This is one of my favorite papers ever. Cedar is a descendant of Mesa, and Mesa was a source of inspiration to Niklaus Wirth for Modula and, consequently, Oberon. You can see the similarities in Cedar’s UI to Oberon’s.
Also, why is it in 1984 they had a workstation with a 1024x808 display and yet I had a midrange laptops were stuck at 1024x768 until like 2016??
I’m legitimately curious if there is any non-Xerox parallel GUI evolution. Even the stuff mope divergent from the Macintosh like Lisp Machines takes after Xerox, even in little ways like scrollbars. (Athena/X11 scrolling? Xerox.)
I think the closest to that title would be the Pointer Environment on the QL, which seems like someone was described a GUI, but had never used one before.
Honestly, Oberon is pretty different from the GUIs developed at PARC. It was definitely inspired by Mesa and it took the tiling window management and “tracks” idea from Cedar, but it has no icons, no direct manipulation, no overlapping windows (neither did Cedar really, from what I understand), etc. etc.
But yeah, the entire world of GUIs is divided into pre-PARC (The Mother of All Demos, specific applications, etc) and post-PARC.
I suppose one could argue that non-WIMP GUIs are distinct enough from PARC-inspired UIs that they could be considered distinct?
What would you define as non-WIMP? Maybe stuff like PDAs, but the Newton has much of the hallmarks of a windowed interface, and Palm OS is borderline a small Mac OS for touchscreens. And iOS still uses a variant of Palm OS’ home screen.
Home screen aside, I don’t think iOS owes much to the Newton. Its “direct manipulation” style is taken pretty directly from existing multi-touch research concepts, most famously Jeff Han’s 2006 demo. Magic Cap was an early example of direct manipulation. You might also be interested in ZUI’s such as the Kansas networked collaborative environment (“a big flat space”, sorry for the paywall!). My favorite left-field example might be Raskin’s Archy system, although it’s hard to say how far that really got.
I’m sure you’re already aware, but if not, Raskin’s Archy system was preceded by the Canon Cat, which had a lot of the same features in simpler form. The Canon Cat actually got produced and it was pretty neat.
(I got >< this close to getting a Canon Cat of eBay. Somebody swooped in and got it for like $1 more than I had bid. Humph.)
Raskin’s book The Humane Interface has most of the good ideas in that system, though. His son Aza was doing some interesting stuff before he went off to Mozilla and was never heard from again… (j/k, I just stopped paying attention)
You’d better not have bought it off of eBay circa 2017, or I’m gonna be mad. ;)
The Humane Interface is a good book (I read it cover-to-cover and then loaned it to a friend of mine and never got it back and he moved to Oregon…).
THI had some problems, IMHO, with its core concept of metamodes and keystroke algebra (I don’t remember the exact terms because, again, my book is MIA). The fundamental argument in favor of its metamodes relies on keystrokes/half-keystrokes and so on, but has an underlying assumption that all keystrokes are of equal cost…which isn’t true. The human hand can’t hold down certain keys and hit other keys easily. It’s why, obviously, the LEAP keys on the Cat are where they are but I would imagine that even then certain combinations of keys would be a pain.
I also feel like the Cat/Archy paradigm lacks discoverability that is afforded with a more traditional WIMPy UI…
That being said, the concept of the metamode itself is definitely interesting and worthwhile and we have kinda abandoned it.
I feel the biggest flaw of the Raskin system is that it really doesn’t have good answers for anything more than users doing word processing. All the metaphors feel in service to editing text, but almost nothing else.
I’m not sure it’s fair to call that a flaw, since it was a pretty deliberate design decision, as least as far as the Cat was concerned. But you could say the same for emacs!
The Cat handled the discoverability issue by having a small and fixed number of functions, the names of which were printed on the front of the keys. One of them brings up the manual while you hold it. Overall I think it’s way more discoverable than vi or emacs (or any of the keyboard-centric window managers), but I suppose that’s a low bar. It’s definitely not anything close to as extensible, although you can get it to execute Forth.
If there’s a market for these ideas any more, I’d expect it to be somewhere in the neighborhood of DIY mechanical keyboard enthusiasts (who seem to be reinventing chord keyboards) and those who like keyboard-driven editors, WMs, and browsers. Such “expert” users seem pretty OK with sacrificing discoverability for quick access; indeed I’d claim they’ve sort of over-accommodated. Meanwhile, regular folks are lucky to even have a keyboard on their device at all nowadays.
