Diary of a Disaster: General Magic Goes Poof (2004)
grosen.dk
grosen.dk
- Magic Cap came from the early-1990s expectation that the future involved intelligent agents that would travel around the network, do work (e.g. searching for the best price on a flight or new shoes), and then report back when the user was connected again.
In practice, this meant that you could send email to another Magic Cap user that included executable code, like an animation or a little game, that would run in a sandbox on the user's device.
Even today, we still don't really have this -- you can't stick JavaScript (or a Java applet) in an email and expect the receiver to be able to run it.
- The Motorola Envoy ran on Motorola's low-bitrate, high-power (and now defunct) Ardis wireless network. It was robust enough to work from within an airplane!! Because almost nobody had them (in 1995), the flight attendants would not stop you from sending and receiving email wirelessly while flying. Of course today, even if Ardis still existed and still worked from 30,000 flight, they would be telling you to put your phone into airplane mode...
- I did not have a Motorola Envoy, but I did have a Sony Magic Link that I used for email from 1995-1996. It was amazing! You could leave it by your bed, and when you woke up in the morning, it would make a (free, local) call to AT&T PersonaLink to exchange email. I have fond memories of lying in bed in my PJs in early 1996, having just woken up, and flirting via email with a high-school friend who had already arrived at school at the computer lab. This all seemed like an incredible luxury at the time.
Of course "checking email when you first wake up" became much more common with the end of per-minute fees from major ISPs (AOL ended theirs in December 1996), the rise of laptops and pervasive Wi-Fi (~2001), and eventually the rise of cellular handhelds with email (~2005).
You say that like it's a bad thing?
NB I can see that it might make sense to allow "executable" messages - but they potential for security and/or privacy problems look pretty bad.
I'm not sure running agents is all that more a problem than the JavaScript in a web browser. The agent folks were definitely thinking about security and even CPU usage.
We did have it -- it was called VBScript in Outlook and it was one of the biggest (actually probably the biggest) security disasters in the history of the internet.
Tcl also had several agent extensions back in the 1990's.
I remember they were very proud of the fact that the programming language was fully object oriented, and drank deep of the OO kool-aid. So no C-style direct memory access AFAIR.
"We find that a phone link is actually faster than IR! The poor IR performance is particularly wounding to Woz, who signed away his IR patents and technology to General Magic."
a lot of bug discussion and ignoring Woz
"Wozniak is visibly upset that the product has not been fixed and that the stock market is so gullible. Perhaps be is also miffed that be was not even offered stock for his gift of the IR patents some three years ago."
If they treated Woz they way this article says, then I cannot help but think they were probably going to fail. Its a true shame because I still think agents would be a very interesting way to organize programs even though the remote communication part has been taken over by http-based APIs.
https://en.wikipedia.org/wiki/The_Mythical_Man-Month#The_sec...
(Not to say the original Mac API wasn't awesome in its own way -- you can do a lot of really cool stuff. A hacker's delight, in ways that the original Windows API wasn't.)
One more story to add to the "Woz: a real mensch" pile.
// ok Apple WTF are we doing trying to autocorrect Woz to Wiz?
Isn't that insider trading?