miguel.de.icaza@gmail.com (Miguel de Icaza) (Miguel de Icaza)
What Maintaining SwiftTerm Taught Me About AI Contributions

If you maintain an open-source project right now, your pull request queue probably looks very different than it did two years ago.

I maintain SwiftTerm, an open-source terminal emulator library that is now seven years old. Today, I estimate that about 80% of the pull requests I receive involve AI in some capacity. Looking back, though, this didn't happen all at once—it came in distinct waves.

Disclosure: as a Mexican national that routinely writes tedious and typo-ridden English blog posts, I had Antigravity fix my writing and repeated tedious and repetitive points (as my wife often points out).

Wave One: Vibe-Coded Band-Aids and Fork Experiments

Before people even started opening AI-assisted pull requests on my repository, I would browse the GitHub network graph just to see what forks were popping up and what problems users were struggling with. Early on, what I found was a lot of "vibe coding"—folks patching around issues in ways that were fundamentally wrong, slapping surface-level band-aids over deeper terminal state bugs. I would fix or implement the feature, despite the fact that the user had not submitted the fix - for the sake of helping all my users.

In one fork I stumbled across, someone had clearly unleashed an autonomous agent to try to fix all of my open issues in rapid fire. Luckily, they never actually submitted those PRs—they were likely just experimenting—and it was only my nosy self browsing the network graph that spotted it. Still, looking through those commits confirmed the early reality: the fixes weren't worth merging. They solved surface symptoms without understanding the system.

The Rise of the Promptsplainer

Alongside those early waves came a new open-source archetype—one that many communities had to deal with: the Promptsplainer.

Some folks are really eager to particiapte in improving a project and LLMs have opened new doors, but they lack some basic etiquette.

A Promptsplainer is someone who asks an LLM a question, gets back a confident-sounding explanation or a sprawling patch, and suddenly believes they can override decades of hard-won domain expertise. They don't actually understand the underlying quirks, race conditions, or protocol edge cases of the codebase they are touching. When a veteran maintainer points out a flaw, the Promptsplainer doesn't go investigate the code; they just copy-paste the maintainer's review comments straight back into their chat session and paste whatever the bot spits out next.

You don't need a massive daily flood of Promptsplainers to feel the damage. Even the occasional encounter can fill a maintainer with rage - and with good reason.

Wave Two: "Send a Patch" Finally Became Real

Here is what has changed dramatically over the last year. We went from AI fixes being flat-out wrong or lazy band-aids to pretty much landing in the right ballpark.

To appreciate why this matters so much to me, you have to remember what maintaining an open-source library looked like just a few years ago. Back then, I would get a steady stream of bug reports—and if I had the spare time, the patience, the kids went to sleep early, and my Netflix queue was empty, I might actually sit down and fix the bug.

For decades, open-source communities had a semi-joking, semi-exasperated stock response whenever a user asked, "Why doesn't it...?". We told them, "Patches welcome," or "Send a PR." Most of the time, we knew they wouldn't, because diving into a complex codebase from scratch was too steep a climb. The perfect storm between unpaid volunteer work, uneducated takes and exasperation.

Like many others, I have witnessed the change from getting bug reports to getting detailed pull requests. And I am personally grateful for this.

My philosophy on incoming contributions comes down to a simple reality: every PR represents a real pain point caught in the wild. It means a real user ran into a problem in their day-to-day work, reported or tracked down the issue, found a working solution, and felt the pain acutely enough that they took the time to package it up and upstream it.

As a maintainer, that is an infinitely better signal than my own laundry list of wish-list items sitting in the issue tracker. Even if the implementation needs additional work, someone just handed me proof of a real-world problem along with a concrete starting point to fix it.

Multi-Agent Code Review on My Side of the Table

The catch is that many coding models and agents still have a bad habit of just doing the fix—taking the shortest path to make a test pass without stepping back to look at the whole library. And not every contributor has learned how to tune their agents for deep engineering review, or has access to the best models for it.

The first pass of AI-generated code often needs work, and that is okay. Just as we spent years developing the discipline to master our programming languages, debuggers, and in some cases, write m4 macros, we now need to develop the muscle for using AI tools effectively—learning how to steer them, verify their assumptions, and audit their output.

For me, building that muscle meant putting AI on my side of the table to bridge the gap. Before I even bother reading a contributor's patch line-by-line, I religiously run /code-review in Claude and /review in ChatGPT [1].

These review passes routinely catch far more than narrow syntax nits—they surface genuine architectural flaws, subtle API breakage, rare side effects, and terminal conformance problems that would take me hours to trace manually [2].

