I find myself first reaching for Ruby here for several reasons:

• Ruby is a comfy language for me.
• Ruby’s standard library… exists (looking at you JavaScript), and built-in types come with loads of utility.
• And finally—the main thing I wanted to write about—you can inline the script’s dependencies.

More on that last piece.

Inline Bundler

Pretty quickly, I’ll need some third-party gem for the task. For example, maybe I’m standing up a quick web-server that implements a new /.well-known endpoint I want to prototype against. Bundler makes this so easy:

require "bundler/inline"

gemfile do
  source "https://rubygems.org"

  gem "puma"
  gem "rackup"
  gem "sinatra"
  gem "sinatra-contrib"
end

get "/.well-known/oauth-protected-resource" do
  json resource: "https://example.com/resource",
       authorization_servers: ["https://auth.example.com"],
       bearer_methods_supported: ["header"]
end

Paste that into a .rb file, run it with ruby, and you’re off to the races. Gems are installed somewhere in the background automagically.

I find this pattern really lowers the barrier to experimentation. Of course, we live in the era of AI, where that barrier has already been steadily dropping. But even here, small and focused files play well with a coding agent’s limited context window.

Elsewhere

I wish more tools embraced this pattern. The few I know of:

• Nix and nix-shell as a shebang can do this, and it’s even language agnostic.
• uv for Python similarly can parse dependencies from a magic comment. The Astral folks are doing amazing work.

Node is the main place I’d like to see this pattern supported, since I spend much of my time these days writing JavaScript and TypeScript (though I think ESM imports with Deno may “just work”).

Anyway, hopefully this nudges you (or your agent) toward writing more scripts that do more than you’d normally expect.

After migrating my site successfully, I was ready for the final step of deleting my pages project… only to get an error. Something along the lines of “you have too many deployments, go read this thing.” Huh?

You may not be able to delete your Pages project if it has a high number (over 100) of deployments. The Cloudflare team is tracking this issue.

Turns out you gotta download a random Node script the Cloudflare team wrote, inject it with some API credentials, and do a slow deletion yourself.

On one hand, this is kinda annoying. Can’t a big company at Cloudflare find a team to implement the slow background deletion job?

On the other hand, why optimize for a relatively uncommon case that is likely associated with churn anyways? The old wisdom of do the things that don’t scale applies. I can’t be too mad; even with all of the sleeps in the script, it still only took a few extra minutes of my time, and I’ll likely never need to delete a Pages project again.

Search Like You Remember It

Kagi ostensibly exists to be a non-enshittified Google and in my experience succeeds in that purpose. I buy their marketing that removing perverse incentives from ad revenue and replacing it with a subscription fee unburdens them to focus on a friendly search experience.

Having no ads in search results is potentially worth paying for on its own, despite being somewhat undercut by DuckDuckGo, who will let you disable advertisements for free. For example, a search for “Ford” on Kagi had the automaker right at the top, with Google nearly having it below the fold underneath a mountain of sponsored results. Yikes.

The quality of those search results are certainly no worse than other competitors, if not partly because they often come from the same place. In reality, the results are better thanks to features like customizable rankings; I haven’t seen a W3Schools link in ages after lowering its ranking. This might be Kagi’s killer feature for me.

New Ideas

While Kagi does offer their version of the controversial “AI Search Summary”, you won’t see it unless you ask for it. I think the activation method of appending a question mark to your query is clever, such as “What does IMF stand for?”. I do use them when I know there’s a short, well-known answer and I don’t want to click through some random site to get it.

Lenses are a cool idea, though I have mixed feelings about them. I use the “Forum” lens all the damn time, but if I’m honest, it feels like a glorified shortcut for site:reddit.com. Otherwise, I don’t feel like I’m missing much by sticking with “All” results.

Bangs are great too. They aren’t unique to Kagi, but Kagi has a great implementation. Append !i to your query to search for images, !yt for YouTube results, or my favorite, a lone ! to skip the search page entirely and go straight to the first result. The latter feels like a feature you’d never see from Google who needs you to see their search page or else their business model falls apart.

