2026-01-22 01:16:04
[Editor’s Note: I asked Antony to detail the process by which he created his interactive crime novel, Can You Solve The Murder?, which required some specific tools and a lot of intricate planning. —J.S.]
I’m an author, primarily writing crime and thriller novels. I’m best known for the Cold War spy movie Atomic Blonde, which was based on my graphic novel. I also write the Dog Sitter Detective murder mysteries, the Brigitte Sharp spy thrillers, and most recently the interactive novel Can You Solve the Murder?
Hang on — an interactive novel? That’s right. You see, in addition to all the above, I also write video games. I grew up loving both books and games of all kinds, and was fortunate (read: old) enough to have been a young boy when the original Choose Your Own Adventure, Fighting Fantasy, and Lone Wolf books were first published.
Those series are examples of what came to be called ‘gamebooks’, because they’re both forms smushed into one; books where the reader plays an active part, directing the story by making choices, like playing a game. (These days we sometimes get fancy and call them ‘interactive novels’.) This is achieved by dividing the story into numbered sections and sending you to read different sections depending on your choices.
If you’ve ever played a text adventure game on a computer…
…Then you’ve essentially played a gamebook, just with all the page-flipping done for you by the computer. The modern ‘Visual novel’ form is basically the same thing, too.
We call these types of story a branching narrative, because the choice map often looks like a tree, with each new section branching off into further choices.
At their heart, all video games are essentially branching interactive experiences. You make a choice — whether in text, or with a joystick, or by pressing an action button — and something happens in the game reflecting that choice. The presentation of those choices, and the results, is nowadays enormously more sophisticated and complex than it was in a 1980s text adventure. But at their core, all games are about players making choices and the system reacting to them.
Back in the late ’70s & ’80s, computer games looked very much like that screenshot above. They weren’t the interactive movies with ultra-realistic graphics, hours of cinematic music, and dialogue voiced by professional actors to which we’ve become accustomed. Instead, they were basic, often poorly-written, and any graphics were extremely primitive. Those early limitations allowed gamebooks to thrive. They were almost always better-written than computer games, with richer and more evocative text, and were also often illustrated by professional artists.
As video game technology accelerated, though, and digital interactive experiences became ever more realistic and immersive, gamebooks fell out of fashion. They became a curio, remembered with nostalgia by enthusiasts but largely forgotten by the mainstream.
I was one of those enthusiasts from the very start. In fact, I was so taken with gamebooks that eleven-year-old me even had a go at creating one of my own, written on a manual typewriter and illustrated with a ballpoint pen, which I then made my friends play. Sadly, the ambitiously-titled Hellfire of Death’s Caverns is lost to the mists of time, but looking back, it certainly explains a lot about my career.
You see, a love of gamebooks introduced me to early role-playing games and fantasy board games, which in turn led to an interest in game design… which ultimately led to me working in videogames as a writer and narrative designer. That career has run in parallel with my fiction writing for the past twenty years1, and more recently I’ve also established myself as an award-winning crime author.
I love writing and plotting crime and mystery fiction. I love writing and designing games. So I kept wondering: was there some way to combine the two?
That’s when I remembered the many hours I spent as a boy with my head buried in gamebooks. Was it possible to write a crime story in that format? How would I handle clues, red herrings, and all the other elements that people love about murder mysteries?
And how on earth do you plan a book like that, anyway?
While answers to many of those questions would take months of work to figure out, the last — how do you plan a gamebook? — was one to which I already knew the answer thanks to that early typewritten effort, and my later experience in game design: you build a huge flowchart.
The original gamebook authors drew their charts and maps by hand, on taped-together pieces of paper, with markers. If you’ve ever mapped a text adventure game in a notebook as you played, you were effectively doing the same thing in reverse.
Nowadays, though, we have better tools. Digital tools. Surely one of them would be better suited to making such a flowchart. But which one?
