Fyi, some of those photos are much later than 1981. I wrote added the Arabic/Hebrew support to the Star document editor which probably was about 1984 and showed in one of the photos. I was still in college in 1981 so that could not have been the year.
Thank you for all of your comments on this article. Getting so see high-quality images of this foundational artifact of modern computing was a pleasant enough surprise on its own. Getting actual historical commentary from someone who worked on the product is amazing.
You already answered most of the questions I had, but do you remember how it managed multiple character sets? I'm especially curious about how it supported Japanese Kanji, given how many characters that has. Some quick searching suggests maybe it was XCCS[1]? I had no idea that Xerox was so far ahead of their time that they invented a sort of proto-Unicode when many computers didn't even support lowercase letters.
It was XCCS (Xerox Character Code Standard) which is really the predecessor to Unicode. Both initially were 16 but character code spaces (Unicode was later extended) and Unicode developed by many of the same folks in Unison with Apple.
I was at one time on the XCCS committee and involved with Unicode development before it was called Unicode when I was at Both Xerox and Apple.
A lot of the Japanese feathers were developed with Fuji Xerox, namely vertical text support but the typing method virtually keyboard support was done by the international team headed by Joe Becker who was the Architect of all the multilingual stuff.
Albert in his comment preceded my involvement some trust what he says about earlier development over what I remember.
Sadly, the hardware was underpowered. The same hardware was sold for InterLisp workstations but it was so slow as to be frustrating. Like all D machines they could also boot into the Smalltalk and Cedar/Mesa environments.
A problem was that the InterLisp, Smalltalk, and Mesa development was done on the beefier Dolphin and Dorado (an ECL barn-burner) machines and as far as I can tell never tuned for the wimpy machines. I only used them when I tested some custom microcode on them; I did all my work on the faster machines too.
XDE, the Mesa development environment was run on the same hardware as Star. Star’s biggest issue was that it had far too little real memory and was always thrashing. XDE did much better than Star on a small memory footprint and did not feel slow and generally did not get into a thrashing state. The issue was not the language or the OS, Star and XDE ran on the same Pilot OS, but the architecture of the Star software and the amount or simultaneous this it did in its integrated application which increased its memory foot print.
You are perhaps confusing Cedar rather than Mesa which normally ran on the larger machines.
The other machines you mention were far too large and noisy to use in an office or normal work environment. NO ONE for example had a Dorado in their office, it was always in another room.
If you look close one of the photos labeled Star was really XDE. Rather than object Icons it had a process bar (running applications) at the bottom. It’s the photo with only the XDE mail tool window open.
There's a guy, Bill Fisher, who's battled with Xerox Legal for decades to get them to open source the Star code. He says it runs lightning-quick on modern hardware.
The Dorado came along later and played no part in Star, as knucklehead says. Dolphins were referred to as "dog-zeros" because they were so flaky.
Open sourcing would also require open sourcing all the development tools. Also open sourcing the original code while interesting to look at would be impossible to extend (not impossible but very hard).
It would make the most sense to open source BWS (the basic workstation code used for the later version of Star) with all the tools. It had a ton of well written documentation that was all given to universities to use along with XDE.
Here is one interesting manual for Viewpoint (which I called Basic Work Station in other posts) that was the rewritten version of Star that it was open to others writing apps. It’s a good document with examples or the Mesa APIs for usage. This is from 1986.
I don't think I'd ever seen that document. Thank you!
I worked on the Basic Workstation team; my part was the desktop and folders code. As I recall, the whole BWS rewrite was the result of Robert Ayers coming along and trying to rewrite the bottom layers of Star from scratch. The framework he came up with became the basis for BWS.
I was one aisle over on the editor team. Your memory matches mine about how the BWS came about. It was about opening the system up to allow apps to easily be added unlike the original “trait”-based system that was closed by design.
Traits was the object oriented multiple inheritance archetypes layered on top of Mesa that Star was based upon and required a static analysis step of all objects in the system to optimize object layouts—this is why it was a closed system. After the changeover to BWS, a few years later, only the document editor, and all object types represented in documents (graphics, tables, equations…) continued to use traits.