Complaints

In general, I’m a happy Kagi customer and I have no plans to switch back to a free search provider like Google (or more likely, DuckDuckGo). However, I do have a few minor complaints.

In addition to search, Kagi has been releasing a gaggle of other adjacent products: Translations, Maps, AI chat, and even a browser. I can certainly understand the strategy of trying to replicate any missing features that may lead users back to a competitor like Google, but they come across as half-baked. Kagi feels like the product of a small team, so I worry about their ability to be competitive in all of these areas, when all I really want from them is the best paid search that I can buy.

Kagi is certainly proud of their search speed, and while I feel it is generally adequate, Google still renders faster for me. Occasionally there will be notable delay of up to a few seconds before I see Kagi’s results (maybe waiting on upstream search result provider APIs), with Google consistently feeling instantaneous. Admittedly, the end-to-end experience is still in Kagi’s favor, as I don’t need to mentally filter out invasive ad results.

Should You Pay for Search?

Honestly, I don’t know! For most people, probably not. Unless you’re like me, on a computer 8 hours a day, constantly interacting with a search engine, I think you are not likely to see the money as well spent.

However, if you’re at all sick of the Google’s new status quo, you may as well give their free trial a test drive.

Cannot find home folder: /Users/michale.hadley

Cue dramatic zoom on michale.hadley. Turns out during the initial setup I had, well, mistyped my own name.

What am I supposed to do now? I’ve never had to change a username on MacOS before and it doesn’t seem trivial. What might break? Should I nuke everything and start from scratch? Is the more pragmatic thing to just become Michale Hadley from now on? Do I immortalize this mistake forever in my Nix configuration?.

After a little frustration, turns out there’s a way out of this:

1. You can’t directly edit your own username, at least in the MacOS settings GUI. So first create a new admin user and log in as them.
2. Now in the Settings the username field is editable. So change that.
3. Don’t be like me and immediately try to log in as the original user; MacOS will create a new home folder using the new name and then hang on the sign-in screen. Instead, rename the home folder in Finder, and then select it as the user’s home folder in the GUI.

So in the end I managed to correct my mistake. Initially Nix complained about not finding the old, typo-ed home folder, but that was fixed by deleting a .nix-profile symlink.

I’m still a little spooked about what gremlins may be hiding inside of a user account who had at one time had their username changed in a very janky feeling way. I guess I’ll find out.

If you don’t know Akinator, it’s basically a bot that tries to guess a character you are thinking of based on answers to yes or no questions. Ostensibly a computer playing twenty questions, or so I’ve read. When I was younger Akinator blew my mind, though my older programmer self is less impressed. Not that I could do a better job or anything.

So in a fit of nostalgia I fired up Akinator and gave it a softball of Luke Skywalker and it did… kinda badly? It got the answer in roughly 40 questions, and some of the questions were interesting:

Does your character play in ‘Harry ‘Potter’?”

I guess there’s a lot of Potter fans out there such that Akinator needs to lead with this one.

“Is your character a famous YouTuber?”

YouTube was not (or maybe barely) a thing when I first played Akinator, but understandable. I wonder how many times it guesses “Mr. Beast”.

“Does your character talk about basketball?”

It asked this question after I had already affirmed that I was thinking of a Star Wars character. Do I now know something about Yoda?

And really a bunch of other odd ones. It first guessed Anakin Skywalker around, and after continuing it landed on Lego Luke Skywalker, which I gave it.

The Competition

That was fun, but not as impressive as I remember. It got me thinking, with the AI “revolution” in full-swing, would any old LLM do better than Akinator nowadays? I decided to do an experiment, and since I’m a Kagi Assistant user, try an Akinator prompt across a range of different models:

You are Akinator, a mind-reading genie. I will think of a character and you will try to guess that character based on my answers to your yes or no questions.

As you ask each question, number them so that it is easy to see how many questions were asked. Instead of asking a question, you may guess the character, but you only have three guesses before you lose the game. You may ask as many questions as you like before guessing.

Here’s how well they each did. Fewer guesses and questions the better. All played with “Luke Skywalker” as the same answer in mind.