There are quite a few apps and services out there with which to build flowcharts. I was already familiar with Miro, an online service used by many game designers for brainstorming and building ‘paper prototypes’ (while there’s no paper involved these days, the name has stuck). I used Miro to plan a couple of proof-of-concept interactive short stories, which I wrote to get a handle on the format and test whether I could make it work.
I could, and friends to whom I sent the prototypes enjoyed them, so I began planning a book-length mystery.
But Miro is a subscription service, with only a small number of ‘boards’ available to free accounts. I was already at my limit, and reluctant to subscribe — believe me when I tell you that very few authors actually make a living writing novels! So I put aside the decision of what to use for a moment and focused on pitching the book itself. After all, if it didn’t sell, there’d be no need to build charts.
I needn’t have worried. Can You Solve the Murder?2 went to a bidding war, and was eventually acquired by Transworld in a two-book deal with a frighteningly short deadline. The matter of choosing a flowcharting tool had now become rather urgent.
My criteria were simple, but specific. I wanted to:
But have the option to snap-align objects to one another
Draw arrows that snap to destination objects with auto-applied directional heads, and that can be moved and modified freely later
Easily apply text labels to arrows that represent a choice (most do, but not all) and have those labels move with the arrows when modified
Draw multiple such arrows coming from a single object
Have multiple arrows also arriving at a single object
Mix object shapes, and be able to easily resize and re-color them however I wish
Write text directly into any object, and have the text be easy to modify: size, color, style, etc
Overlay shapes on top of one another (in order to track section numbers and clues discovered)
Finally, I needed it to feel intuitive and simple to use. While that’s not really a quantifiable metric, it was important to me. I knew I’d be spending many, many hours making and using these charts. As a former graphic designer, not to mention lifelong Mac user, I want my tools to get out of the way and let me work, not make me wrestle them into submission.
I spent the next couple of weeks trialling many different charting apps and web services. I won’t go into detail on them, but suffice to say that while several had the features I was looking for, none of them felt particularly easy or intuitive to use…
…Except Miro, to which I kept returning. So I sucked it up and used part of my book advance to pay for a subscription.
Ultimately, the best tool had to win out. I’m a big believer that using inferior tools simply isn’t worth the hassle and frustration they cause, no matter how cheap they may be, and while Miro still isn’t perfect (it doesn’t do per-character text sizing, and object color modification could be better), it remains the tool best suited to how I build gamebook flowcharts. In some ways, using Miro reminds me of a good vector design app, like the venerable Aldus FreeHand3 or modern Affinity Designer, with an ease to its interface and UI that at times becomes almost invisible (the highest compliment I can pay an interface).
So, how do I use all those features I mentioned to build a gamebook flowchart? Well, let’s take a look.
Colors: Green boxes represent ‘core story’ sections, which all readers will see, while yellow boxes are sections that will only be seen by readers who make particular choices. Blue boxes are ‘ghost’ sections — they don’t exist in the book, but here they note where a choice will send you to a different part of the chart. Dark green boxes are good endings, while black boxes are — you guessed it — bad endings that require you to return to #1 and try again.
Overlaid objects: Each main colored box has a numbered red circle overlaid on it at top-right — this is the number of that section in the book, to which you flip when directed. Some boxes also have a blue circle bottom-right for Clue Numbers, which I’ll explain below.
Text inside any object: In addition to text in the main boxes, the number codes in those red and blue circles are contained directly within the object, rather than being a separate text box placed on top and grouped. This sounds like a small thing, but it makes them much easier to manipulate, edit, and move around the board, and when you’re dealing with literally hundreds of these objects, it makes a big difference.
Grid alignment: Although most of the boxes are aligned with one another by choice, note that section #52 is not. This tangle demonstrates the value of not being forced into a grid:
Freeform arrows: Many boxes have multiple arrows branching out of them, and #46 even has multiple arrows branching into it. That’s an essential requirement. Again, that tangle reinforces the need for fine control over arrows.