What always gets me is how much these old GUIs got out of a simple black-on-white, 1 bit per pixel, modest resolution display (1024 pixels wide). That was so unimaginably luxurious back then, and it did look good, thanks to careful tweaking at the pixel level.
Interesting though they might be, these low quality static images give little away about the actual state of the software. How much of what is shown represents real software functionality, how much of it is heavily stubbed software (i.e. implies more functionality than written), how much are contemporaneous mock-ups (i.e. pixel art), and how much of it is not even contemporaneous?
Images 02 and 08 show slightly different desktop icon arrangements but utterly identical application states. Do the applications actually work and can they be used to create the documents shown? Or are they mock-ups to demonstrate possible functionality?
Image 05 appears to be, at best, a collage of images and not a user interface. The apparent differences in pixel pitch between the different elements looks strange and suspicious.
Image 10 is an entirely different operating system.
The first 12 images are all real working software but from different times and systems (XDE, Star and possible Interlisp.)
As an FYI, one actual shows an overlapping window in Star, actually a property sheet. Most people didn’t know that Star support overlapping windows but hardly used them and users could not freely move the windows anywhere. This was a usability decision not a technical one. Many years later that restriction was removed.
The one that looks like the Lisa is the only one I, confused about.
Mind the substitute system font. This (img. 5) isn't just a screenshot from a real Lisa, but either a handmade drawing, or maybe some kind of emulation (which I deem rather unlikely), or a prototype image leaking from Apple.
Perhaps it was a mockup made to show the greater real estate picture of the Xerox Star screen (and maybe capability also – given the "high resolution" car) compared to the Lisa one? Like 1024×808 vs 720x364?
Your suspicions are (almost entirely) unfounded, except for 10, which is XDE, as knucklehead says).
The software was real and worked, for some value of "worked." It was very slow, though. Dave Smith, who demo'ed it at NCC, said he had to be careful to only do things that worked.
As for 02 and 08: that is a real table, which I also used for Records Processing, and was extremely slow. I talk about that extensively in Inventing the Future.
There are many academic papers published about the software, including the graphics.
It was also not slower than the Lisa which came out later. I remember going to a Lisa demo at UCLA and it was super well scripted as someone would click on the screen while talking about something and the 20 seconds later start talking about what they clicked on so that the computer and the demoed were in sync. It was a master class in deception about speed.
> The software was real and worked, for some value of "worked."
What does "worked" mean? The screenshots imply a LOT of functionality. Can it take a two-dimension table of data and generate a stacked bar chart? Can the word processor interleave and flow a varying number of columns of text correctly? Are there any useful tools for doing any of these things within a graphical interface?
I'm not saying that basic tools don't exist, but I've not seen any screenshots showing typeface options, layout options, etc. I haven't seen anything that confirms that these documents could be made visually rather than using a markup syntax.
If the Star really can do all these things, it's deserving of some serious exposure in the form of documentary film. These polaroids really aren't doing it justice.
I would have been amazed for the time. I had an Atari 600XL a few years later and it was amazing to a 14 year-old. The Xerox Star 8010 would have been like something from 100 years in the future to me.
An interesting detail is the substitute Lisa system font.
(I'd conclude, this is either a handmade drawing, or a screenshot of a prototype before Apple settled on the final font. Maybe used as a comparison showing off the use of bitmapped images on the Star, something not available on the Lisa? And/or, as indicated by another comment, comparing screen sizes?)
Also, mind the slight angle at the top edge, where the screen content and the monitor meet. I guess, this is not a screenshot but a montage.
The screenshot seems to be based on one from a Lisa brochure here: https://archive.org/details/AppleLisaBrochure/page/n12/mode/... (page 13 if link doesn't go to the right page, the LisaProject example). Seems to be some sort of remake as beyond the brochure example being very low res things like text spacing are different (circles cutting off bits of text inside them is obvious one on the Alto shot).
The Lisa was in development at the time, so apparently someone at Apple shared it. Strange thing to show off in your Xerox demo! (showing how much larger the screen was is my guess)
Lots of Lisa history on that. The only real thing if I remember correctly that Lisa copied from Star was the desktop itself after viewing a the NCC.