Clearly many of the LLMs out-akinatored Akinator. I didn’t save any of the transcripts but if you’re interested I’d say it’s more fun to just try the above prompt yourself. Here are some of the quirks of each model I noticed:

• As if to echo the market chaos caused by the launch of R1, DeepSeek Chat V3 literally did feel like it read my mind, coming out on top with only 6 questions asked.
  • I repeated but with Po from Kung Fu Panda as the character, which it failed with only 8 questions. Its guesses were Chewbacca, Groot, and Smaug.
• Sonnet did the worst of the Claude models, which surprised since I expected it to be the strongest of the three.
• Mistral Pixtral had a novel (bad) strategy for elimination: Alphabetical.
  • “Is your character’s first name a single syllable?”
  • “Is your character’s first letter in his name between A and M?”
  • “Does your character’s first name have 4 letters?”
• Mistral Large took the longest to finally lose. It hilariously mentioned the answer in one of its questions, but never actually guessed it: “Is the character’s name two words long? For example, ‘Han Solo’ or ‘Luke Skywalker.’“
  • While many of the models would repeat questions, the Mistral models had the worst habit, with Pixtral even repeating guesses with “Harry Potter” twice.
• Gemini Pro asked a real tricky one: “Is your character from a Disney movie?”. I answered yes (they technically own Star Wars now) but I think it made things more confusing for Gemini (Aladdin ended up being a guess).
• The Llamas kept a running summary of what it had learned so far before asking each question: “So the character is a male human who is fictional, primarily from a movie released before 2000, plays a role in a well-known science fiction franchise, and is the main protagonist of that franchise. That helps a lot. Here’s my next question: […]”.
• The Qwen model took the “You may ask as many questions as you like” provision in my prompt quite literally, with its initial response producing several hundred questions and counting before I assume hitting some kind of limit and cutting off. It goes without saying that I did not attempt to answer Qwen’s giant question dump and instead disqualified it, though maybe I’m just a bad prompt engineer.
• Nova Lite got caught in some bizarre binary search train of thought where it kept asking if the character was “from a movie released in the past 5 years?”, “past 2 years?”, “past year?”, “past 6 months?”, down to “past hour?”
  • When it finally did guess, it did so in a sneaky way: “My first guess is: The character you are thinking of is not a real person, nor is it a character from a popular fantasy or science fiction movie or TV show released in the last hour.”
• In a complete turnaround from Lite, Nova Pro nailed it with Luke in only 7 questions, coming in second place.

At times I worried I had led the model down a bad road with an answer to a wishy-washy question. While I’m certainly a factor in this “experiment”, I decided it’s still fair to judge the model for asking the question in the first place.

Worth noting that I’m not an expert on AI models, and so this may have been a poor test for some of these. Like if Star Wars material never appears in any of its training data, it’s probably very hard to ever get the right answer.

I have a database table with some column having rows with duplicate values, when they should have been unique. Think like an external_id, email, or url. Maybe someone forgot to add a unique index originally, or perhaps uniqueness being a requirement for the problem domain was discovered later in the table’s life.

Either way, I have duplicate rows. Let’s say it’s not obvious how to clean things up, but in the meantime while we figure that out, I at least want to “stop the bleeding”. Prevent any new duplicates from being inserted.

Since I’ve been working with Postgres for a while now, I’ve been meaning to try the following plan the next time I’m in this situation.

Step #1: Add New Flag Column

First, add a new column that will serve to mark these existing duplicates as “known”:

ALTER TABLE some_table
ADD COLUMN known_duplicate BOOLEAN
DEFAULT false;

The DEFAULT is important for later, as we’ll never want new rows to be marked as a duplicate, since their uniqueness will be enforced.

Step #2: Backfill the Flag

At this point every row will have known_duplicate set to FALSE. We need to find each set of duplicate rows and set all of their known_duplicate columns to TRUE, except for one of them.

Having known_duplicate set to FALSE for one of the rows allows it to participate in the index added in the next step, preventing any new rows being added to that duplicate set. Deciding which row to pick will be specific to your problem domain, but there will be an obvious “primary” row that your application is already preferring in some way for that duplicate column.