Arrow labels: These show which section a choice (or conditional: see below) directs you to read next.
The blue circles are part of a system I call Clue Numbers. Many sections of Can You Solve the Murder? ask the reader to write down a Clue Number. In this example, the text of section #23 would end with the instruction:
Write down D2 in your notebook, then turn to #46
Later, if you decide to keep walking and thus turn to section #9, a conditional check is made. The text of that scene would end as follows:
If you have D2 written in your notebook, turn to #32
Otherwise, turn to #58
This mechanic is vital to the success of a clue-based mystery like Can You Solve the Murder? because it effectively allows the book to ‘remember’ your choices, and the information you’ve gathered, throughout the story — but does so in a way that doesn’t give the game away, because the reader doesn’t know what those codes mean. (I do, of course, and track them all in a spreadsheet).
It’s impossible to complete Can You Solve the Murder? without using Clue Numbers. Thus, it’s vital that I can easily assign them to the appropriate sections and move them around if need be when building the book’s flowchart.
One nice element of using an online service is shareability. This was necessary for Can You Solve the Murder? because of something called ‘structural edits’. In fiction writing, this is what happens after you submit your first draft, and the editor then suggests changes such as removing/adding subplots, modifying characters, perhaps re-ordering some scenes, and so on. All these things change the structure of the book, hence the term.
Making such large-scale changes to a gamebook manuscript would be absolute hell.
Every choice in a branching narrative sets off a cascade of subsequent choices and reactions, meaning structural changes ripple downstream throughout the book. In a regular novel, revising a manuscript to remove a secondary character can often be accomplished fairly easily. In a gamebook, though, it might require re-plotting half the story, creating new sections, re-ordering existing paths… a nightmare that could take weeks or even months to untangle.
To avoid this, I gave my editors password-protected view-only access to the flowchart board, something Miro makes very easy (I could also have given them full editing access, or not required a password), and asked them to make their structural edit notes based on the flowchart. None of us had ever worked on a book like this before, and doing so required a lot of mutual trust, especially with strange requests like this! But it worked out, and part of that was down to how easy Miro makes sharing and viewing.
You can also easily export a Miro board, or part of one, to a variety of formats, including JPG, PDF, and even CSV. While I haven’t needed that facility for Can You Solve the Murder?, I use it often in my video game work.
The truth is that I barely use a fraction of Miro’s power. There are Miro wizards out there who can make it do all sorts of things: presentations, prototypes, slideshows, you name it. I’ve worked with some of them at game studios. Apparently, even the NFL’s digital team uses it to plan gameday strategy for their apps and services. Miro’s YouTube channel has many such case studies, along with tutorials.
Me? I don’t need to do any of those things, and wouldn’t know where to start. But that leads me to the final thing I like about Miro: it doesn’t force that stuff on me.
One of my favorite and most-used apps is the writing software Scrivener, which I’ve used since it first launched in 2007. (I’m writing this very article in it.) A common refrain amongst its advocates is that most people only use 20% of Scrivener’s features… but we all use a different 20%. Miro feels very much the same.4
I expect most of you reading this aren’t writing interactive novels. But you might need to create charts or diagrams of some kind, and if so, I recommend Miro. Yes, there are free alternatives out there, but I haven’t found any that are as flexible and intuitive to use.
As for the book, Can You Solve the Murder? was published in summer 2025 and quickly became my best-selling novel ever. A sequel, The Forest of Death, is due this year… and yes, I used Miro again to build its flowchart.
2026-01-22 00:00:00
Enough about the icons, Lex tells us about his big night, Dan talks about his e-reader and Moltz is also on this episode.
2026-01-21 01:22:11
Apple’s Time Capsule was always a strange item in the company’s hardware lineup. Coming out in 2008, after the introduction of the sleeker version of the AirPort Extreme Base Station, the Time Capsule combined Wi-Fi gateway features with internal hard drive storage that Time Machine could access using the Time Capsule as network-attached storage (NAS).