Software architecture wise however the two systems shared many concepts (but not tools or languages) that probably are traced back to the same roots at Xerox but otherwise were independently developed. For example Apple used Clascal (object oriented pascal) while that added an object oriented layer to Mesa. (Mesa itself was pretty much cloned as Modula 2 later if anyone wants to know more about it.)
Mesa was really a great language that was really only missing first class support for Object rather than the way it was takes on which made star a closed system. It had to do with the ‘trait’ system and the static analysis that had to be done to determine the optimal layout or class objects. It was also overly complex because of multiple inheritance. The layout was important because object size directly related to filed document sizes because the filled document was just a relocatable memory mapped image! Basically a filed heap.
Star was pretty much rewritten latter to open it up using the Basic Workstation Software although the document editor itself still used the trait system. But even then the one feature that ‘required’ multiple inheritance was rewritten to not require it (it was the tables feature).
Easy, specially if you compare his evolution as language designer in pursuit of minimalism, from his point of view both Xerox systems were too complex.
You can get hold of his opinion on this ACM session.
As nobody person in the world of computing, I don't agree with him though, I would rather have Active Oberon[0] than Oberon-07, his final language in pursuit of minimalism.
[0] - He did not took part in Active Oberon design and evolution, and it is closer to Modula-3 in features.
The image of the Lisa desktop was not running on Star. Star was software, the hardware ran many different Xerox systems. It was running under some other Xerox system/OS. Hard to say what it was. Could have been XDE (unlikely because I don’t see process bar at the bottom) or Interlisp. But defiantly not running under Star.
Almost certainly not real, or at the very least not contemporaneous. The graphics shown on screen have varied pixel pitch which isn't consistent with the Star's monochrome display. Or, for that matter, consistent with the non-square pixels you would expect from a Lisa screenshot.
I was wondering why that window has a menubar at the top! I even went through the other screenshots to see if it appeared in other places--made me wonder if pull-down menus were actually also invented at PARC.
is Jeffrey Smith, the son of Dave Smith, the inventor of icons and one of the designers of the GUI. He's playing MazeWar on a real, working Alto (I have the raw photo in the book). There's a guy in the Seattle area who has two of them in his basement.
The family story is that Dave's wife was pregnant and overdue with Jeffrey, and they dropped in on PARC and looked at MazeWar. The little eyeball gave her such a shock that she went into labor. So Jeffrey was eager to reenact his origin story.
I love the evolution of the document icon! You can see what the designer(s) thought and the different reasons why each one might have been rejected and replaced in turn. I've put together a few "evolution" graphics like this for clients to show how we've iterated to create logos. I would love to see the designers' processes for other ubiquitous things like this.
These could have been 4x5 polaroids, they had pretty good quality, as good if not better than any 35mm or medium format print.
As proof point, there was a type (50) where you could get a negative as well as a print. I don't recall being concerned about the quality of the negatives, the grain was pretty small. The only downside was the exposure to get a good negative wasn't one where you got a good print. So if it could get a good negative (2000dpi) for enlarging, it can make an awesome print (300-400dpi).
They are very well done. We used a 4x5 with a Polaroid back in a college Physics of Photography class (which was a blast). The creamy, grain-less images made going back to enlarging Tri-X even more painful.
I'm here for this. I didn't even know B&W Polaroid film existed and I guess that's part of the reason it looks so good. It's not your uncle's pictures of Vietnam from 1975 that became all washed out with time.
Very likely they used the larger than regular sized Polaroid such as that found in the 95 and 110 series cameras. Size would have been 3.25 by 4.25 inches. The b&w film would give a very finely detailed image.
For a good primer on how/why the user interface was the way it was you can read this Byte article. These images are from the one of the early Star releases.
Is that the same optical mouse you could find in Xerox-Olivetti 8088 PCs? What the pictures do not show is the raster pattern mouse pad it required. Basically same tech as in ordinary mice but the optical encoders looked directly at the surface, not at grated discs rotated by a rubber ball.