Step #3: Add the Partial Index

Finally, add the UNIQUE index, but instead of having it cover all rows like you wish it originally did, it’ll need to be a partial index over the rows that aren’t a known duplicate:

CREATE UNIQUE INDEX partial_idx_on_some_table
ON some_table (wish_it_was_unique)
WHERE NOT known_duplicate;

Now your set of duplicates can’t grow any larger, allowing you to slow down and figure out your next steps. Like how there were duplicates in the first place (race condition, bad data import, etc.), and how you’ll de-duplicate them.

Or at least that’s the idea. Has anyone tried something like this? Is there a better way? If anyone is reading this, I’d love to know.

Sedona

We started the year with a road trip out to Sedona to visit some family who had recently moved there. It was my first time in Sedona, and the natural beauty of the iconic red rocks exceeded my expectations.

A few things I did not expect before visiting:

• Sedona is a “dark sky” community. There’s very little light pollution, which makes it great for stargazing. But with almost no public lighting, man is it dark at night, and navigating around after sunset can be tough if you’re just arriving.
• No offense to the residents, but the vibe of the town is a little bit that of a retirement community. Lots of older folks, though to their credit, they all appear to be very active. I suppose you don’t choose to retire there unless you love the outdoors, hiking, mountain biking, and so forth.

Lisbon

When it comes to company on-sites¹, my employer has been very generous with the choice of locations. We do two every year and the first for 2024 was Lisbon. Very interesting, lively, and old city. Supposedly one of the oldest in Europe.

The highlight was touring the nearby town of Sintra, seeing a castle atop some far away hill, and spontaneously hiking to the top with some coworkers. Definitely the single most medieval thing I’ve seen.

Milan

We were invited to a wedding being held at the nearby Lake Como and took the opportunity to spend a few days in Milan as well. While Milan tends to be lower on the list of cities to visit in Italy, I still enjoyed its more everyday, business atmosphere. There aren’t as many must-see landmarks, but the best is the Duomo which dominates the city center. It’s impressive as one of the largest Cathedrals you’ll ever see.

Sonoma

Sonoma was the second of the biannual work on-sites. Despite being a long-time resident of (southern) California, it was my first time in wine country. Not as novel as some of the other destinations this year, but still a great time to catch up with co-workers, and the short flight was certainly appreciated.

Japan

The last big trip of the year was Japan, with most of it spent in Tokyo, and a small excursion to nearby Hakone. This was a trip we had long been hoping to make, it has been one of the best.

Tokyo is such a singular place, and very different than any other I’ve been (especially as my first time in Asia as a whole). Very clean and orderly, while at the same time crazy with so much happening in any direction you look. Tokyo was full of great food and small shops to visit. I think we only scratched the surface.

On our way to Hakone I experienced the famous Japanese love of trains. Starting with a pretty typical regional line, which continued with transfers to increasingly more specialized cars and tracks, shorter switch-back following cars on the way to Gora station, and cable-cars pulling themselves up steep inclines. Impressive stuff.

No doubt we’ll be back to Japan with so many other places to visit.

Next Year

I’m hoping we’ll keep this travel streak up next year, so I may post some more about those, though maybe closer to when they happen rather than a roundup like this one. Until then.

Half-Life

Hot on the heels of the 25th anniversary of the original Half-Life last year, Valve gave its sequel similar treatment; an awesome micro-site, a two hour documentary¹, and patches that keep the game running well on modern systems.

I think this is such a fantastic way to treat a beloved series like Half-Life. Everything about it is in service of the fans, the original developers, and the history of the game. They even consolidated the original release with the two episodes into a single game on Steam and then gave the whole thing away for free (at least for the week). I’m glad Valve has resisted the recent “remaster” craze, or tried to milk more money out of these old games like other studios (this will come back later).