Armed with that feedback, I can point contributors toward concrete architectural fixes—or, if someone disappears after opening a PR for a critical bug, I can clean up their branch and land it myself.

Some maintainers lament that reviewing these PRs feels like talking to someone who just copies and pastes feedback into a remote LLM. But we are still communicating with a human being on the other end. That gives us a real opportunity not just to coach contributors on how to use these tools responsibly, but to model and adopt better engineering practices across our industry. Given that the economics of top-tier reasoning models aren't ideal for everyone doing casual open-source work, I’m happy to foot that bill on my end for the sake of keeping my library healthy.

The Friction of One-Size-Fits-All Rules

This brings me to why I’ve been watching the broader debate around AI policies in large projects like Gnome and Godot with a lot of sympathy—and a bit of concern.

At Godot's massive scale, protecting volunteer reviewers from unvetted submissions and the occasional Promptsplainer is a completely understandable instinct. Nobody wants their Netflix time eaten up by drive-by code the author didn't bother to understand.

The challenge is that drawing hard lines around how code was written inevitably creates friction for experienced contributors who are already doing the architectural heavy lifting. Recently, I tried to upstream a small, production-tested bug fix to Godot as a trial balloon—one of roughly 60 to 80 improvements my team has built and tested in our own tree. Because I disclosed my AI workflow honestly, it ran afoul of their strict guidelines and was rejected.

When a policy polices provenance rather than quality, two unintended things happen:

  1. It penalizes transparency: Honest contributors who disclose their tools get blocked, while others are quietly nudged to keep quiet and pretend they typed every character by hand.
  2. Good fixes stay locked in private forks: Rather than wrestling with policy friction, teams like mine simply keep useful patches in our own internal trees. Because AI has made maintaining forks trivial, it doesn't slow us down—it just means the broader community misses out on fixes that could have helped everyone.

Accountability Over Provenance

The Promptsplainer is a real headache, and open-source projects are right to push back against uncomprehending, drive-by contributions.

But the antidote isn't banning the tools—it's requiring human accountability paired with better review tooling on the maintainer's side. At the end of the day, as the maintainer of a library, it is my role to make the editorial decisions about what goes in—and the people who depend on my library trust me to make those calls. Whether a patch started in an LLM prompt or was typed out by hand doesn't change that responsibility. If a contributor understands the problem, has tested the fix, and can engage thoughtfully in review as a human being, the models they used in their editor shouldn't disqualify their work.

Of course, scaling that kind of review culture when a handful of core maintainers are already stretched thin is a structural challenge all its own.

[1] With Claude I use Opus 5.5 alongside /review in GPT 6.2 with Sol-xhigh. In Claude, /code-review actually launches multiple distinct agents in parallel, each focused on different goals and common engineering mistakes. Looking forward to Antigravity to get this.

[2] In fact these reviews are so thorough, that I use /code-review on my own changes before shipping. In a seven-year-old codebase, it is impossible to keep every historical invariant in your head at once, and those agents regularly catch edge-case scenarios I had completely forgotten about.

Posted
Richard Stallman
Google enshittificaton with supposed intelligence

Google Search has been unjust for many years, as it required users to run nonfree JavaScript code that it sent to users' browsers. But enshittification with supposed intelligence has made it much worse. It has become Google's Smirch.

Posted
Richard Stallman
Strongest El Niño in centuries

* Strongest El Niño in centuries adds urgency to UN climate talks as vulnerable countries face "compounding" crises.*

Such a damaging El Niño, with its damage, was inevitably going to happen some day. Its happening in the time of a climate defense summit is good fortune.

Posted
Bram Cohen
The Rules Of Dominoes

Robin Linus has done some truly heroic work showing that verbatim Poker can be done in Bitcoin script as it exists today with no extensions. This is still very much a prototype. Much further along than a prototype Nokitlan is supporting games on Chia, including dominoes (this is built on the Chia gaming work of been doing as the fundamental building block and interoperability substrate). Which got me thinking: What are the best dominoes rules to play on Chia? There are many different dominoes variants, with not that much standardization. The tournaments are a four player version which has a very bridge-like feel and is obviously a complete nonstarter for remote play because it’s just begging for cheating via collusion. For two players there isn’t even a completely standardized set of rules, so that’s my excuse for pondering variants, which I’ll do now.

The double-six set of dominoes is a pretty good start for the medium, because if two players alternate discards that’s 28 turns total max is which is more than ideal but somewhat workable. But there are some things which are problematic:

Thanks for reading Bram’s Thoughts! Subscribe for free to receive new posts and support my work.