Time Capsule was discontinued in 2018, and Apple will drop support for it with the release of macOS 27 this fall. Apple’s announcement has prompted many people—including Six Colors subscriber Mattias—to replace a long-used (but long-in-the-tooth) backup solution:
My mom still uses my old AirPort Time Capsule as a Time Machine backup. Since it won’t be supported any longer, what’s the best replacement?
Matthias wondered about a simple Network Attached Storage (NAS) device as a replacement, but notes, “I never had to touch the Time Capsule. Don’t think it’d be the same with a full NAS.”
It’s a problem. Many people have continued to rely on a Time Capsule for simplicity’s sake. Plug it in, configure, and (one hopes) forget about it until you need to retrieve a file or restore a volume. While hard disk drives don’t last forever, they keep ticking away for years and years before—often suddenly—giving up the ghost.1
At some point, all Time Capsule owners will need to face the problem, even if they’re not planning to update to macOS 27 El Cerrito (or whatever it’s called). So let’s think about a solution.
Time Capsule tried to solve some interrelated problems: while you can use Time Machine to back up to a different Mac on your network, that required you to have a Mac you can leave on all night, or at the very least one you didn’t want taxed by performing network backups while you were working away on it.
Time Capsule was always on, could be located anywhere you could plug it in, handled connections over Wi-Fi and Ethernet, and allegedly just worked. As any of us who owned one recalls, it often didn’t just work. Backups would fail, requiring erasure of the entire internal drive with no option to recover older backups.2 There was no Disk First Aid for Time Capsule.
This became extra galling as the models advanced to 2 TB and 3 TB capacities in their final versions. While you could add an external drive, that seemed to defeat the point of a standalone unit, particularly when you remember the port was USB 2.0 or 480 Mbps, far below internal hard drive rates.3
Nonetheless, we didn’t have better options when the Time Capsule came out, and for years afterwards. Internet-hosted backups are a good supplement, but I don’t think there’s a drop-in replacement that boasts the same ease, even if reliability were not an issue.
What can you do today to replace a Time Capsule and provide the functionality it offered? You have effectively three choices:
Time Machine’s key attributes are that you can set it and forget it, that it prunes older backups over time without intervention, and that it’s integrated into macOS for easy (by some definitions) retrievals of older versions of files and deleted files.4
I have a desktop Mac, and use cloning software to make a nightly duplicate of my startup volume and Backblaze for Internet-based encrypted backups. Most of my documents are on Dropbox or iCloud Drive, so I felt I had sufficient redundancy. I balked at using Time Machine, as it was so fiddly and unreliable for a long time, requiring erasing external drives or poking at low-level settings.
The other issue was how inefficiently Time Machine interacted with hard drives. During active backups, whatever CPU management Apple said was in place, my computer would be affected, and I’d also have to hear the whirs and whines of active drives.
At some point (I’d say about four or five years ago), Time Machine improved in quality, and solid-state drives started to drop in price for compact 1 TB USB 3.x models, which could connect at 5 to 10 Gbps; later 2 TB SSDs dropped in price enough to fit my budget, too. I’m embarrassed to show a photo, but here are three of the five SSDs connected to my Mac: four are 2 TB, and one is a dual 4 TB RAID 0 configuration that gives me an 8 TB volume.5 Two others are plugged in elsewhere!
Using an SSD plugged directly into a Mac could be a solution. Even if you don’t have a desktop Mac, plugging in an SSD on a regular basis when you’re using the laptop on a desk or other location is feasible, too: Time Machine automatically picks up when an associated volume is mounted.
The improvement I saw in Time Machine and the switch to SSDs made it feasible for a desktop Mac to serve as the networked destination for four family laptops, including mine. So I didn’t go down the NAS path. There are inexpensive NAS options that can be remotely administered, if you’re helping a relative or need remote access, and support the protocols needed for Time Machine. Synology has a tech-support note on using Time Machine with its devices. And I recommend looking at Tailscale for ensuring you can always tunnel in; the company also has NAS support. (See my long background look at Tailscale at TidBITS, and a specific usage case here at Six Colors.)
You might also look into third-party tools that perform various kinds of cloning and archiving operations, though you need an external drive, a networked Mac with an accessible drive, or a NAS to make use of them:
All of those apps offer flexibility in backup destinations. I’ve used ChronoSync to archive files on one of my virtual Linux servers over SFTP to my Mac, for instance. And you could archive critical files with encryption on cloud-storage systems, too.
I do wish someone would release a Time Capsule successor that was nearly single-purpose and had two or four slots for NVMe M.2 format SSDs. Rather than a highly configurable NAS, just give us something that plugs and goes—and can be upgraded and repaired!
Take Control publisher Joe Kissell offers a deep dive into Mac backups that can help you find and execute the right strategy in Take Control of Backing Up Your Mac. It was updated in December for Tahoe, as well as the latest changes in the hardware and software he covers.
[Got a question for the column? You can email [email protected] or use /glenn in our subscriber-only Discord community.]
2026-01-20 08:56:55
Late last year, a bunch of people, knowing that I love e-readers, asked me if I was going to try the Xteink X4, a $69 tiny, no-frills reader. Like, seriously tiny—so small that it comes with a stick-on magnet ring and will just snuggle into the MagSafe area on an iPhone 17 Pro Max.
The moment I saw that the Xteink X4 didn’t have any lighting, I decided I wasn’t going to bother—I spent way too much time clipping book lights to Kindles so I could read in the dark. But in a moment of weakness (possibly fueled by several Hazy IPAs and high altitude) over Thanksgiving, I bought one. It was $69! Why not take a chance on a weird little e-reader?
When the Xteink X4 arrived, I got my answer: I was deeply unimpressed with the hardware, which has two rocker switches on the front as well as a rocker on the side and a power button and a recessed reset button. Look, I love e-reader buttons, but eight is too many. Of course, there was the aforementioned lack of lighting, meaning you need to use it outside or in a well-lit room. Also, the whole premise of this thing seems to have been that it sticks on the back of your iPhone, but that’s not true unless you have absolutely the largest iPhone available.
But I’ll give it this: it was tiny.
The software was the real tragedy, though: It was really bad. Hard to navigate (so many buttons) and, tragically, just bad at being an e-reader. I couldn’t turn off forced justification, another deal-breaker for me. I tossed it on my desk and figured I wouldn’t write about it because why kick this little thing when it’s down? (I did say some unkind things about it on Upgrade, I’ll admit.)
And then a funny thing happened: Dan Moren sent me a message saying:
Finally got my Xteink ironically after hearing you slag it on Upgrade. So I flashed it with the community-made firmware, which by all accounts is better.
I had seen several people report that the Xteink worked better with some community-built firmware, but I hadn’t tried it—mostly because I firmly believe that suggesting that someone buy hardware only to immediately replace its firmware with someone else’s fix is not really the endorsement it sounds like.1
But Dan’s message intrigued me, moreso when he pointed out that I could install the new firmware directly from a web page using Chrome. Loading a web page and clicking a button seemed like a very low-effort way to see what all the fuss is about, so I went ahead and installed CrossPoint reader.
What a difference. The CrossPoint software draws labels next to the four rocker directions on the front of the Xteink x4, so you know what each button does. It parses ePub files properly, offers font and justification control, and will even display the cover art of the book you’re reading. There’s even support for uploading books via Wi-Fi!
Is it as sophisticated as other e-readers? Absolutely not. Most readers are either Android or Android-based; the Xteink X4 is an ESP32-based thing, so it’s incredibly bare bones. But it works.
