Chasing the Unified Messaging Dream

Microsoft gave it a try in 1999 and failed. The same tantalizing possibility returns every few years: a single place to communicate rather than an ever-expanding cacophony of apps, each with its own quirks. Are we any closer to this hope a quarter century later?

Microsoft’s attempt with the launch of MSN Messenger is hardly a footnote at this point. But, I still remember the promise at launch: a single application that could communicate with one’s Microsoft contacts and those on the dominant AOL Instant Messenger (AIM) platform. AOL was not amused and the effort fizzled out relatively quickly.

During my days using Linux on the desktop, I landed on another attempt at the same thing: Gaim (now Pidgin). It promised to bring together the wide variety of messaging platforms available at the time, so if some of your contacts used ICQ and some used AIM and some used Yahoo’s messenger, you could reach them all.

In the mid-2000’s, I abandoned my search as everyone I knew had discarded “instant messaging” for increasingly affordable and common SMS texting and web-based tools like early forms of Facebook’s Messenger.

This was largely an improvement, as neither required both users to be online at the same time in order to work. But neither were amenable to a universal communications tool. Facebook’s messages were within a single website; texts were bound to a specific phone.

History does have a way of repeating itself, though, and over time new platforms emerged that ended up feeling something like a split between the late 90’s instant messaging and SMS. Smartphones accelerated the fusion with tools like the separate Messenger app ballooning Facebook’s messaging presence and Apple’s iMessage moving its users transparently off of the aging SMS system to a modern, secure system without any fuss.

If iMessage were available on non-Apple platforms, I’d consider myself set. The so-called “blue bubbles” have everything one might want in a modern messaging tool. Every message is end-to-end encrypted and verifiably so. Messages have delivery confirmation, conversations show when someone is writing, group messages work reliably, and everything synchronizes securely between all of your Apple devices. It’s about perfect.

Except it only works for the half or so of folks I communicate with that use Apple products. I’ve said before both that I thought Apple should make iMessage cross-platform and that they have the right not to. Unfortunately, they’ve leaned into that right, missing a fantastic opportunity for expansion.

Meta (nee Facebook) remains a giant in the field of messaging, but it too is less than universal. Messenger will run most anywhere, but I know plenty of people who are justifiably untrusting of Meta and thus unreachable on its platform.

My church launched during the pandemic and electronic communication has remained an essential part of its functioning, so as it grew and gained some non-Facebook and non-Apple users amongst the family, I went looking for a good alternative to cover everyone. Texting could work, but group messaging on SMS is unreliable. We landed on the excellent, free and open source Signal.

Signal has been a massive blessing for us. Signal offers the near-universal platform support of Messenger and pairs it with the trustworthiness of Apple. Most (though not all) of the church agreed to sign up. Signal is easy to use, supports reliable messaging groups, and has far better security than our purposes require.

Yet, there’s a catch. First, with each of these options I’ve laid out, some people are on one, but not another. There are still people I can only reach by iMessage or SMS text, so I almost always have Apple’s Messages open. There are folks who the only reliable way to reach them remains Meta’s Messenger — it too is frequently open for me, though meaningful messages come in just infrequently enough I can forget to check it. And then there’s Signal, where more of my pressing communication happens today than the others or e-mail. The fragmentation is worse than it was years ago.

Second, I prefer to correspond on my laptop or iPad with a keyboard connected. Call me old fashioned, but typing on a phone’s onscreen keyboard when I can type faster and be more productive on a computer just isn’t ideal. All three platforms work on the computer, but Signal’s desktop client is objectively inferior. Unlike both Apple’s Messages and Meta’s Messenger, Signal’s desktop app is built on the contemptible Electron.

Like all such apps, Signal cosplays as desktop software, but misses the mark. Not every aspect of the interface acts as it “should” if it were a real Windows or Mac app. Electron apps are a kludge, running a full web browser in the background, hogging your computer’s resources and accelerating battery drain. To add insult to injury, when it is closed, it won’t notify you if new messages come in.

It’s a subpar experience.

Signal can’t fix the fragmentation, but they could at least solve the other problem for Mac users. They have a fine iPad app and only Signal’s stubborn refusal prevents it from running on the Mac, not anything technical. The organization could click one button and immediately give the world an almost native Mac app with properly working notifications. But, the project believes all Windows, Mac and Linux users should have the same experience, no matter how inferior.

Third, Signal suffers from one of its organization’s greatest strengths: it is a security-conscious non-profit. In a combination of the utmost safety for end-to-end encryption and the limitations of how much data a non-profit can afford to store, Signal does not offer a way to synchronize your past messages between devices.