There’s a lot of go-fish-like cheating, where a player is required to do something based on information only they can see. Go fish is a terrible game, it has tons of cheating by accident which small children will of course do and then learn that that’s a good thing to do. At the very start of the game a player has to do discard with a double or can use any tile they want out of their hand if they don’t have one. It’s much simpler to just make the double six be the first tile on the board at the beginning of every game.

Passing is only allowed if you don’t have a legal move to play, which is a obviously also a huge cheating problem. This doesn’t come up all that often because usually it’s counterproductive to pass, but adjudicating it after the fact is a disaster, especially in combination with tiles getting pulled from the boneyard later so there’s no way for an outside observer to tell what tiles each person had at each time from later play. It’s much better to simply allow passing at any time. If two passes in a row happen then the hand is over. It wouldn’t be too difficult to implement a rule that a player can’t do a second pass ending the hand if they have a legal move but I suspect that would result in some very cheeseball strategies where one player passes until the other player has very few tiles and then with the greater control of ensuing play easily wins the hand.

Pulling from the boneyard is awkward to implement. Notably the common four player game has no boneyard. All 28 tiles are dealt out at the beginning, 7 each to four players. It does result in some interesting hidden information play, but the lesson from Poker seems to be that when it comes to hidden information less is more, and you want it to stay relevant up until the very end instead of having everything eventually become known and then a straightforward endgame. To that end, and because this is a game mechanic I’ve long wanted to use, I think the best approach is to deal out all tiles at the beginning of the game except for one which is permanently out of play and the players don’t find out until the very end which one it is. Then the game becomes all about guessing what that tile is based on the opponent’s play and there’s lots of room for deceptive play.

There’s some random advantage to the player who gets to go first. That could be counteracted with an amount added to the final pip count like komi in Go but there would need to be some playtesting to figure out what a good value for that would be.

Putting this all together here are the draft rules of Bram’s dominoes. Needs playtesting:

Played with double six dominoes. At the start of the hand a random domino is removed from play, the double six is played out as the start, and the other 26 dominoes are dealt out 13 to each player. The first player to play is selected randomly.

Players alternate putting down a domino which extends one of the two ends of the chain on the board. They must match the number at the end of that chain with the domino they place. A player may alternately pass and must do so if they have no legal placement move.

If a player plays their last tile then they win. If both players pass in a row then their remaining tiles are revealed and which ever one of them has fewer total pips wins (possibly plus komi added to the player who went first).

Thanks for reading Bram’s Thoughts! Subscribe for free to receive new posts and support my work.

Posted
Tom (Jon Lund Steffensen)
Linux-webhotel: Derfor kører de fleste hjemmesider på Linux-servere

Linux webhotel: hvad det er, og hvorfor det er det mest populære valg

Et Linux webhotel er en hostingløsning, hvor hjemmesider, databaser, e-mail og webapplikationer kører på en server med Linux som operativsystem. Et webhotel betyder i praksis, at flere kunder deler serverressourcer som CPU, RAM, lagerplads og netværk, hvilket også beskrives i grunddefinitionen hos Webhotel – Wikipedia, den frie encyklopædi.

Linux blev første gang offentliggjort af Linus Torvalds i 1991, og operativsystemet har siden udviklet sig til en central platform for webservere, cloud-infrastruktur og containerteknologi. I 2026 er Linux fortsat tæt forbundet med klassiske webteknologier som Apache, Nginx, PHP, MySQL, MariaDB og SSH-administration.

Linux webhotel er populært, fordi det passer til de teknologier, som mange hjemmesider faktisk bruger. W3Techs har i flere år målt, at WordPress driver over 40 procent af alle websites globalt, og WordPress’ standardmiljø bygger typisk på PHP og MySQL/MariaDB, som historisk har været tæt integreret med Linux-servere.

Et Linux webhotel skal ikke forveksles med, at brugeren selv skal kunne bruge Linux på sin egen computer. De fleste kunder administrerer deres hosting via kontrolpaneler som cPanel, Plesk, DirectAdmin eller hostingudbyderens eget panel, mens Linux primært fungerer som serverens tekniske fundament.

De største fordele ved Linux hosting: stabilitet, sikkerhed og ydeevne

Linux hosting forbindes ofte med stabilitet, fordi Linux er designet til flerbrugerdrift, procesisolering og lang oppetid. Enterprise-distributioner som Debian, Ubuntu LTS, AlmaLinux og Rocky Linux tilbyder sikkerhedsopdateringer over flere år, og Ubuntu 24.04 LTS har eksempelvis standard-support frem til 2029 ifølge Canonicals officielle livscyklusmodel.

