i do and make and write and make and do things with and for and as pastagang. it has taught me quite a lot, and one of these things is the slippy mindset: if youre always ready to lose all of your code and not be fucked, you are very slippy. this mindset is extremely good in most ways i can mention. i am a very the fog is coming kinda person, so i used to constantly and constantly and constantly worry and back it all up and oh god what if the storage medium dies. no. i would prefer to keep my code for the most part since it is a bit inconvenient to make new, but i would probably not be fucked nearly as much as i thought. the bits that were too depended-upon are slowly falling into disuse as i try to be more slippy. tuddylib is dead. my web server is python -m http.server and i could write a new http server of my own in minutes if python's one becomes nonexistent. it might bother me a little, but thats it. and the slippy mindset isnt even necessarily about code that isnt mine dying, so in fact almost everything would just stay how it is. revpfw3 is almost the only thing where i am a bit fucked if it disappears, but that would actually be good because it really needs a rewrite. v3 is pretty good but its got issues that v4 wont have.

i try to avoid anything that is too heavy nowadays. no nodejs. no big http servers. as few things that arent a single binary as possible. this website runs on two shell scripts and the http server integrated into python3.

nearly nothing i write is important enough to mess me up if it dies, and most things i write are small enough to be renewable in a day or two. this is good. i must not depend on what is rare and hard to create. the biggest issues at the moment are fedi and forgejo, both of which i depend on relatively bigly but are hard to recreate in a way that remains compatible with the old (unless i settle for much less).

looking at the bigger picture i also depend on the linux kernel but that isnt particularly relevant to slippiness since its not my code and im sure some copy of archlinux somewhere will survive even pretty significant catastrophes and i can clone it from there. this is in fact the case with most of my own code too but i mostly dont need this assurance because i need very little of my code, if any.

in general this mindset is extremely good because it allows me to be reckless and i like being somewhat reckless. it kind of reminds me of those silicon valley startups that say they move fast and break things, except in my case its not marketing and its much more irrelevant if i break things.

i host a git forge. the ssh for it is sometimes a bit flaky for a few minutes. i could fix it but i dont really care. i can wait a minute, and it teaches me to not take it for granted all the time. i also apparently host the pastagang page now (at least for a time) and while i will probably try to not give it unneeded downtime i dont really put in any effort into that. if it dies it dies, and thats good actually. the less you try to resist the less it hurts.