slow per say, one has to take into account pulling the feed into the cache layer and having payload rendered in a browser. the round trip of data from out in the wild back to the server involves network calls (cpu-bound), cache-stuffing (cpu+memory-bound), and disk read/writes. by the time it's done it can be a costly operation compared to a db-backed data store.
to give things a perspective, try measuring the time of running curl on top1m even on a fast link it can be up to 1s
you're reading from cache, so it's quicker. memory will always have significantly faster iops vs disk-bound read operations. also recommend giving the codebase a look. there's always room for contributors. i'm planning to take a crack at a few issues.
i think there's an rfc for adding
pagination to the hardest problems in computer science. if there isn't, there should be :-)
pagination would be a good way to work around this but without js being involved i'm not sure how it would be done.
displaying a flat text file is in principle not problematic at all. the crux of the situation is the scale-factor.
accurate. the backend has to be able to catch the event from the browser which there's a disconnect unless you have some sort of client-side hook to trigger the backend to advance the next pagination call.
there's an ancillary note to take into consideration. post-arpa days
(1987..1994) if one wanted, one could reach the end of the internet. in those days backhauls were t1 and the data contained within the internet could be held on a hard drive.twtxt/yarn in general allows me to reach the end. at least for the day. there's no infinite feeds to overwhelm the visitor which in my view is quite nice compared to the endless pit that is the current state of the web.
we are very very small. but we are profoundly capable of very very big things. --hawking
node01.dc01.happycrunchybits.lol seems legit?
https://github.com/foxlet/macOS-Simple-KVM here's a repo i found that might be useful
nsa still knows it's me searching for old thinkpads and temperature tolerances for thermal paste :-)public search instances are likely a good call. i still use duckduckgo myself as i have work and research to be done, and it's a better alternative than google in my personal view.
no apologies needed, we're all friendly here
i hope this makes some sense. at first i didn't understand it until i read the techspec and tore apart the client code then it was a
this is brilliant moment.there's quite a bit of philosophical thought that goes into how things work here. this is not twitter, and that's by design.
twtxt and client yarnd would have to be redesigned from the ground up.now that being said, let's say there's a post that turns into a yarn and people respond to it frequently it may be more prevalent and show up in your feed if you are indeed engaged in said yarn.
@nexeq there's a couple of ways to look at this. first your feed is what you make it. based on whom you follow. there's also this concept which i don't think has been officially named, but the idea is the a timeline such as in a timeseries moves along as life does. thus what you see today may not be here next week
dc01.<noidea>.tld
@prologic that's the spirit!
cloudflare is a mitm!!! well ffs, what do you expect? they are proxying global traffic for you. this is how things work. if you want to use their service you have to accept some basic perceived risksthe way i see it. if they housed lulzsec for 10+ months without a takedown my concern for my pod and infrastructure is minimal should i choose to use them.
how do you like inwx as a domain registrar?
i really miss freenode
love || hate situations i'm afraid. when implemented properly and the plants properly watered it can work just fine. but the documentation like many things are lacking, especially a year or two ago when many were trying to shoehorn it into docker. ended up solving this problem after a few days but synapse was resource intensive.dendrite works (or did), but the number of issues not being addressed is concerning and their api/system components are not in full parity with synapse (yet).
that's quite the broad generalisation. moxie seems to be speaking for the people. while the subset of operators who run their own systems are tiny compared to the global internet this is like saying
people don't want to drink coffee, they are addicted to the caffeine i'm long past even recognising caffeine in my system but i love a good cup of coffee as a focal point
.dotfiles today :-)
circa 2011-2014 were very nice. got my hands on one and loaded linux on it. lasted forever until i dropped it. while i'm not a personal fan of the ecosystem, i respect it as they do make things quite easy and mostly does just work. if i were doing business i'd give them a go
.dotfiles would be a fun adventure :-)
s/stupid/clever/ with the mesh network for my infrastructure. goal is to global load-balance and give say an .au visitor a traffic path close to them while someone in the .us or .nl would have their own. all that noted, any leads on some decent vps providers in .au, .uk, .jp, an .is ?
- ensuring you persist your metrics data as it's nice to be able to look back for patterns (docker volume to tier2 storage like rusty-spinners is ace here)
- get to know grafana. use some of the dashboards that are available, edit them to learn how the queries and designs work.
- if you hook up alert-manager, give gotify a try.
if you have questions maybe we can have a chat on our weekend call
i've nothing up at the moment as i'm still trying to wrangle my machines and auto-deploy swarm. i've been so polluted by kubernetes for years the simplicity of swarm is forcing me to unlearn adhd-plumbing
twitter3e4tixl4xyajtrzo62zg5vztmjuricljdp2c5kshju4avyoid.onion
doesn't surprise me alec worked on this, he did the same at facebook. great work even if i won't use the platform.