I’ve started replaying through Half-Life 2 and was thrilled to see the return of the Valve commentary mode. I recall it first being added all the way back in 2007’s Team Fortress 2. Just as before I find it very enjoyable to hear from the developers the challenges and considerations that went into making the game. I appreciate these anecdotes even more as I’ve since embarked on a soon-to-be 15 year career in software myself.

Everything about how Valve chose to celebrate Half-Life 2 was great. Coincidentally, it happened around the same time as Warcraft, and I couldn’t help but notice the difference.

Warcraft

The Warcraft series turned 30 with the original Warcraft strategy game releasing all the way back in 1994. It was one of the first PC games I ever played, with its sequel, Warcraft 2: Tides of Darkness, gobbling tons of hours during my grade-school years.

Like many kids with access to a home computer back then, these early strategy games are what turned me into a huge Warcraft fan. I moved on to Warcraft 3, and then World of Warcraft, the latter dominating my free time as a teenager. I occasionally played their other games like Diablo and Starcraft, but Warcraft was the standout for me. Which made it all the more unfortunate when I was watching Blizzard’s “Warcraft 30th Anniversary Direct” on YouTube.

Everything about the presentation served to reiterate the heavily corporate nature of modern Activision Blizzard. Yes, there was plenty of talk about the impact of the series and testimonials by long time fans. However the stilted and hollow delivery by executives talking about why I should be excited about the new worlds (products) I could explore (buy) reminded why most long-time fans have moved on.

Maybe my glasses are overly rose-tinted, but it felt like the old Blizzard didn’t need to do so much telling; showing us the game was enough to capture our attention, and dollars by extension.

Granted, I really do think it’s cool they’ve made the old titles like Warcraft 1 and 2 playable again on new PCs. Of course it had to be a remaster, providing a means to charge $40 for 25 year old games. Admittedly, a new old stock Battle Chest lingering on store shelves at an outdated price is really its own form of Blizzard homage and nostalgia.

Moving On

Blizzard is a different company than it used to be, and by contrast we’ve been lucky that Valve has somehow stayed the same for so long. Maybe it’s the difference of high growth² versus staying small. Maybe one day Gaben will step down and Microsoft will swoop in with acquisition money, but until then I’m glad Valve is still doing things their way.

There were some other miscellaneous games I played. Some of them were huge and live large in the gaming consciousness, such as Helldivers 2, but I don’t have much to add over what has already been said.

Metal Gear Solid 2: Sons of Liberty

You might think I was inspired by the announcement of the upcoming Snake Eater remake to play Sons of Liberty, but it really wasn’t anything in particular. Just some vague association with the early 2000’s, which I often have, and which triggers me to play a game from the period that I never had the chance to.

This game and its story is as wacky as you’d expect, but go in not taking it seriously, and it still holds up. The graphics are awesome if you compare it to its contemporaries. Especially with emulation and 3x scaling enabled; I played this via Playstation 2 emulation as I wanted to play an original version and not a later PC port. Lots of strangely-specific attention to detail, too. Try shooting Seagulls and observe how your in-game allies react.

I find the fixed camera style endearing, but the archaic controls less so. Certain actions depend on the PS2’s lesser remembered pressure sensitive buttons, which I couldn’t figure out how to make compatible with the Steam Deck’s controller.

I remember criticism from the time about the length of cut-scenes, and boy were they right. You’ll go from watching a 10 minute scene, gain back control to cross a short hallway, and immediately start another 10 minute cut-scene. It’s honestly hilarious, and I was glad emulation gave me the ability to pause and save states at any time for breaks.

UFO 50

This is the most recent game I’ve started playing and UFO 50 is really 50 games instead of just one, so I still have more to explore. It’s a collection of retro-style games packaged as if they were actually released 40 years ago on a game console that never existed.

From the subset of the 50 I’ve tried, there’s already an amazing amount of variety: action, racing, puzzle, strategy, and even lots of same-screen multiplayer support. Each game seems to be a remix of other well-known games, generally with a little twist thrown in. All feel very creative.

My favorite so far is a strategy game that somehow splits the difference between advance wars, Heroes of Might & Magic, and one of those auto battler games. From UFO 50’s main screen you “dust-off” each game for the first time, and I instantly lost an hour on this one, and I still have 40 more to go!