I much later found again the same mouse pad and started using it for nostalgia reasons. Sadly the mouse itself, IIRC, was not directly compatible with standard serial mice.
My dad worked almost his whole career as a Xerox (and later outsourced) service tech and I remember having all kinds of cool stuff discarded by them at home. Not many kids had their 486 full of SCSI discs.
It originally shipped with a mechanical mouse. Later on Xerox changed to an optical mouse with crappy cheap buttons. It did rely on alternating patterns but blue jeans worked well. :-)
No, it had a rubber ball in it at the introduction.
I even have a paragraph or two in Inventing the Future about how we brought a box of spare mice to the NCC, since the balls would get dirty from all the people handling them.
The tools is not what you think it is. IP here stood for InterPress. It’s capturing an image from the desktop as an Interpress file, the predecessor (again) to PostScript with the same imaging model as PostScript. It’s what Dorris used to print out here creations within XDE.
The documentation is on page 154 of this tools document:
The Bitmap To Printer Tool.
Purpose: To capture in an Interpress file a rectangular portion of a bitmap. Sources include the display in the same volume or debugged volume.
Window: The tool has 3 subwindows. The first (top) is a msgSW where the tool says some things. The second has to do with the Capture Phase, the third with the Layout Phase and the fourth with the Send Phase of the tool's operation
Capture Phase - specify a bitmap rectangle:”
What most everyone here missed, just like Steve Jobs when he say Smalltalk, was the entire network of computers that were released with the Star System they sold a separate network print server, file server, mail server… the were all separate Machines connected over Ethernet. When I started Xerox is ‘82 they had over 5000 computers interconnected across the world (and you could debug any computer on the network remotely and it would pull source code/symbols remotely as well for debugging!)
Apple did not do any of that when they released the Lisa or the Mac. But you also have to remember this was target for large businesses and not for small businesses, like the Lisa, or home users like the Mac.
Bill Atkinson was the principal designer and developer of the graphical user interface (GUI) of the Apple Lisa and Apple Macintosh. He was the creator of the ground-breaking MacPaint application and its novel interaction techniques, and he also designed and implemented QuickDraw, the fundamental toolbox that the Lisa and Macintosh used for graphics. He also conceived, designed and implemented HyperCard, the first popular hypermedia system. He invented many now-common interaction techniques, including the menu bar, the marching ants selection used in painting programs, the Selection lasso, FatBits, pull-off menus (where a menu becomes a palette), and many others.
Recommended Readings:
Bill Atkinson, "Joining Apple Computer", Facebook post, April 27, 2018. Facebook, or local pdf:
How many of those Android users are "Android developers are running Android Studio in Android, or office clerks are editing their documents on Android."?
probably not many? but even if the answer were literally zero it would still be the case that most people use android now instead of desktop operating systems, just not everybody
they just don't run android studio or work as office clerks
It always annoys me how people say Apple invented the modern GUI when stuff like this has existed since the late seventies. Xerox were way too ahead of the time with projects like the Alto.
> It always annoys me how people say Apple invented the modern GUI when stuff like this has existed
I suggest watching a video of how the Star worked and one of how the Lisa worked and then comparing the two to the way Windows, Macs, and Linux work today.
The Star paradigm of mixing the use of function keys on the keyboard with mouse movements is definitely not the way things work in the modern GUI.
For example, moving an icon on the desktop:
Star: Click icon you want to move with mouse. Press move key on keyboard. Click destination location with mouse.
No one knew the right way to do stuff. Star had a noun/verb model of doing actions that all objects you could select responded to in the same way. It did not use control key combinations. It was actually much easier to learn and more consistent than the Lisa or Mac. But it did require a unique keyboard.
It also used a two button mouse (3rd if you coorded but Star did not use that, XDE did). Mac Lisa had one button but a second button was simulated with a control click. I found the Max model much harder for text selection over large area with click and drag a PITA. The two button metaphor of click and extend the click with a second button click much easier in practice for most editing. But both Xerox and Apple’s systems were well done and consistent.
By the way the Star keyboard was excellent. The tactile feel was awesome and much like a Selectric typewriter with a slight curve to the keyboard.
