Slashdot Mirror
Hypertext Creator: Structure of the Web 'Completely Wrong'
angry tapir writes "The creator of hypertext has criticized the design of the World Wide Web, saying that Tim Berners-Lee's creation is 'completely wrong,' and that Windows, Macintosh and Linux have 'exactly the same' approach to computing. Ted Nelson, founder of first hypertext project, Project Xanadu, went on to say, 'It is a strange, distorted, peculiar and difficult limited system... the browser is built around invisible links — you can see something to click on but you’ve got nowhere else to go.'"
“[My approach] would be entirely different from today's documents where you look at one page at a time and you can see a ribbon or beam connecting documents together,” he said. “Having to refer to a paragraph and a sentence in an e-mail is just so barbaric when you could just strike it out and make the connection between sentences.”
Is it just me, or is this just completely incoherent? What the hell is he talking about?
you dont know whats inside until you get there and look around, sometimes you have a good idea whats there and can be predictable at websites (or stores) you frequently go to, but when opening unknown URLs (or visiting new stores) you have no idea until you get there
Politics is Treachery, Religion is Brainwashing
You mean that thing that's supposed to be superior to the World Wide Web, but that's been in development hell for the last fifty years? (Duke Nukem Forever, most delayed software ever? Ha.) Someone needs to tell this guy that it doesn't matter how superior your invention is if no one ever sees it. Like Steve Jobs said, "Real artists ship."
Dislike the Electoral College? Lobby your state to join the National Popular Vote Interstate Compact.
He lists several very abstract complaints, without giving an example of at least one way in which he thinks it could be differently, and done better?
I'm not in complete disagreement with him that the web could be improved. For one thing, we've given website creators so much control over presentation, that there's no standard 'look' to hyperlinks anymore - ever been to a website and not even *realized* that one of the elements in the page was a link to something else?
Also, there's too much problem of link obfuscation - the problem of the user having absolutely no idea where a link will take them, because when they hover the mouse over the link, it just shows some useless javascript, or the site designer used some javascript to make something which is not a link behave like a link, but not actually give the user any feedback about where it goes to, or the link is rendered by Flash, and Flash never tells you where a link goes. I just hate that.
But, I'm not really sure that's what this guy was talking about. In fact, his complaints were *so* abstract, I have no idea what he was complaining about?
Yet another visionary wanting to do something different just for the sake being different. It's become popular lately to claim that particular industries or areas are doing it "all wrong", because naturally, if their whole process is "wrong", and you know the "right" way, then you're a genius right?
In reality, some things haven't changed in a long time because we've figured out something that works well. Every time I hear one of these "revolutionary" interface ideas they work well for the couple of examples that their creators can cite, but typically fall flat when you try to then adapt it to the entire world of computing.
"People who think they know everything are very annoying to those of us who do."-Mark Twain
Correction, what Xanadu "is" ;)
http://www.xanadu.com/
It's basically an MDI for browsing where links open horrizontally and scroll with the page. It's a clugy attempt at what he is talking about.
-Rick
"Most people in the U.S. wouldn't know they live in a tyrannical state if it walked up and grabbed their junk." - MyFirs
If you think of Xanadu as a highly available redundant P2P document system mixing in TBL's Semantic Web and adding more automation, you get a bit closer to what Ted Nelson was trying to do with Xanadu.
http://xanadu.com.au/general/faq.html
Section two of the FAQ covers what a Xanadu system was supposed to entail.
This article (originally on Wired) covers some of the controversies that have broiled up:
http://aether.com/archives/the_curse_of_xanadu.html
If you can find Nelson's 1982 Datamation article it is pretty interesting but I couldn't find it anymore after some quick Google searches (YMMV).
Basically he's arguing for bringing that even further. For example, in Wikipedia, there's a separate main page and discussion page, and on the discussion page, things have to be categorized and organized. Under his model, instead of having this, a person that wished to lodge a new comment or question about a certain bit of text would highlight the text they were commenting on, do some sort of drag-and-drop-like operation to comment on it, and write their comments.
Then another user browsing the comments would do so by browsing the main document, and hovering over a word they would be able to see a list of comments made on "phrases" (using the term loosely) containing that word.
I suspect the people that dream of wikifying the legislative process have similar ideas about granular live documents, as that would be where this type capability would be needed most obviously.
Someone had to do it.
man who says impossible shouldn't interrupt man who does
So Teddy boy comes up with a concept, theorizes around, accomplishing (near) zilch building his ivory towers out of clouds for 20 years and he's complaining about the 50 million bazillion websites people have made, some of them actually useful? Jeeze, at least pretend to be relevant by helping pound a stake through the heart of Flash.
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
I think what he's suggesting is this:
Many documents are composed of parts of other documents. If I write an essay I might quote from source texts, scientific papers, other people's work on the subject, interviews I've conducted, etc, and I'll add my own ideas around this. At the moment, I duplicate (retype) any source material and provide a link to it. The material I've linked to doesn't automatically link back. Instead, I could make a link using his system which includes the text from the version of the document I look at, and provides a two-way link.
It's a nice idea, but unless you can make it easy to create documents with all these links (and ensure they don't need any maintenance) I don't see how it would catch on.
Wikipedia's software is close in some respects -- you can include pages (but not, AFAIIA, selected bits of pages) in other pages. There aren't links in the UI, but it would be trivial to add them.
One thing they've mentioned on many occasions is that 404 errors bug the shit out of them. In the Xanadu system, all links were two-way, and you couldn't end up with a broken reference like that.
How would it be possible to not have 404's unless every document took control (ie. a copy) of every document to which it linked (and subsequently would have to link to everything linked in those linked pages, ad infinitum).
That seems to be the obvious flaw in everything this guy has talked about for 50 years. XanaduSpace is really no different from a web browser with regular links, all that it does is load all linked pages simultaneously and display the linked documents in the background of some 3D view. Real browsers don't do this because they have to deal with the reality that the linked pages are hosted remotely and therefore have latency and bandwidth issues which need to be balanced with the likelihood of a user wanting to actually follow that link.
XanaduSpace's entire concept seems to be predicated on the assumption that all linked content is immediately available and immutable. This obviously cannot work on non-trivial amounts of data. Either it would mean having the entire Internet on your local computer or, slightly more realistically (but altogether more scary), having some kind of central Internet server/database/authority that maintained control of all published documents. Short of an international fascist uprising I don't see that happening.
Spelling mistakes, grammatical errors, and stupid comments are intentional.
I think the bigger problem with his system is that it would only work if all the source material was kept on the same server. Or at least if there was a common document provider to serve it.
The way the web works today doesn't allow this. Sure, you could fetch some text part from a remote server somewhere, but what if that site goes down? Or what if your document contains 100 snippets from 100 servers? Just imagine the load times.
At least now, when presented with a hyperlink, the user has an expectation that it might be broken, but even then the locally stored text remains accessible.
And then we didn't even mention copyrights...
If you look at the rules he proposes you'll see that half of them are about restricting access and creating profit venues for the publishers.
Ted Nelson's view is a web where you have to pay for each page you visit. We have seen too much of this lately
For an example of a wiki that has better (but still limited) support for transclusion, see Wagn.
The problem with true hypertext as described by Ted Nelson is that it's a hard problem to solve. Your document editor really needs to be aware of the transclusions, or else you need some really complicated diff algorithm to work out your changes and then apply them properly.
That said, we probably would have seen a working example, if the Xanadu Project hadn't suffered from project management disasters. (Waterfall model, development in secret, second system effect, name an antipattern and they probably did it.)
It's also rather sad that his books are hard to obtain and not on the web, so people are generally unaware of how much actual useful work was done and how good the concepts were.
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
And that is how it should be. If the cited text changes, you do not want your text, which refers to the old version, to suddenly apparently refer to the new version.
Continue reading what I wrote:
I could make a link using his system which includes the text from the version of the document I look at
(Alternatively, I might choose to link to the latest version.)
Thanks for the link. The idea is brilliant and radical (and for perhaps the first time a youtube video where the comments underneath made sense ;-) ).
However structure of paper document he accuses of being limiting reflects how our brains are geared to work. Having all those parallel hypertexts and floating links would be quite distracting - cross linking on wikipedia for example is distracting enough on its own. Footnotes, references and asides are what they are for a reason - they are not the actual subject of the document - and hence should not distract the reader whose brain can process only one stream of thought at once.
Besides, as someone else note above, I can't see how this would scale with more than handful of documents. Who's to say what the URI for a piece of text is and where it lives? Does modifying one its "hyper references" modify every instance? And he needs to stop using cheesy terminologies like flinks (floating linnks, apparently!) if he wants to be taken seriously.
If you look at the rules he proposes you'll see that half of them are about restricting access and creating profit venues for the publishers.
Ted Nelson's view is a web where you have to pay for each page you visit. We have seen too much of this lately
Let's go down the checklist to see how well the WWW complies or has a mechanism TO comply (as in, without forcing someone at knife point... or... Cranky Old Man Cane in Your Chest point):
Every Xanadu server is uniquely and securely identified. - Not Done
Every Xanadu server can be operated independently or in a network. - Local, Intranet, Internet, Done
Every user is uniquely and securely identified. - SSL, Done
Every user can search, retrieve, create and store documents. - Google, Done
Every document can consist of any number of parts each of which may be of any data type. - HTML5
Every document can contain links of any type including virtual copies ("transclusions") to any other document in the system accessible to its owner. - Done
Links are visible and can be followed from all endpoints. Pingback, Done (unless he means forcing reverse linking... HAHA, screw THAT!)
Permission to link to a document is explicitly granted by the act of publication. - Done, we just can't convince the RIAA/MPAA of that...
Every document can contain a royalty mechanism at any desired degree of granularity to ensure payment on any portion accessed, including virtual copies ("transclusions") of all or part of the document. - Done (it says "can" contain "a royalty mechanism", so yes, there is not restrictions on the WWW that force a document to explicitly NOT contain a royalty mechanism)
Every document is uniquely and securely identified. - URI, Done
Every document can have secure access controls. - SSL, Done
Every document can be rapidly searched, stored and retrieved without user knowledge of where it is physically stored. Google (ever really know the drive letter of website pages you search for?), Done
Every document is automatically moved to physical storage appropriate to its frequency of access from any given location. Amazon EC2, Google, Facebook, Load balancing, blah blah blah, Done
Every document is automatically stored redundantly to maintain availability even in case of a disaster. Raid1,5, Done (unless he means forced mirroring, again SCREW THAT)
Every Xanadu service provider can charge their users at any rate they choose for the storage, retrieval and publishing of documents. - Rackspace, Done
Every transaction is secure and auditable only by the parties to that transaction. Part Done, SSL isn't the norm. But switch to SSL only, and Done.
The Xanadu client-server communication protocol is an openly published standard. Third-party software development and integration is encouraged. - Done
Beyond that, there's a few good points left. SSL should be standard as proven by FireSheep/Facebook debacle. Um... More people need to mirror... oh gee, I guess there aren't really any points left, unless you wanna force backlinking. And, with all do respect, he can shove that up his Xanadu! We have enough ads and spam without being force to replicate links back to link farms.
I8-D