djsundog

djsundog, 8 days ago

we just got back from a checkup my wife needed (all good!) and driving in the rain all the way there and back, so it's time to chill for a few before I go do some more fallen tree removal.

in between the storms and the fallen trees and the general ennui of...waves hands...y'know...anyway, between all of that, we've been working on a cool little project that's almost ready to get some actual use in the next few days, so I'm gonna ramble a bit about that and the thinking behind some of the technical choices I've made along the way.

djsundog, 8 days ago

the project in question, currently under the working name All Whomst Wander, is a geolocation-based monster-finding-and-collecting game similar to things that Niantic Labs have built like that one game full of pokemans that lots of people play.

djsundog, 8 days ago

we're building a similar type game, but one that will encourage/assist/entertain folk while they explore New Ellijay and surroundings. the specifics of the game aren't super important at this stage - and in fact aren't even super well specified yet ha - but I want to chat a bit about the engine that makes (will make) the game happen.

djsundog, 8 days ago

in order to build the game, we need an engine for the game, and that's the bit that I'm currently working on (and getting ready to start internally testing over the next few days).

as @ajroach42 and I (and others in our crew of comfy co-conspirators) talked through the idea, we decided to focus on living our values and our ideals as we build the thing, because no corp ever will and folk deserve some thoughtful intent.

djsundog, 8 days ago

so, it's going to be a web-based frontend, not a native app tied to a specific mobile platform.

it's going to use as little library code as we need to get the thing done, and no huge frameworks or dev tooling.

the backend is in PHP.

the backend database is sqlite3.

so far, the backend uses no external library code. no composer. no ORM.

djsundog, 8 days ago

the data in the database is as sparse as it can be and still provide the functionality we want to offer.

when you log in, the data comes from the Players table (and the PlayerAuth table), not Users or Customers or Clients, but Players. we're making a game for people to play; that's the model.

intent.

djsundog, 8 days ago

during registration, we ask you for a nickname, an email address in case you find yourself needing to reset your forgotten password, and a password.

not your real name. not your first and last name. not your address or country or phone number.

we don't need any of that for the game, so why collect it?

djsundog, 8 days ago

I've been questioning every best practice my fingers move toward typing during the initial dev process and stopping myself: "is this a best practice for the right reasons?" and then deciding on an implementation based on my pondering.

so yes, using solid password hashing with a salt and enough complexity to make it difficult to brute force user passwords.

but still no ORM, no enterprise-class database management server. just sqlite and php.

djsundog, 8 days ago

"sundog, it'll never scale!"

friends, we do not care. we don't want to scale it, we want to play.

we have a small town. if there are 10,000 people here at once the place has been overrun and is elbow to elbow. if all the elbow to elbow folk all want to play the game at the exact same time, php and sqlite will still handle that just fine (ok, and nginx).

djsundog, 8 days ago

we're going to release the engine as open source, in hopes that other folk take it and build games with it for other places.

why wouldn't we? we're not competing with other places. someone in florence, alabama is not going to suddenly run to New Ellijay just to play this game. we'd love someone in Florence to localize it there.

heck, we might even be able to set up a webring of folk who've set up games on this bare bones web-based game engine we're building. play 'em all, have an epic road trip!

djsundog, 8 days ago

we're not going to charge people to play and we're not going to run any scammy contests in the game and we're not going to try and get people addicted to playing the game.

we're going to use it as an artistic expression and as a way to introduce folk to the bits and pieces of New Ellijay that we want to highlight.

and if people are having fun and playing the game, they might spend some money at some of the places they go, and that'd be great! supporting the businesses in the area is crucial if we're going to keep folks housed and fed and working.

that's the big business model. knock-on purchases. not ads, not selling player data, not a pivot and an exit to private equity.

djsundog, 8 days ago

so why am I blathering on about this?

because I'm proud of it, already, even before it's finished.

and I should be! we should be proud! it's good!

but beyond that, there's no magic, and that's the part I really want to drive home.

we're not building anything super fancy. we're using as much off-the-shelf stuff as we need and as little as we can, and we're assembling it, and then we're gonna load our database with fun cryptids in neat places and make it so folk can go meet them and befriend them and feed them and stuff.

and you can do this type of thing too.

djsundog, 10 days ago

@ajroach42

so, hot take, but I think the answer is "other human curators" and I think it can work even though it sounds like an ouroboros type impossible system.

the trick is that the curation focus of one curator needs to source from curators with slightly different yet often (for various definitions of often) overlapping foci, and at least some curators need to be sourcing from direct sources - the people making the thing that needs discovering.

djsundog, 10 days ago

@ajroach42

and I think we have to duplicate work a lot to avoid over-amplification of a subset of voices creating a quieting effect on the majority of voices.

if I curate music recommendations, ideally our curation system should not allow my voice, or any voice, to dominate that space. some mechanism to turn off the attention tap once a curator hits some audience size kind of thing.

djsundog, 10 days ago

@whitneymcn @ajroach42 webrings are absolutely a reasonable model for human-scale discovery and definitely a form of curation-based discovery - someone somewhere is deciding which sites get added to any given ring, based on whatever basis they choose.

djsundog, 10 days ago

@whitneymcn @ajroach42 add a "directory of webrings" layer (of which there should be multiple directories obviously) and promote strong local per-site search capabilities and you're well on your way to a decentralized yet discoverable web imho.

djsundog, 10 days ago

@whitneymcn @ajroach42 directories are just curation too! but I feel like directories should probably have a lighter touch than a webring? like, opt to add more pertinent listings rather than find more ways to limit the scope. but maybe that way lies madness and web crawling again, and I'm not convinced we really need much of that.

djsundog, 11 days ago

@StillIRise1963 ok, but consider: we could implant Elon into the brain worm hole of Kennedy and be rid of them both