One can do end-to-end encryption of messages while also keeping them in sync by storing the fully encrypted data on the provider’s server. That’s what Apple does (albeit in a painfully slow way) and what Meta has begun doing as well (surprisingly more efficiently). I wish Signal offered a plan to do the same, because if you upgrade to a new device (or don’t use one for more than a few weeks and then return), there’s no easy way to get the missing messages onto it.

Signal gets a job done for me, but these shortcomings, and the overall fragmentation of contacts on other systems, leaves me wanting something better.

The company best known for being the commercial backer of WordPress, Automattic, recently acquired several tools aiming to create that “better.”The premier one the others are merging into is Beeper (the second Eric Migicovsky-founded project I’ve written about in as many weeks).

Beeper is best known for being the app that claimed to bring Apple’s iMessage to Android. This was ill-advised, as Migicovsky’s initiative depended on tricking Apple’s servers into creating accounts for, and storing all the data for, users who weren’t Apple’s customers. That ethically dubious stunt did a disservice to his broader project, a significant part of which is open source and had a bigger goal: full, unified messaging.

Under its new corporate steward, Beeper can connect to most platforms (outside iMessage) and synchronize those messages between your devices and computers via one, familiar looking interface. It’s a great concept.

This is a journey through the valley of caveats, though, and Beeper’s weakness is a huge one. In an era when most messaging services utilize end-to-end encryption, Beeper breaks that to work. Instead of the encrypted messages arriving on your computer or phone and being decrypted only by you, the Beeper service inserts itself in the middle, decrypting messages (with your permission, of course) and repackaging them for the Beeper software to use.

That’s a really bad idea. I have no reason to distrust Automattic particularly, but the whole point of the move to greater encryption is that I don’t have to trust anyone — Signal, Apple, Meta or anyone else — to handle my data carefully because they can’t read it. Beeper asks users to trust them with your data (and, by extension, ever more dystopian governments that might legally compel Automattic to let them take a gander).

I have zero interest in doing this and neither should you.

To their credit, Beeper is upfront about their system’s weakness and they offer a possible solution: users can run that “bridge” themselves, using the open source server they provide. This is obviously for those with a fair amount of tech-savvy, but it is a valid solution for those able to set it up. If I have a server that retains my data, I can sync and connect securely without concern because it is my server, not Automattic’s or anyone else’s.

With some more effort, I may get Beeper’s backend working, but my story ends elsewhere for now. My Beeper “bridge-manager” attempts were felled by the system’s two-factor authentication code. Like so many providers, Beeper sends a code by e-mail to log in. And the backend, even when self-hosted, must login to Automattic’s server. The login messages would always arrive for me after they expired.

That glitch accompanied Beeper’s continuation of one of my Signal complaints: it used the contemptible Electron to create a non-native desktop application. So, for now, I moved on.

But Beeper got me a good part of the way there. They’ve actively sponsored and participated in the Matrix open-source community. Matrix’s system is the infrastructure behind Beeper, providing an end-to-end encrypted messaging system one can self-host without dependence on Beeper’s server or anyone else’s.

Because Beeper is built upon the Matrix design, the software the company has collaborated on developing with the broader Matrix community benefits regular, Beeper-less Matrix as well.

Matrix’s primary developer, Element, also does what Signal ought to do and allows their app for iPhone and iPad to also run on Mac. They have one of those annoying Electron apps, but you can skip right past it. The iOS/iPad OS app, Element X, is actually beautiful and, unlike Signal or Messenger, hardly draws any resources when running on my computer.

It is not perfect. You have to “chat” with a chatbot you set up in order to start conversations with a new contact on, say, Messenger. The setup of the server that relays messages also took me several hours to configure and merits a column of its own, assuming it keeps working. But it has nudged me towards that seemingly unreachable goal: a nearly unified messaging tool, lacking only Apple’s side of the equation.

Maybe that’s a taste of a future where we will benefit from using Matrix directly. Like the fediverse of social media and e-mail, Matrix servers talk to one another, so Matrix users on other servers can contact you on your own. It could be a better messaging platform than any of the currently popular ones. Not yet, but maybe.

Maybe it’s a taste of what Beeper will be. After all, Automattic is continuing to improve its self-hosting tools and has promised on-device decryption, which would keep end-to-end encryption fully intact. With its Matrix foundation, that would boost the underlying network, too.

Either way, it points to something we should hope for: messaging where it doesn’t take managing umpteen different apps just to reach our close friends, family and business associates. A quarter century is enough time to solve that.

Full Disclosure: Tim does own some Apple (AAPL) and Microsoft (MSFT) stock.