Sikkerheden i et Linux webhotel afhænger ikke kun af operativsystemet, men også af patching, isolation, firewall-regler, backup og adgangsstyring. I praksis bruger mange udbydere teknologier som CageFS, chroot-miljøer, ModSecurity og Fail2ban til at begrænse skader, hvis én konto bliver kompromitteret.

“Security is a process, not a product.” Citatet tilskrives sikkerhedsforskeren Bruce Schneier og bruges ofte som en kort forklaring på, hvorfor hosting-sikkerhed kræver løbende vedligeholdelse frem for én enkelt teknisk løsning.

Ydeevnen på et Linux webhotel påvirkes af webserver, PHP-version, caching, disktypen og antallet af kunder pr. fysisk server. Ifølge PHP-projektets egne udgivelsesnoter har nyere PHP-versioner typisk forbedringer i ydeevne og hukommelsesforbrug, og derfor bør et webhotel i 2026 som minimum tilbyde aktivt understøttede PHP-versioner som PHP 8.2, 8.3 eller nyere.

Linux er også velegnet til automatisering, fordi serveradministration kan udføres via shell-scripts, cron-jobs, SSH og API-baserede deployments. Hvis du vil forstå operativsystemets tekniske rolle bedre, giver Hvad er Linux? – En komplet guide til det åbne styresystem en bredere forklaring på kernel, distributioner og open source-modellen.

En praktisk fordel ved Linux hosting er, at meget dokumentation, fejlsøgning og community-viden allerede er skrevet til Linux-baserede servermiljøer. Stack Overflow Developer Survey har i flere år vist, at Linux er blandt de mest anvendte platforme blandt udviklere, hvilket gør det lettere at finde løsninger på fejl i Apache, Nginx, PHP-FPM, Composer eller MySQL.

Linux vs. Windows webhotel: hvilken serverløsning passer til dit behov?

Linux og Windows webhoteller kan begge drive professionelle hjemmesider, men de passer til forskellige teknologistakke. Linux vælges typisk til PHP, WordPress, Laravel, Drupal, Joomla, Node.js og open source-databaser, mens Windows hosting især er relevant ved ASP.NET, Microsoft SQL Server og klassiske IIS-baserede applikationer.

Windows Server kræver normalt licenser, mens Linux-distributioner ofte kan anvendes uden licensbetaling til selve operativsystemet. Microsofts officielle licensmodel for Windows Server er kernebaseret, hvilket i mange hostingmiljøer gør Windows dyrere at drive end en tilsvarende Linux-platform.

Kriterium Linux webhotel Windows webhotel
Typiske teknologier PHP, MySQL, MariaDB, WordPress, Laravel, Apache, Nginx ASP.NET, C#, IIS, Microsoft SQL Server
Licensmodel Ofte open source uden OS-licens pr. kunde Kommercielle Microsoft-licenser kan påvirke prisen
Kontrolpaneler cPanel, DirectAdmin, Plesk, egne paneler Plesk, IIS Manager, udbyderspecifikke paneler
Typisk målgruppe WordPress-ejere, PHP-udviklere, mindre virksomheder, bureauer Organisationer med Microsoft-stack og .NET-applikationer
Kommandolinje SSH, Bash, cron, Git, Composer PowerShell, IIS-værktøjer, Windows Scheduler

Valget bør afhænge af applikationens afhængigheder frem for en generel idé om, at den ene platform altid er bedre. En WordPress-side uden særlige Microsoft-krav vil næsten altid være enklere at placere på Linux, mens en intern applikation bygget i ASP.NET med SQL Server kan være mere naturlig på Windows.

Når pris, teknologistak og supportniveau skal vurderes samlet, kan det være nyttigt at sammenligne konkrete hostingpakker på tværs af diskplads, PHP-versioner, backup, oppetidsgaranti og supportkanaler; oversigter som findwebhosting.dk kan i den sammenhæng bruges til at skabe et mere struktureret beslutningsgrundlag frem for kun at sortere efter laveste månedspris.

Netværkskvalitet er også en del af hostingoplevelsen, selv om den ofte overses i sammenligninger mellem Linux og Windows. For danske brugere kan lav latenstid, god peering og en stabil internetforbindelse have mærkbar betydning, og emnet hænger naturligt sammen med principperne i Billigt internet: 7 måder at få lavere pris uden at miste hastighed.

