Nelson makes no distinction between the Macintosh GUI and Xerox Parc's Smalltalk environment. He calls them both "the PUI". They are different! Between 8:15 and 8:21, Nelson says: "Xerox Parc took away the right to program". That is absolutely and categorically false. Smalltalk is one of the most powerful end user programming environments ever developed. It was Steve Jobs, not Xerox Parc, who took away the right to program.

I haven't used the Alto, but — were you truly able to reprogram the whole system? Or were there only certain areas that were reprogrammable?

Some fascinating history here that I hadn't heard before. • Doug Englebart wept when Xerox Parc broke up the NLS/Augment team. • "Unix shell scripts, invented by Frenchman Louis Pouzin" -- https://en.wikipedia.org/wiki/Louis_Pouzin

Did PARC folks not have any involvement in the creation of the Star? If so, does that mean the "desktop metaphor" didn't come from PARC after all?

The Alto definitely had a "paper metaphor". It had a high resolution bitmap display, it displayed black text on a white background, you had a choice of fonts. You could mix text and graphics arbitrarily, just like paper. They did this in an era when paper was a much richer visual display medium than computer monitors. They were consciously simulating paper, and Nelson is attacking them for this. But the user interfaces they created using this paper metaphor looked graphically better than anything else at the time. Some PARC people were seconded to SDD for the development of the Star. By my recollection, the "desktop" metaphor, as I understand it, with file and application and folder icons on the deskop, and a trashcan, first appeared in the Star. I could be wrong, but I'd like to see primary sources that prove otherwise.

I was the original author of WordStar 2000, the sequel to the pioneering WordStar product which was a massive success. When visiting the offices of the company, they had purchased a Xerox Star machine, which was $50,000 at the time. Although Xerox had a crack R&D dept., the upper management of Xerox was full of idiots who had no clue how to enter a new business area. Through management incompetence, they failed to take advantage of their huge head start. Commodore was a typewriter company that also failed to understand computers and blew it bad. If you put a bad jockey on the fastest horse you will still lose. Leadership matters more than anything, and Xerox had a long series of bad leaders, just like IBM had after Watson retired. By the way, Prof. Wirth of ETH who invented Pascal, Modula-2 and Oberon took a sabbatical at Xerox Parc, and after that built his own computer called the Lilith machine which was very interesting, and was the closest thing to interchangeable parts and not a strong boundary between OS and applications that Nelson would have liked. Nelson also didn't mention that 3Com which pioneered ethernet also came out of PARC. Xerox has the unfortunate distinction of the company with the most inventions not commercialized due to their stupidity. The management was basically running a gold mine that discovered diamonds and threw them away.

An alternative take on the “who ruined the computer”: The problem was neither taking away or granting full modification access. It was the failure to recognize that both are necessary. We need both barriers and means to securely cross them. From this perspective, everyone on either far side of the debate would be responsible for this deep problem.

@Doug Moen it does seem like PARC was the site for the invention of cut and paste, something Ted Nelson strongly opposes and with very good reason: https://futureofcoding.slack.com/archives/C5U3SEW6A/p1583620045067100?thread_ts=1582854486.016000&cid=C5U3SEW6A

Yes, cut & paste was invented by Larry Tesler at PARC. He later went to Apple to work on the Macintosh. In TextEdit on the Macintosh, I can select some text, then drag & drop the selected text to another location. It seems like this is the kind of UI that Nelson is advocating. But I generally don't use this feature, because it becomes awkward if the text has to travel a long distance, especially if I have to scroll a long distance in the document to find the destination for the text I am moving. PARC-style cut copy & paste is simple, easy and works. People have implemented "better" systems but they always more complicated, and fail to catch on. Can you provide a link where Nelson demonstrates a better UI?

Nope.. It's more than just the UI used to manage the clipboard hidey hole. It's partly that there's just one hidey hole, but even that doesn't quite capture the breadth of his criticism. Ted's asking for the OS to provide a framework for parts of documents that can be given mutable but persistent locations, akin to a desktop or bedroom floor. They can be moved around, they stay in one place unless moved around, and you can see where they moved from. Seen through this lens it's immediately clear to me -- in a way it never was before -- why he cares so much about transclusion.

