry on the past forever.

published on 2017-12-06

to the frontpage