Cats don’t show up on eBay very often, but if you have an Apple IIe, there’s this guy who makes replica SwyftCards…
This is one of my favorite papers ever. Cedar is a descendant of Mesa, and Mesa was a source of inspiration to Niklaus Wirth for Modula and, consequently, Oberon. You can see the similarities in Cedar’s UI to Oberon’s.
Also, why is it in 1984 they had a workstation with a 1024x808 display and yet I had a midrange laptops were stuck at 1024x768 until like 2016??
I’m legitimately curious if there is any non-Xerox parallel GUI evolution. Even the stuff mope divergent from the Macintosh like Lisp Machines takes after Xerox, even in little ways like scrollbars. (Athena/X11 scrolling? Xerox.)
I think the closest to that title would be the Pointer Environment on the QL, which seems like someone was described a GUI, but had never used one before.
Honestly, Oberon is pretty different from the GUIs developed at PARC. It was definitely inspired by Mesa and it took the tiling window management and “tracks” idea from Cedar, but it has no icons, no direct manipulation, no overlapping windows (neither did Cedar really, from what I understand), etc. etc.
But yeah, the entire world of GUIs is divided into pre-PARC (The Mother of All Demos, specific applications, etc) and post-PARC.
I suppose one could argue that non-WIMP GUIs are distinct enough from PARC-inspired UIs that they could be considered distinct?
What would you define as non-WIMP? Maybe stuff like PDAs, but the Newton has much of the hallmarks of a windowed interface, and Palm OS is borderline a small Mac OS for touchscreens. And iOS still uses a variant of Palm OS’ home screen.
Maybe Magic Cap?
Home screen aside, I don’t think iOS owes much to the Newton. Its “direct manipulation” style is taken pretty directly from existing multi-touch research concepts, most famously Jeff Han’s 2006 demo. Magic Cap was an early example of direct manipulation. You might also be interested in ZUI’s such as the Kansas networked collaborative environment (“a big flat space”, sorry for the paywall!). My favorite left-field example might be Raskin’s Archy system, although it’s hard to say how far that really got.
I’m sure you’re already aware, but if not, Raskin’s Archy system was preceded by the Canon Cat, which had a lot of the same features in simpler form. The Canon Cat actually got produced and it was pretty neat.
(I got >< this close to getting a Canon Cat of eBay. Somebody swooped in and got it for like $1 more than I had bid. Humph.)
I’m the proud owner of a Canon Cat. :-)
Raskin’s book The Humane Interface has most of the good ideas in that system, though. His son Aza was doing some interesting stuff before he went off to Mozilla and was never heard from again… (j/k, I just stopped paying attention)
You’d better not have bought it off of eBay circa 2017, or I’m gonna be mad. ;)
The Humane Interface is a good book (I read it cover-to-cover and then loaned it to a friend of mine and never got it back and he moved to Oregon…).
THI had some problems, IMHO, with its core concept of metamodes and keystroke algebra (I don’t remember the exact terms because, again, my book is MIA). The fundamental argument in favor of its metamodes relies on keystrokes/half-keystrokes and so on, but has an underlying assumption that all keystrokes are of equal cost…which isn’t true. The human hand can’t hold down certain keys and hit other keys easily. It’s why, obviously, the LEAP keys on the Cat are where they are but I would imagine that even then certain combinations of keys would be a pain.
I also feel like the Cat/Archy paradigm lacks discoverability that is afforded with a more traditional WIMPy UI…
That being said, the concept of the metamode itself is definitely interesting and worthwhile and we have kinda abandoned it.
I feel the biggest flaw of the Raskin system is that it really doesn’t have good answers for anything more than users doing word processing. All the metaphors feel in service to editing text, but almost nothing else.
I’m not sure it’s fair to call that a flaw, since it was a pretty deliberate design decision, as least as far as the Cat was concerned. But you could say the same for emacs!
The Cat handled the discoverability issue by having a small and fixed number of functions, the names of which were printed on the front of the keys. One of them brings up the manual while you hold it. Overall I think it’s way more discoverable than vi or emacs (or any of the keyboard-centric window managers), but I suppose that’s a low bar. It’s definitely not anything close to as extensible, although you can get it to execute Forth.
If there’s a market for these ideas any more, I’d expect it to be somewhere in the neighborhood of DIY mechanical keyboard enthusiasts (who seem to be reinventing chord keyboards) and those who like keyboard-driven editors, WMs, and browsers. Such “expert” users seem pretty OK with sacrificing discoverability for quick access; indeed I’d claim they’ve sort of over-accommodated. Meanwhile, regular folks are lucky to even have a keyboard on their device at all nowadays.
Cats don’t show up on eBay very often, but if you have an Apple IIe, there’s this guy who makes replica SwyftCards…