Do I recommend that people rush out and get the Xteink X4? No, I don’t. It’s fun to mess around with, and if you’re looking for a super-tiny e-reader that you can keep in a pocket or bag and break out in well-lit spaces at the drop of a hat, it will suffice. It won’t sync with other readers or your phone, so consider it the digital equivalent of throwing a paperback in your purse or backpack.
What I am enthusiastic about is the potential for interesting e-readers. Amazon seems comfortable shipping Paperwhites that are boring and featureless, Kobo’s innovation seems to have slowed as well, and everyone seems distracted by the possibility of finding a new market with E Ink-based note-taking devices like the Kindle Scribe.
But there’s still room for weird. The Android-based Boox Palma is shaped like a phone, but it’s an e-reader. At $250, it’s not cheap, and it’s a bit too big, but who’s to say where experimentation with smartphone marks and E Ink screens might lead? And coming from the bottom up are devices like the Xteink X4, with basic software running on ESP32 hardware.
If I were Xteink or any similar hardware developer, I’d be looking hard at giving support to the CrossPoint project and then focusing my efforts on making a device with simpler controls (fewer buttons!). Adding lighting and potentially a touchscreen would make this interesting, too. There are a lot of directions this sort of product could go—so let’s get to experimenting.
In the meantime, I’ve loaded a bunch of DRM-free public domain books from Standard Ebooks on my Xteink X4 and am letting it just hang around the house. I can pick up an Agatha Christie on a sunny afternoon and just enjoy a little bit of reading time. There’s a lot of potential here.
2026-01-20 06:33:44
This week we’re following up on what the Apple-Google AI hookup means and reacting to Apple’s announcement of an oddly shaped creative apps bundle.
2026-01-16 02:54:25
It might be weird to describe myself as an “authentication enthusiast,” but I guess I’ve never claimed to not be weird. I’ve written a whole lot about passwords and passkeys, so it shouldn’t come as a surprise that I’m a big fan of Apple’s Passwords app. It lets you easily store your authentication details, share them with others, and even view the history of changes to your accounts.
Previous to Apple offering features like iCloud Keychain and Password Autofill, I relied on 1Password to store a lot of this information, but in recent years I’ve transitioned in large part to Passwords. But you’ll note I said “largely.” There are still a few things that I use 1Password for and while Apple is generally good about ticking off the lowest hanging fruit and leaving third parties to offer more niche products, I’d argue that authentication and security are important enough to our everyday lives that the Passwords app can afford to take on more responsibility.
So, maybe it’s time for a power user feature. cracks knuckles SSH keys! You know them, you love them. If you don’t know them, you should love them. Like passkeys, SSH keys are credentials that rely on public-key cryptography to simplify connecting to remote servers and computers without the use of passwords.
And before you dismiss this as something that’s just for those of us who enjoy diving into Terminal, lots of services and sites let you use SSH keys, from GitHub to apps like Edovia’s screen-sharing app Screens and many more. Again, like passkeys, their use helps make our lives more secure and more convenient.
Managing these credentials, however, can be a headache. In part because they can be stored or viewed in many places: in your user’s home directory on macOS, synced via iCloud Keychain, in macOS’s Keychain Access app, the command-line ssh-agent tool, and even some third-party apps like, yes, 1Password can handle them.1
A veritable surfeit of solutions. Too many, really. I’d love to be able to have all my keys stored in a user-friendly interface like Passwords, which would hopefully work under the hood with the command-line tools as well as providing a system for more easily using the keys. 1Password seems to provide the best implementation here, where you can set it up to have requests for your key pop up a dialog box where you can use biometrics or your main password to authenticate.
Just as Apple eventually supported (or at least didn’t actively hinder) Touch ID for sudo on the command line, it’d be great to see Passwords embrace SSH key management for those of us who need it. Which, honestly, is all of us.
.jpg){kind=link}