Optimal kompatibilitet: hvorfor WordPress og PHP fungerer så godt på Linux

WordPress fungerer særligt godt på Linux, fordi platformen er skrevet i PHP og typisk bruger MySQL eller MariaDB som database. WordPress.org anbefaler i 2026 fortsat moderne PHP-versioner, HTTPS og MySQL/MariaDB-kompatible databaser som centrale krav til drift.

PHP har historisk været tæt knyttet til Linux, Apache og MySQL gennem den klassiske LAMP-stak, som står for Linux, Apache, MySQL og PHP. I moderne hosting er LEMP-stakken også udbredt, hvor Nginx erstatter Apache, men Linux forbliver operativsystemet under webserver, database og applikationslag.

W3Techs har i 2026 fortsat PHP registreret som et af de mest anvendte server-side programmeringssprog på offentlige websites. Det gør PHP-understøttelse, PHP-FPM-konfiguration, OPcache og versionsvalg til konkrete kvalitetskriterier, når man vurderer et Linux webhotel.

WordPress-ydeevne afhænger ofte mere af serverkonfiguration end af temaet alene. Et Linux webhotel med SSD eller NVMe-lager, aktiv OPcache, HTTP/2 eller HTTP/3, objekt-cache og korrekt databasekonfiguration kan reducere svartider, især på WooCommerce-butikker og indholdstunge websites.

For udviklere giver Linux hosting en arbejdsgang, der passer til moderne softwareudvikling. Git-deployments, Composer, WP-CLI, cron-jobs og SSH-adgang gør det muligt at opdatere kode, rydde cache, køre databasekommandoer og automatisere gentagne opgaver uden at være afhængig af manuelle uploads via FTP.

  1. Kontrollér først, at webhotellet understøtter en aktiv PHP-version, fordi ældre PHP-versioner ikke modtager samme sikkerhedsopdateringer. PHP 8.1 nåede ifølge PHP-projektets officielle tidsplan slutningen af aktiv support i 2023, mens nyere versioner bør prioriteres i 2026.

  2. Undersøg derefter, om databasen er MySQL eller MariaDB, og om versionen matcher applikationens krav. WordPress og mange PHP-frameworks fungerer bedst, når databaseversionen ikke halter flere større udgaver efter.

  3. Vurder til sidst adgang til SSH, cron og backups, fordi disse funktioner påvirker både drift og fejlsøgning. Et billigt webhotel uden ordentlig backup kan blive dyrt, hvis en opdatering ødelægger en produktionstjeneste.

Linux hosting er også relevant for mindre specialprojekter, hvor en hjemmeside skal integreres med eksterne API’er, datakilder eller beregningsmoduler. Eksempelvis kræver softwareprojekter som Integrering af en valutaberegner i dine softwareprojekter ofte stabile servermiljøer, planlagte jobs og sikker håndtering af eksterne forespørgsler.

Sådan finder og sammenligner du det rigtige Linux webhotel til din pris

Det rigtige Linux webhotel bør vælges ud fra arbejdsbelastning, krav til support og forventet vækst. En lille informationsside kan ofte køre fint på delt hosting, mens en WooCommerce-butik, et medlemsområde eller en Laravel-applikation kan kræve mere RAM, bedre CPU-prioritering eller en VPS.

Pris bør ikke vurderes isoleret, fordi de billigste pakker ofte har begrænsninger på samtidige processer, inode-forbrug, databaseydelse eller backupfrekvens. I hosting betyder “ubegrænset plads” sjældent ubegrænsede ressourcer, da fair use-politikker normalt regulerer CPU, RAM, I/O og e-mailtrafik.

Hostingtype Typisk anvendelse Fordel Begrænsning
Delt Linux webhotel Blogs, mindre virksomhedssider, simple WordPress-sites Lav pris og enkel administration Delte ressourcer og mindre kontrol
Managed WordPress-hosting WordPress-sider med behov for support og cache Optimeret drift og automatiske opdateringer Ofte dyrere og mindre fleksibelt
Linux VPS Webshops, API’er, Laravel, Node.js, specialsoftware Større kontrol over servermiljøet Kræver mere teknisk ansvar
Dedikeret server Høj trafik, særlige compliance-krav, tung databehandling Dedikerede hardware-ressourcer Højere pris og mere drift

Oppetid bør dokumenteres med målinger, ikke kun markedsføringsløfter. En oppetid på 99,9 procent tillader cirka 8 timer og 46 minutters nedetid om året, mens 99,99 procent reducerer den teoretiske årlige nedetid til cirka 52 minutter.

Backup er et af de vigtigste punkter i en hostingaftale, fordi fejl ofte skyldes opdateringer, brugerhandlinger eller kompromitterede plugins snarere end total servernedbrud. En god backupordning bør som minimum beskrive frekvens, opbevaringsperiode, geografisk placering og om kunden selv kan gendanne filer og databaser.

Datacenterplacering påvirker svartid, databeskyttelse og supportmæssige forhold. Hvis målgruppen primært er i Danmark eller EU, vil hosting i EU ofte gøre det lettere at håndtere GDPR-krav, fordi dataoverførsler uden for EU kan kræve supplerende juridiske vurderinger efter EU-Domstolens Schrems II-praksis fra 2020.

Supportniveauet bør matche kundens tekniske kompetencer. En erfaren udvikler kan klare sig med SSH, logs og dokumentation, mens en mindre virksomhed ofte har brug for dansk support, hurtig gendannelse af backup og hjælp til DNS, SSL-certifikater og e-mailopsætning.

Et seriøst valg af Linux webhotel bør omfatte en teknisk tjekliste før køb. Tjeklisten bør mindst dække PHP-versioner, databaseversioner, SSL, HTTP/2 eller HTTP/3, backup, e-mailgrænser, inodegrænser, staging-miljø, SSH-adgang og mulighed for at opgradere uden nedetid.

Typiske udfordringer ved Linux webhotel og hvordan de håndteres

Den mest almindelige udfordring ved delt Linux hosting er ressourcebegrænsning, hvor én hjemmeside rammer grænser for CPU, RAM, I/O eller samtidige PHP-processer. Symptomerne er langsomme svartider, 508-fejl, 503-fejl eller periodiske timeouts under trafikspidser.

Plugin-konflikter og forældede temaer er en hyppig årsag til problemer på WordPress-baserede Linux webhoteller. Ifølge Wordfence’ sikkerhedsrapporter har sårbarheder i plugins og temaer i flere år været en væsentlig angrebsflade i WordPress-økosystemet, hvilket gør opdateringsdisciplin vigtigere end valg af operativsystem alene.

E-mail fra webhoteller kan give leveringsproblemer, hvis SPF, DKIM og DMARC ikke er korrekt konfigureret. Google og Yahoo skærpede i 2024 kravene til masseafsendere, og i 2026 bør selv mindre afsendere bruge domæneautentificering for at reducere risikoen for spamfiltrering.

Migrering fra ét Linux webhotel til et andet kræver planlægning, fordi DNS, databaseeksport, filrettigheder, e-mailkonti og SSL-certifikater skal flyttes korrekt. En typisk DNS TTL på 300 sekunder kan give hurtig omlægning, mens ældre opsætninger med høj TTL kan give længere overgangsperiode.

Logfiler er centrale ved fejlsøgning på Linux hosting, fordi de viser fejl i webserver, PHP og applikation. Access logs, error logs og slow query logs kan afsløre tunge databaseforespørgsler, manglende filer, bots, brute force-forsøg og fejlslagne cron-jobs.

En praktisk driftsrutine for Linux webhotel bør indeholde månedlig opdatering, ugentlig backupkontrol og kvartalsvis gennemgang af brugerkonti. Denne type rutine reducerer risikoen for nedetid, fordi mange hostingproblemer opdages, før de udvikler sig til akutte fejl.

Ofte stillede spørgsmål

Hvad er et Linux webhotel?

Et Linux webhotel er en hostingløsning, hvor din hjemmeside kører på en server med Linux som operativsystem. Kunden administrerer normalt siden via et kontrolpanel, mens Linux håndterer serverens processer, filer, databaser og netværksfunktioner.

Er Linux webhotel kun for tekniske brugere?

Nej, et Linux webhotel kan bruges af både begyndere og udviklere. Begyndere bruger typisk et grafisk kontrolpanel, mens udviklere kan få ekstra værdi af SSH, Git, cron-jobs og kommandolinjeværktøjer.

Er Linux bedre end Windows til WordPress?

Linux er normalt det mest naturlige valg til WordPress, fordi WordPress bygger på PHP og MySQL/MariaDB, som traditionelt kører meget effektivt på Linux-servere. Windows kan godt bruges til webhosting, men det vælges oftest, når applikationen kræver ASP.NET, IIS eller Microsoft SQL Server.

Hvilken PHP-version bør et Linux webhotel understøtte i 2026?

I 2026 bør et Linux webhotel understøtte aktivt vedligeholdte PHP-versioner som PHP 8.2, 8.3 eller nyere. Ældre PHP-versioner kan give kompatibilitets- og sikkerhedsproblemer, især når WordPress, plugins og frameworks opdateres.

Hvad skal jeg kigge efter, når jeg sammenligner Linux webhoteller?

Du bør sammenligne PHP-versioner, databaseversioner, backup, oppetid, support, lagerteknologi, e-mailgrænser og mulighed for SSH-adgang. Prisen er vigtig, men den bør vurderes sammen med ressourcegrænser og driftsvilkår.

Kan jeg flytte fra et Windows webhotel til et Linux webhotel?

Ja, det er ofte muligt at flytte fra Windows til Linux, hvis hjemmesiden bruger teknologier som PHP, WordPress, MySQL eller statiske HTML-filer. Hvis siden derimod afhænger af ASP.NET, IIS-specifikke funktioner eller Microsoft SQL Server, kræver flytningen teknisk tilpasning.

Posted
jwz (Jamie Zawinski)
My friends call it the big D
Periodically, the "d" key on my keyboard stops working. Always and only "d". Power cycling the keyboar█ █oesn't fix it. Plug the keyboar█ into a █ifferent computer: works fine over there. Rebooting the Mac fixes it. An█ I have now █iscovere█: the mere act of launching NoMachine (without typing anything) fixes it -- for a while.

It just █i█ it again, 5 minutes later, go█ █ammit.

Previously, previously, previously, previously.

Posted
jwz (Jamie Zawinski)
Periodic reminder that "Moderate Democrat" is how you spell "Republican" in San Francisco.
For historical reasons, it is impossible for a Republican to get elected in this city, so all the Republicans register as Democrats and call themselves "moderates". What they actually are is "neoliberals", or, if you like, methane-breathing space zombie free-marketeers. The people to the left of them call themselves "progressives". And "centrist" means somewhere between "Libertarian" and "MAGA".

Therefore:

S.F. Dems vote to oppose billionaire tax, defying state party:

The vote fell along a neat progressive/moderate divide. The local chapter took the step two months ago to alter its bylaws to allow itself to diverge from the state party's endorsement on California ballot measures. Previously, the local chapter's convention was to follow the state party on statewide ballot measures and candidates.

Prop. 40, or the billionaire tax, would levy a one-time, 5 percent tax on some 200 billionaires who lived in California on Jan. 1, 2026. Most of the money would be spent on health care, patching up a $30 billion hole in Medi-Cal due to federal cuts after President Donald Trump's "Big Beautiful Bill." The rest would go to education and food assistance.

Previously, previously, previously, previously, previously, previously, previously.

Posted
jwz (Jamie Zawinski)
Afterglow
Classic After Dark screen savers on modern macOS:

Afterglow is an emulator built specifically for running After Dark modules -- the original 68k module code, not a remake or port -- with no dependencies on ROMs or classic Mac OS. [...] Powered by the Musashi 68k CPU emulator, Afterglow reimplements just enough of the Macintosh Toolbox and OS APIs to run After Dark modules. No Apple ROM or classic Mac OS installation is required.

You can also browse and install software releases directly from the Internet Archive without leaving the app.

Previously, previously, previously, previously, previously, previously.

Posted
jwz (Jamie Zawinski)
Afterglow
Classic After Dark screen savers on modern macOS:

Afterglow is an emulator built specifically for running After Dark modules -- the original 68k module code, not a remake or port -- with no dependencies on ROMs or classic Mac OS. [...] Powered by the Musashi 68k CPU emulator, Afterglow reimplements just enough of the Macintosh Toolbox and OS APIs to run After Dark modules. No Apple ROM or classic Mac OS installation is required.

You can also browse and install software releases directly from the Internet Archive without leaving the app.

Previously, previously, previously, previously, previously, previously.

Posted
jwz (Jamie Zawinski)
libvlc
Dear Lazyweb, if you have succeeded in getting VLCKit to build on macOS 14, specifically BasicPlayerWithPlaylist, please let me know how you did it.

I downloaded the zip, did "compileAndBuildVLCKit.sh -x", which took like 4 hours but eventually completed seemingly-successfully. But then BasicPlayerWithPlaylist.xcodeproj can't find VLCkit.framework or VLCKit/VLCKit.h. This sounds like a build settings problem (Framework Search Paths had "../../../build/Release" on it but it seems to be in "../../../build/macOS" instead? But adding that didn't help it find the .h). Trying to build VLCKit.xcodeproj instead gets me complaints about missing static-libvlc and/or install-iphone-device/libvlc-device-static.a (I am not interested in iOS at all.)

