Hacker News Reader

Comment by jgrahamc | Hacker News Reader

The most interesting thing about this article, to me, is that part 7 is often skimmed and not understood. Bush describes the user building his own 'trail' of linked item (books, articles, etc.) and being able to store that trail. He was not envisaging the web as we have it (consisting of links chosen by authors), but of the ability to form your own web of connections.

It's in section 8 that he describes the something similar to the web. I've always felt that better tools could be built for part 7.

How many times have I lost the thread of thought while surfing the web and been forced to re-search to find my way back?

View on HN

Yes.

I also regret that we don't have Ted Nelson's original idea for hypertext, in which incoming links were also visible by default.

Of course in reality the spammers would have a field day with that -- every popular page would have about a jillion incoming links for dubious pharmaceuticals and dodgy gambling operations, but still, one can dream.

Can't you hack that together with a [link:] query on Google? You can even use Google's ordering algorithm as a spam filter.

Interesting idea. It might make a nice plugin, assuming it's can be done in accordance with Google's TOS.

I haven't looked at that part of it for a while, but I remember there are restrictions on automated queries. Making the user click a button or something might get around that.

link: has been incomplete for years and years; when I heard of link: I thought it'd be so awesome, and then I encountered the brutal reality...

You can approximate something similar though on sites you control: set Google Analytics to record referring URLs and after a few time-units, dump the set per-page to construct a manual trackback.

Don't do this if you care anything about SEO. It will make your site appear to Google as part of a reciprocal link scheme, and it will get nuked from the index.

Marking the links as rel="nofollow" should prevent that from happening, shouldn't it?

Well, I was kind of assuming that you would be doing a little filtering out of the worst links - only substantive discussions of your page.

I have also wondered a lot about the incompleteness of "link:<mywebsitesite>" command. Google's documentation is not very elaborate regarding that. Google webmaster on the other hand gives a more complete link of external links to the site.

My guess its deliberately incomplete because of:

1) Google's competitor search engines (bing, blekko, duckduckgo) to not easily get Google's Internet graph.

2) Competing websites should not get each other's complete link graph.

Hey thanks! His ideas are interesting. "HTML is precisely what we were trying to PREVENT— ever-breaking links, links going outward only, quotes you can't follow to their origins, no version management, no rights management". Would be useful, for sure.

I recently re-read his book Literary Machines after first reading it in the late 80s, and it struck me how simultaneously visionary and impractical his ideas were. His version of the web was centralized(), with the organisation running it covering its costs by franchising-out access points. yes, it had transclusion and versioning, but it also had rights management and automatic billing for quoting (transcluding) other people's work - all imposed by a central authority that owned the entire "web".

( it had to be centralised, otherwise the contract necessary for transclusion would not work in the real world)

Nelson is a very smart visionary, but I'm glad we ended-up with the messy, decentralised web we have today.

Obligatory link to The Curse of Xanadu at Wired, back in the day: http://www.wired.com/wired/archive/3.06/xanadu.html. A real hacker nightmare story.

Awesome, that is interesting - I was not aware of that. I was looking around to see if I could find his book. I found a PDF - don't think it's the entirety - but will give it a read this weekend.

I wonder if some kind of medium ground could be put together. Thanks for that article, I'll check that out!

The book is hard to get hold of now. There are used copies on Amazon, but the prices are stupidly high. I bought a dead-tree edition from Eastgate Systems [1] a few years ago at a reasonable price, but it seems to be out of stock now.

[1] http://www.eastgate.com/catalog/LiteraryMachines.html

It is called TrackBack and it is a standard feature on blogs.

Yeah, I know about trackback (and trackback spam).

Nelson's version didn't require any special stuff, though. His idea of the hyperlink was bidirectional by default. He also had a scheme that would let you "transclude" text from other documents. This was different from quoting, in that the transcluded text was live (i.e., when the original source changed, the transcluded text in the other document would also change).

> He also had a scheme that would let you "transclude" text from other documents.

And even at the time, he planned to modify copyright law to allow this.

(I believe launching a successful land war in Asia was next on his list.)

This reminds me of a browser extension I'd like to see. It could watch over your browsing and construct a graph of the sites you've visited, then allow you to spatially navigate your navigations. Throw in some auto-categorization and you have a nicely organized breadcrumb trail that you can then use to explore new spaces & fill in gaps in your knowledge.

^This! When I'm jumping into a new area, I go through the basics, and then it starts to get hard to see what I've already covered. Something that would make it easier to see that I've missed a section due to the "Squirrel" effect, or just the density of what I have seen, would be a major help.

TrailMeme [http://trailmeme.com/] is a step in this direction.

Reply on news.ycombinator.com