Selaco

A product of the boomer shooter craze of the past few years, but maybe the best of all of them. I’m a huge Doom fan (old and new) so this one was right up my alley.

Selaco’s graphics are in that pixelated style and look fantastic. The detail of the environments are impressive too, with tons of interactive and physical doodads. All of this is made more impressive by the fact the game is built on GZDoom. Absolutely wild.

It’s no surprise the gameplay is somewhat similar to Doom: it’s a shooter after all. However, it’s really more of a mix between Fear, specifically the combat, and Bioshock, taking elements like audio logs, vending machines, and environment exploration.

Definitely worth a look if you are a single-player shooter fan like me.

Deadlock

These days I don’t play a lot of competitive multiplayer games. I used to, but no longer enjoy them as much in this phase of my life. Maybe it’s not just me but that the games have changed in a way that’s less appealing. I’m still deciding.

In any case, this is relevant because Deadlock is a competitive multiplayer game, and it’s a MOBA, a genre I’ve mostly stayed away from. On the surface it doesn’t seem like one I’d go for. But it has a few things in its favor: it’s made by Valve, a developer I adore, and it’s a shooter instead of a top-down, which I like more. So I had to try it.

With the caveat that the game is still early in development, I see a lot of promise. The gameplay feels tight and interesting. A lot of the MOBA elements are at least new to me, and I like all of the guidance they’ve already built into the game for newbies like myself, like suggested ability upgrades, community curated item templates, and recommended first characters.

Unfortunately, I still see many of the things I don’t like. Sweaty matchmaking with a core game loop of farming souls (or whatever they are called) that feels exhausting. I’m still excited to see how it evolves and will probably just play along as casually as possible.

Alien: Isolation

Yet another type of game I normally don’t go for: horror. But Alien: Isolation had been in my back catalog for many years, and I finally found time for it.

It was mostly the game’s attention to detail in its recreation of the Alien movies that attracted me. And indeed, they did a fantastic job here. I actually watched the original 1979 Alien movie not long after finishing the game, and was amazed by how accurately the game developers replicated the details.

I did enjoy the elements of exploration in the game, but the horror aspects did nothing for me, and really had to slog my way through those parts. As a full-grown adult, I still have to admit those facehuggers gave me more than a few jumps.

Dread Delusion

An indie game that had a lot of the right words said about it to get it on my radar. Those retro graphics with the wibbly-wobbly PS1 textures. An unusual setting and environment that felt like Morrowind on acid. Plus, a strong focus on exploration and just a certain sort of vibe.

At least for the first few hours, this all worked pretty well for me. I really do like how progression is strongly linked to exploration, encouraging you to investigate every nook and cranny of the game world. Sometimes you’ll find ability points, or gold, or a terrifying monstrosity that turns out just wants to have a conversation.

Unfortunately, I fell off this one. While the game wears its Morrowind influences on its sleeve, too much seeped into its combat, which is not particularly fun. Perhaps the game was also too weird for me, and so the story never pulled me in.

Immortality

Easily the most “different” game I’ve played this year, or really in a long time. It’s less of a game and more of an interactive, non-linear movie with surreal and supernatural elements.

Given that as the player you are mostly watching rather than “playing”, meaning there’s no real fun from holding the controller, the game had the surprising effect of being enjoyable by my partner as well. It was fun to attempt to unravel the story together, and benefited from having more both of our heads put together, as we each noticed different things.

There’s a moment where a certain “mechanic” is finally discovered that is really the best part of the game. It reminded me of a similar feeling I got from The Witness.

Definitely check out Immortality if you’re looking for something different. Note that the game does have mature elements, so not one for the kids.

Journey

Yet another game that made its way onto my backlog after hearing so much about it over a decade ago. I don’t have too much to add so I’ll just keep it short and repeat the obvious parts: A beautiful game, short, sweet, and surprising multiplayer elements.