Posted
Bram Cohen
Math Tidbits

Maximal Fences

I got nerd sniped by this video and with some help from my friend Claude was able to optimize some of the existing records and set some new ones.1

Infinite Tetration

There are a bunch of videos talking about how the infinite tetration function is only defined in a specific range and strangely only hits finite values at the ends of the range, but they don’t show what it actually looks like as a function, so here it is:

Thanks for reading Bram’s Thoughts! Subscribe for free to receive new posts and support my work.

You might wonder what’s going on at the end points here. On the right it’s going completely vertical. There’s nothing actually funny happening on the left, that’s just where the formula for it using the W function splits into two value. But we can make an analytic continuation of it like so:2

Since we’re defining tetration we can make the rules and there’s no reason to limit it to the happenstance constraints of W. But this still leaves the question of what’s going on on the right. It turns out it does extend, but going up, like this:

Looking at it this way raises the question of what the inverse of this function is. It turns out it’s y=x^(1/x). So the natural way of defining this stuff is to start with that function then wonder what happens if we take its inverse and wonder what happens at the points where the W function splits. But in that direction it seems like a bunch of bizarre leaps and generalizations which require justification, where if we instead work backwards from infinite tetration they’re strange phenomena which need explanation.

The big lesson to take away is that if you see W turn up you’re probably looking at the inverse of something, and being able to support that is how W has worked its way into the canon of standard functions.

Unit Distances

After thinking about the unit distance conjecture I came to the surprising realization that for any asymptotic bound it must be possible to constrain the points to being in a ‘barbell’ of two discs exactly a unit distance apart and only take a constant factor hit based on the diameters of the discs. Writeup over here3. I was hoping that this would result in better bounds but it works out that the amount of constant factor on tightening the disc radii exactly cancels out the reduction in number of points covered by their areas. But it does seem to imply that you can find solutions whose points are within arbitrarily tight disc pairs by taking subsets of ever larger solutions. So maybe the optimal solutions all have this form and have some parameterizable epsilon how big the discs can be where any value works.

I don’t know if this observation is novel or is something obvious to people actively studying the problem and they just haven’t written it up because it hasn’t lead anywhere. But I found it interesting and surprisingand hopefully you do too.

1

It turns out Claude was taking my suggested search areas and implementing them verbatim using SLSQP which explains why nobody else had already done it with a bare prompt.

2

It’s possible to do an analytic continuation into the complex numbers which produces all kinds of interesting stuff but I’m not about to make a half hour long video with lots of 3d animations. If somebody else does that I would much appreciate it.

3

I was discussing this with Claude as I worked through it and Claude’s profound inability at geometric visualization made it not follow anything I was saying until I got to this very specific statement and then it was able to prove it immediately in a much more elegant way than I had worked out.

Posted
jwz (Jamie Zawinski)
DNA Lounge: Wherein someone tries to sell our Prince scraps
There's someone on eBay trying to sell unused tickets to our 2013 Prince shows for $700 each. (I don't begrudge them the effort, but I'm still not linking to it -- and I realize that "asking for" is not the same as "receiving".) But what they don't know is, I have around 2,000 of those tickets left, and any time someone signs up for our Patreon and I mail them their membership card, I usually toss into the envelope a bunch of stickers and a Prince ticket or two.

So, you know, if someone were tempted to give this guy $700, maybe instead they could give DNA Lounge $20 or $30 a month. Then in addition to maybe getting a little souvenir while supplies last, they'd also be helping keep this place alive so that memorable shows like that can keep happening.

I've got uncut sheets of these things too, so if that's your jam, make me an offer...

The reason I have so many extras is because of how big printers work. Or worked, at the time, anyway. We needed like 3,200 tickets, but these things were stamped in large galleys, so if you ordered 3,200 they ended up printing 5,000 and threw the rest away. And for a show like this, we did not want those discards falling into the wrong hands and ending up on the scalper market, so we had them send us the entire batch from the print run.

A couple of years ago I asked the World's Biggest Prince Fan whether this stuff was worth anything, and their answer was (my loose interpretation) "Yes definitely, but only if you put a bunch of effort in to trying to act like an auctioneer, and are cagey about how much you have" and that sounded like a bunch of work and also kind of dirty, so I never bothered. (It would require more finesse and salesmanship than just dropping them on our merch page, in other words.)

But hey, literally nobody but me currently possesses one of these uncut sheets, sooooo... Scarcity!

[ Butterfly meme ] "Is this... Captialism?"

Posted

Planet Debian upstream is hosted by Branchable.