That image is an absolutely fascinating historical artifact. He's describing (in a related paper) the essence of web sites in detail in 1965.
It's difficult to even imagine how few people were thinking of such things when they were so far removed everyday life and culture.
Tim Berners-Lee I thought was much more well known for starting a revolution with the world wide web. I hate to say but the web seems a bit less of an insight when you know that Hypertext had already been so well established for almost 25 years.
The one that really blows my mind is Vannevar Bush -- July1945 - "As we may think" - https://www.theatlantic.com/magazine/archive/1945/07/as-we-m... If you can look past the way he describes technologies of the day, there's an amazing amount of foresight in how access to information, its modalities and interconnectedness would become critically important.
- Automatic, redundant cross-network storage of all documents (a built-in Wayback Machine)
- A micropayments system for content creators to sell access to what they create
- Strong identity built in (a necessity if you want to also have payments built in)
... and much more.
Xanadu is (in theory anyway) a much more sophisticated system than the Web is even today, never mind the Web as it was when TBL first designed it. Back then it was so much more crude than Xanadu as to barely merit comparison.
But in a way that was TBL's genius, because, while the Web was much more crude than Xanadu architecturally, it was something that could actually be implemented with the early-'90s technology he had on hand. Xanadu, a much grander vision, was so much grander that it has defied fifty years' worth of efforts to implement it. (Back when Wired was worth reading, they ran a good piece on the state of those efforts circa 1995: https://www.wired.com/1995/06/xanadu/ It's still defying implementation today, more than twenty years later.)
So we ended up with the Web we have, which kinda sorta works except for all the ways that it really doesn't. (Paying content creators, for instance.) Ted Nelson saw all those problems coming, all the way back in 1960, and he tried to come up with a system that would head them all off. He hasn't succeeded (not yet, anyway!), but you have to admire both the vision and the attempt.
I favorited your comment so I can point people to it who have trouble grasping "worse is better". This is a much better example than clos vs scheme or linux vs microkernels.
Though when dealing with people who reject "worse is better", remember that some people may not grasp the idea, but others may grasp it very well and just consider it a bad thing - a problem to work around.
I mentioned "Worse is Better" on this site a few days ago and received this comment:
> "Worse is better" has nothing to do with it. Stop saying it every time anything you think is bad comes up. It's incredibly arrogant, the way people basically spam "worse is better" like they're in Twitch chat every time C, Unix, HTTP, or anything else that they perceive as imperfect comes up.
You're not being offensive. It's an important, non-obvious, under taught engineering principle. I understand the phenomenon he's describing and it's unfortunate he can't tell the difference.
Great point on TBL. Maybe it's a matter of style rather than greatest insight. Many great thinkers and theorists have died before there work is fully realized, and some are just pragmatic geniuses like Edison who can see the path to making things happen.
Both styles can offer so much to humanity, but I'd bet most entrepreneurs would go with Edison as a cofounder before Nelson.
The Xanadu site is worth reading. What a brilliant mind! The practical way to realize that vision, given the constraint of legacy content, could be graph databases.
I'm surprised that anyone could believe it required insight. Having lived through that period, nothing about it seemed insightful or surprising at the time. The only time I remember being surprised was when I saw a UPS truck with their URL painted on the side, in 1993. I remember thinking it was a couple of years ahead of time.
Great idea. My only suggestion is to have SEQme sort of symbol attached to the link that indicates that it expands the text and does not link to other material.
This sort of expanding text is used heavily in hypertext-based interactive fiction,[1][2] but I haven't seen it used outside of that before. Nice work!
Here's a long-term AI/ML goal - interpolate between Simple Wikipedia and regular Wikipedia article on the same topic and use StretchText to allow the reader to decide their level of depth.
I want this for learning faster, summarizing a page, but it's hard to teach something (to me). Like have it pull all the text in plain text, you read the page and pull the information that is new/important to you. Which this is subjective/different per person, but if the "AI" is unique to you then I think that would make sense.
But for this to work, you'd have to have a history/scope of your current knowledge.