@Kartik Agaram Yes, I understand your point about "classic" cut and paste, and why transclusion is important. As for tracking the provenance of each clipping, it's not true that "no writer cared" about this in pre-computer days. Consider the Zettelcasten system, which accomplishes exactly this using paper. Or consider the "associative trails" in Vannevar Bush's "Memex" machine (proposed in 1945), which I know was a big influence on Nelson.

Sure. But it's not the same as "hundreds of years" of all writers using it.

@Kartik Agaram said:

It's partly that there's just one hidey hole, but even that doesn't quite capture the breadth of his criticism. Ted's asking for the OS to provide a framework for parts of documents that can be given mutable but persistent locations, akin to a desktop or bedroom floor. They can be moved around, they stay in one place unless moved around, and you can see where they moved from.

I get to enjoy this exact mode of working when I edit video and audio. For instance, here (below) is a screenshot from the Orca podcast episode. At the bottom, in the clip editor, you see the full source audio recording. In the middle of the screen, in the main "arrangement" view, you see all the ways I've cut up and rearranged that source audio into the final product, stripping out some pauses where they felt bad, adding in pauses where they felt good, cutting out a tremendous amount of stuttering and false starts and tangents. When I'm editing thousand-line source files, or documentation, or blog posts, I often use copy-and-paste in a similar (but less richly supported) way — I keep a scratchpad document open next to my main document, and whenever I cut something but don't know where it's going, I paste it to the scratchpad. Nothing ever stays in the hidey hole, because that's a recipe for lost work. But I often find myself in the situation where I rearrange the document, leave it for a week, come back with fresh eyes.. and realize the flow of thoughts is worse than it was before. I suspect that what Ted is asking for would help with this, by allowing me to "view things as they were" with a little more interactivity than just undo/redo or version history.

I really enjoy reading all your contributions, thanks! 🙂 I won't see cut'n paste the same way now.😄

The conversation here reminds me of this take https://99percentinvisible.org/episode/of-mice-and-men/ on Engelbart's work. Christina Engelbart comments that we computers ended up being the tricycle of the mind rather than the bicycle.

But for Engelbart, ease of use wasn’t the top priority. He wanted the computer inputs to be as powerful possible, and that required some complexity.

Oh to have been alive in that world.

Like Doug Engelbart, whose work he had yet to learn about, Nelson yearned for more than a lazy man's typewriter. They both wanted the freedom to steer their thought paths in new ways. And Ted especially desired the prerogative of changing his mind. He wanted the freedom to insert and delete words and move paragraphs around, but he also wanted the computer to remember his decision path. One of the specs was for something he called "historical backtrack," in which the computer could quickly show him the various earlier alternative versions of his ever-changing text.

From "Tools for Thought" by Howard Rheingold, who I believe was a friend of both Nelson and Engelbart. https://www.rheingold.com/texts/tft/14.html#Chap14

@Ivan Reese said:

I get to enjoy this exact mode of working when I edit video and audio

Exactly. Actually the movies are a big influence of Ted Nelson’s work. His father was a director and his mother was an oscar winning actress. When he describes xanadocs, he mentions that they’re based on EDLs (Edit Decision Lists, a term brought from film edition). The idea is, you create a set of “clips” of anything, and that’s what you cut and paste and rearrange. At that level of detail, it’s possible to work quite fluidly, as film edition software proves.

The undo command was hidden in applications and "There was no operating system undo in any way". I think this is key to Nelson's idea of an OS, which is more like an integrated hypermedia system, where you remix media of same (text-text) or different (text-image) kinds, while preserving provenance and history of changes. While there likely differences between the Alto and the Mac, neither is particularly like a single hypermedia system - you're always working within 'windows', while Nelson talks about drawing lines 'across windows'. I think Nelson is the original 'no silos/no apps' voice. There were still isolated 'apps' on the Alto, no (e.g. Bravo)? And hierarchical file systems, no? The criticism of taking away the terminal may be overstated though. Much of what Nelson says resonates very strongly with me.

Just put your git repo in hg. Easy. And put your hg repo in svn, and put your svn in perforce, and put your perforce.. oh no

And put your hg repo in svn, and put your svn in perforce, and put your perforce..

on your CVS, then on your RCS, then on a name-versionned tar file, then...

@Ivan Reese Its really a shame that Steve Krouse didn't get Juan Benet on the Future of Coding Podcast. Because he could add some interesting perspectives on that story.

I'm a bit late to this chat (just joined the slack group), but Alan Kay gave a demo on a reconstructed Alto VM as a tribute to Ted Nelson: