I used a Blit successor (AT&T 5620 nee DMD) in the 1985ish time frame. It was attached to an AT&T 3B2 host running SVr3 and later SVr4. It was a very cool, minimal, useful and very Unixy way of doing things. The "window manager" software was called "Layers", and allowed a lot of application processing to happen on the terminal. Layers windows behaved (mostly) just like ttys multiplexed back to the host. There was a pretty decent cross development environment for creating little apps, including a C compiler and debugger. There were the normal doodads like a graphical clock, drawing program (pic?), calculator. There was also a pretty neat mouse-enables editor called 'jim' and I seem to remember a sort of troff previewer. Pretty much the coolest thing I'd ever used up until I got a Sun 3/160 in '86 or so running SunOS 3.x and OpenWindows. Game over.
Pretty much the same story for me. When I first joined the Labs, I got a 5630 (the successor to the 5620), also called the 630 MTG. At the time, it was an impressive piece of equipment that was a huge improvement over a regular tty.
I found the video to be somewhat lacking in interesting technical details -- there really were a lot of neat things going on in the background to make it all happen.
Readers waxing nostalgic might enjoy Rob Pike's paper describing the 5620. It provides an in-depth look at what was going on behind the scenes.
Interesting...I've never heard the 630 called a 5630. Very cool. I actually played a little with a 730X (successor to 630 with ethernet and an X server in ROM) back in the dark ages. Layers was pretty rotted by then and it was a terrible XTerm, especially WRT to the hotrods NCD was cranking out. Kinda sad.
Yet another Tech Taking Over The World^TM that didn't. I've got some bubble memory and a Symbolics around here someplace.
I notice that on the Blit, new windows are created by "sweeping", i.e., dragging with the mouse from one corner of where the new window will go to the opposite corner. To this day, that is how new windows are created on the Plan 9 operating system.
(Specifically, you indicate using a command on a menu that you wish to create a new window. Instead of a new window appearing, like it would on any other OS, the cursor changes shape. The shape change is your cue that you can now "sweep", and the new window is created where you sweep, i.e., in such a location and with such dimensions that your sweep is a diagonal of the new window.)
And here is an opportunity to point out my pet hypothesis as to why no one uses Plan 9: namely, the refusal by the leaders of the Plan 9 project to adapt to what later became standard ways of doing things.
Although the Blit preceded the Mac, the Mac became orders of magnitude more widely used. And when Microsoft created an operating system with a GUI, they made opening a new window work like it does on the Mac. Once hundreds of millions of computer users had become used to the way one makes a new window on a Mac, Plan 9 should have changed to the Mac way. This is just basic hospitality. If you run a hotel and the controls for the hot and cold water work differently than the way they work in 99% of the homes of your guests, you hire plumbers to change the controls to work the way your guests expect them to work.
For some reason, the Plan 9 folks, and Rob Pike in particular, never saw the need for being hospitable (maybe because the way his software let the user create windows was invented first, damn it!)
Emacs has a bit of the same problem (inhospitableness, e.g., having unusual names for things, e.g., calling windows frames and calling panes windows), but Plan 9 is the most inhospitable software I've ever encountered.
Deleting text by highlighting an extent of text (e.g., by dragging the mouse), then hitting backspace is another example where Plan 9 behaves differently than every other OS I know about. (Particularly, it causes the highlighted extent to be deleted, like other OSes do, but then also deletes one additional character, which is unusual behavior that a user has to learn -- and then unlearn every time he or she switches back to a "hospitable" OS.)
In contrast, the designers of Gnome and KDE, for all their faults (e.g., being too eager to re-invent things) understood and practiced basic hospitality towards the user.
Setting aside the misdirected sense of entitlement, the primary problem with your post is the assumption that every piece of software is required by some unwritten contract to fulfil several goals, to wit:
- it must look like whatever you're already using
- it must aspire to the maximum possible user base
- it must adapt to latest technological fashions
while these are extremely valid requirements for a commercial product, plan 9 is very much not one. it is a research operating system, and no new knowledge would be gleaned by re-re-reimplementing PARC products.
your definition of "hospitality" seems to involve the willing acceptance of UI trends that some people regard to be disgusting garbage. why is it not sufficient to let people who prefer plan 9's interface use it? why do you make the assumption that user acquisition was in any way an intended result of this operating system's development?
flip your hypothesis over. why has nobody written an interface for plan 9 that meets your expectations? the answers to this question may help you understand why nobody who uses plan 9 cares that it doesn't work like an apple product from 1984.
Plan 9 performs work and enables communication over the internet. It is possible that installing any operating system other than the ones you currently use would be a waste of your time, in case they require you to learn them.