@caesar valid point, and likely right. the yarn ecosystem is microscopic compared to the abyss that is the internet, but that will not stop us from having strong communities and making progress. i'm in this for the long-game
yet. also, i do understand the abundance of customisations available for the android ecosystem. even with an alternate os such as graphene or calyx you're still riding base android. took me a good week to get the device how i like it. now it sits in a drawer until i go out and need to yarn or signal
here's sort of the workflow it serves for me:
inbound email ---> mx for tastycornpuffs.com ---> messaged wrapped in pgp ---> target mx (personal) ---> my inbox
decentralised and disconnected ---> move towards centralisation ---> full centralisation and tight-coupling of systems ---> netizens asking why? ---> move to decentralisation ---> decentralised and distributed (we likely are here now) ---> the future?there's an idea i scrapped together (likely i'm not the first to coin this).
when the system fails you, you build your own system
nil, and every human had the freedom to explore network(s).it's interesting to watch what's occurred over the decades. the internet fundamentally has not changed. only the people and technologies riding on top of it have.
rulership and control over me. i'd rather be engaged with an ecosystem that allows a netizen to stake their claim of the internet, link up with their friends vs the concept of friends and followers that have been psychologically optimised for most people driven by dopamine and fomo
in the end, none of this truly matters. but companies like twitter want to induce a synthetic worry for not being a part of their ecosystem.
@prologic i signed up to namelock my handle (suspension works too), uploaded an avatar, followed some computer scientists, liked a couple of posts, let the account sit for a while, then boom.
@screem once i stop procrastinating and write it would be happy to set you up proper. google voice is still about the only non-sim-bound way of validating most accounts. apparently the bird site still will reject sip-carrier e.164 numbers as seen in my deadlocking test
65516 68162 55401 70333 20849 52655 66910 24952 43174 80113 82214 86079
friends and such.@screem 100% i consider this and also the email thought @prologic mentioned as a
watered-down firewall of sorts. remind me to share this during our next call. or i should write about it?
thingsi may even run osx or windows at that point.
i currently run grapheneos and the user experience is fairly well put together, it's fully google-stripped and you can use many apps you know and love. but i'm not a good pitchman. i've rocked a flip-phone for a decade.
@movq i'm going to give bundlewrap a shot after my day ends. i'm always interested in new config and orchestration tools. on marketing, i don't think anyone is good at it. i sure am not. i couldn't sell free money. :-)
i'm looking to build this in golang since it's wicked fast and as always keep it open so that people can fork and stand up their own system should they choose to.
what are your thoughts?
want dopamine? have a chocolate. i am. cheers!
twtxt lacking such mechanisms forces me to be thoughtful in my posts and replies. similar to if i was writing a letter to a friend far away.
slept on this and i think it's a brilliant idea. here's why.
in a way, sir you hypothesised :)
counter-hypothesis: as humans perhaps we experience an etching period where a specific time in our lives or an era has a deep and profound imprint on our psyche. as the world changes, we adapt but there may be a point where on marks that moment of wonder and never let it go.
i wax poetically about the past and hold onto things i care for. but instead of being frozen in time i prefer to meld older constructs with new. thus satisfying that nostalgic need while deriving new memories that bring the ages together.
i'm off for a coffee and walk, sleep well gents.
too early for a proper screenshot. this was my previous swarm setup in 2020 which i'm reviving. swarm master and workers communicate over a dedicated mesh and traffic routes to an external vps traffic gateway. the reverse proxy that handles things is traefik which i've rather customised for my own pursuits.
this method allows me to host swarm anywhere i want be it at home, work, at the coffee shop off a cellular modem
@adi these were three t480s i rescued from a government auction for near theft prices. no memory, disks, chargers, or batteries but that didn't stop me from giving them a good home :-)
- teardown old thinkpads for service
- optimize storage on swarm cluster
- test deploy yarnd pod on a temporary domain
- go for a long ride
really any mail system (even yahoo and gmail) support pgp. i seem to lean towards encrypting emails and signing from the terminal then simply use a mail client or web browser to send the message. but i get what you're saying, nice to have the automatic bits. helps with people adopting better posture
signal's apps are open-source as well as their server infrastructure. the only thing that's not open is their interface to the intel sgx enclaves (blackboxes). i can accept this risk.
minimal metadata is exposed account id(s), last connection date, account creation date in unix timestamps much less than the facebook
- psyche debt > fiscal debt > tech debt
- learning to let things go is analogous to swedish death cleaning of the self
- don't build a skyscraper to house a post-it note
- procrastination is likely a sign you're manifesting complexity
- never underestimate the potato, it's a mighty fine nightshade
++ on dumping their ca(s)
the internet, in my view has always been nation-agnostic. we are one world, one nation. and like every nation you have good s/citizens/netizens/ and bad s/citizens/netizens/
we don't close roads because criminals drive cars, we just shoot their tires out
although i am looking to upgrade to @ullarah suggestion.
mailbox.org is a good provider, been around a long time. there's another one called mailfence which is also very good and have been around a good while. protonmail is great, but i pull down all my mail and protonmail makes this complicated with a bridge (paid feature).
germany handing over data to foreign governments likely will vary by the relationship and policy between the requesting party and germany itself. due process is still a thing there.
- the pursuit of perfection is a fool's game
- related to the aforementioned, finding the perfect domain name is an abyss-like vortex
- eventually the construct of perfection comes organically
- trying to deadname/squat your id accounts on all platforms says
i care too much about centralised rubbish- note to author, you use too many bulletpoints
that's really not bad considering your region. when i was in the eu it was rather expensive. now i have more bandwidth than i know what to do with so i create traffic crawlers that crawl the top 1m websites to keep the cache hot
i'm also rather inspired by your small website. meanwhile i'm trying to bribe @prologic into learning css for me.