Login
You're viewing the assemblag.es public feed.
  • Oct 10, 2026, 10:00 AM

    @netbsd YES IT DID ❤️❤️❤️

    Enthusiastically happy caps, if that wasn’t clear :)

    I worked on the service component, and also handled all of the service, os, and firmware builds. I sat next to the Bluetooth lead, a close friend, who was working on the port.

    I remember clearly him showing me the first build that booting to a shell, and running `top`.

    Prior to that, it was 100% homegrown, which was huge in its own right. And then this, which was the smallest form factor I’d ever seen a UNIX running on. It was amazing.

    And that was an amazing team.

    💬 2🔄 6⭐ 20

Replies

  • Oct 10, 2026, 10:11 AM

    @netbsd Also many folks who worked on that effort, directly or indirectly, are on fedi. They’re still doing fantastic work.

    💬 0🔄 0⭐ 6
  • 💬 1🔄 0⭐ 0
  • Oct 10, 2026, 11:00 AM

    @ltning @netbsd Sadly, in very limited capacity. They will still boot, and later editions_may_ connect to towers on bands and modulations they support.

    Much of what made them awesome, though, was the tight integration between the cloud and the applications which allowed near-realtime sync of application data, and service-based session proxying which made for effectively always-connected devices at a time when “cloud” wasn’t a thing and connections were spotty.

    There was even an awesome ssh client and accompanying proxy (thanks to @robey [I hope you don’t mind the mention, robey!]) which held sockets and sessions on the backend and meant that one didn’t lose their shell sessions when stuck in the Caltrain tunnel.

    That service doesn’t exist anymore, and being the first of its kind, many of the critical protocols were bespoke and aren’t in use today.

    But, whatever software exists on a unit that boots, still works. Including the first video game I played that actually had great stereo sound design, a platformer called “Bob”.

    💬 1🔄 0⭐ 0
  • Oct 10, 2026, 11:14 AM

    @ltning @netbsd @robey But with all of it said, I think it’s difficult to overstate what an inspired accomplishment the whole thing was. There are others here that can provide a far more detailed history than I can; I joined the company right after the first black-and-white versions shipped.

    I played a really minor role in all of this.

    People here, far smarter than me, invented both the backend, the custom VM and toolchain that allowed developers to write applications in Java, and compiled that down to a highly optimized bytecode that could run performantly on the handheld devices of the time (and went on to inspire Dalvik on Android).

    The synchronized datastore, intrinsic to the OS, allowed cloud<->device synching that, performance relative to task complexity, rivals what we see today. Even the UI rendering engine squeezed an incredible amount of expressiveness into a very small space.

    The web browser mechanism was really neat as well, transcoding web pages and images, **and javascript** on the backend into something that could efficiently be transferred and rendered on the device given the network conditions of the time. Blew my mind.

    Just an incredibly talented, and incredibly kind, group of people. I’m incredibly grateful for that experience.

    💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 10, 2026, 12:36 PM

    @netbsd @ltning @robey Could be. I believe we only ever shipped to Europe with DT in Germany, and Telefonica in Spain (I wanna say there was a smaller French presence as well, but memory is failing me). I _think_ even then Europe had relatively reasonable band standardization, so may be possible, if old Nokia GSM handset can connect, that a Hiptop could establish emergency voice CX.

    Data would be 0% joy though.

    There are multiple levels of authentication and handshake, especially when GPRS comes into play. When one starts a sidekick, and this was true for every model with any type of radio, there was a “dot” sequence one could observe on the connectivity status… stretching my memory, one dot was having a connection to the cell tower, two was establishing a PDP context (or whatever that was called across standards… basically, digital services were allowed), 3 dots was an actual routable TCP connection, followed by “G” which meant connection to the Danger service was established, sessions were synched, spice was flowing).

    To get on the cell net, the carrier had to know that the IMSI/MSISDN was allowed, which required a lookup on a service we hosted. Further, each device was hardcoded to talk to a specific, per-carrier service, by address and key pair, and the service proxied all inet comms.

    💬 0🔄 0⭐ 1