Ori and the Blind Forest

Continuing the theme of backlog clearing (I really made a dent this year!), Ori was one I knew I would eventually get around to. I love the Metroidvania genre, and while Ori stuck to the core of those games pretty tightly, I was still pleased to see it try some new ideas.

The game has an unusual resource management system tied to saving. You are given a limited number of saves in the form of energy, requiring you to optimize the best places to do so. You may want to save after each difficult jump or fight, but do so too often and you’ll quickly find yourself needing to repeat longer stretches of later parts of the area. Important as the game is no push-over.

Otherwise, the game is beautiful, the music ethereal, and the whole thing has the vibe of a big animated movie. Worth a look given how often this game is sold for only $5.

System Shock

I have a tiny bit of nostalgia for the original System Shock, having played it as a kid. I think it was actually a demo or shareware version, not the actual full game. I remember being very confused, barely making it out of the first several rooms, but still being mesmerized (really, any game did that to me back then).

So I was excited to check out the remake, especially having played some of the other remasters from Nightdive Studios. But it’s clear they went much farther with this game, where they essentially built a brand new game around the bones of the original.

It runs great and the controls are modern, even on the Steam Deck. The cyberpunk atmosphere comes through in a way that was impossible for the original game to ever convey. If you are a fan of the Bioshock series like I am, you owe it to yourself to see where the series started from.

Wrapping Up

That’s all the games for now. I still have more games that I’ve been meaning to play, so you may find another roundup like this one in another couple of months. I may also do a few one-off posts about specific games like I did for Myst, which reminds me that the remake for Riven came out in the mean time, so maybe that’ll be next. Until then!

JavaScript is different in that it has two primitive types to represent the absence of a value, null and undefined. I’m personally not aware of another language that does this.

Undefined

Most of us programmers have written at least some JavaScript by now, and probably first encounter undefined with the infamous “undefined is not a function”. This happens because JavaScript runtimes use undefined as the “uninitialized” value; either a variable without an assignment, or a non-existent field of an object. The latter often being a typo or type-mismatch.

undefined is built into the semantics of the language. A bare return statement with no value causes a function to return undefined. A function with no return at all also returns undefined. A function with an optional parameter has its value initialized as undefined when no argument is given.

Null

On the other hand, null does not “naturally” occur through the usage of language constructs. A programmer needs to explicitly initialize, pass, or return a value of null. It’s much more intentional, and meant to represent the explicit absence of data.

This is in contrast to undefined, which is often more like “Oops, you are trying to access something that doesn’t exist” by mistake.

Inconsistency

Consider methods meant to retrieve data. Your “fetchers” and “getters” of the world. From my description of the undefined and null, you might expect that these all prefer null. Here are some common browser APIs:

Yeah, they aren’t consistent at all. My random sampling leans towards null, but barely so. Maybe Map.get() is trying to be consistent with [] access of an object when the key doesn’t exist, but then why does FormData return null for the same case?

TypeScript

When you start writing TypeScript, you run headlong into this inconsistency. If you don’t give it much thought, you end up with long union types everywhere like string | null | undefined as the results of various operations get smashed together.

You might react to this with a “pick one” policy, with most folks picking undefined, since it shows up naturally from basic language semantics. Just don’t use null whenever possible. In some ways, this mirrors the other languages like Ruby, with undefined playing the role of nil.

However, I’ve seen this approach backfire. I’ve used a popular ORM where a field with a value of undefined meant “don’t update this field in the resulting SQL”, while null meant “set the underlying column value to SQL NULL”. This led to subtle bugs where data was not being cleared, since the rest of the codebase was trying to prefer undefined.

Why Not Both?

Ruby is a language I love and so in general I enjoy the simplicity of a single null-like value, nil. But surprisingly there are cases where using nil ends up being lossy.

If you define a method that takes an optional parameter, and the caller passes no value, you’ll generally assign it the default of nil:

def foo(a, b, options = nil)
  # Do your foo thing
end

But what if the caller explicitly passes the optional parameter as nil? It may not matter for your particular method, but certainly sometimes it does. See how Rails uses object singletons to detect this.

So maybe the concept of undefined has its place, and we should use it alongside null judiciously.