Slashdot Mirror
Web Page Entanglement
jason continues:
"By viewing the web through a tangle proxy, you can see the connections and associations left by those who surfed the web before you. By surfing the web using tangle, you also leave behind connections and associations for others who will surf in the future.
When you exit one page and enter another (by clicking a link or performing a search), a two-way link is created between the pages. As users surf through a particular page over time, tangle keeps track of popular ways to get to the page and popular places to go next. These entry and exit links are displayed at the top of each page, sorted by popularity.
Clicking on one of these entry/exit links tells tangle that you think the link is relevant and useful (like a vote for the link) and increases the link's popularity. In other words, if a user thinks of something relevant while reading a page and performs a search for it from that page, tangle gauges how others react to that association over time.
tangle is similar in some ways to the closed-loop hypertext system Everything2, though tangle works for the web at large.
We have several tangle proxies up and running. The tangle proxy software is also available for download.
A note for the paranoid:
Though tangle keeps track of web usage patterns, the focus is not on tracking the habits of individual users, but on tracking the trends of an entire community of users. tangle is GPL'd open source [source here], so you can see for yourself: clicking a link through a tangle proxy simply bumps up the links popularity---user IP addresses are completely ignored."
You'd get the a goatse.cx link on top of every page.
Through goatse.cx, and If we all play our part, we can get gnu.org associated with goatse.cx!
Does this mean that once quantum computers arrive, we will experience quantum entanglement?
Thank you, I'll be here all week :P
"Einstein argued that [...] God is not capricious or arbitrary. No such faith comforts the software engineer." ~ Brooks
It seems pretty poiintless to me...
What?!? Pointless? Think of how the porn industry can apply this technology...
Microsoft does something similar with their Smart Tags. That is, they modify your page without you realizing it. Only with entanglement, it's done on the server, rather than on the browser.
Is there a way to block entanglement?
Brilliant, I can't believe someone hasn't come up with this before. It reminds me of the traveling salesment implementation that models the way ants work. Most ants go the way most ants go, everyonce and a while some ants stray to find a better path.
If this isn't abused by users, I see the net becoming much more efficient for searching for information. You won't have to wait for the search engines to catch up while looking for the most popular page on a topic, because the best (or should I say most popular) pages on a topic will automatically link to each other based on user flow.
Am I missing something here, or am I right in thinking this will revolutionize the way we surf (that is if enough sites do it.)?
Why, o why must the sky fall when I've learned to fly?
If the more popular links are shown first, doesn't it just reinforce their popularity? Once a link becomes popular, is there any way to vote it down?
How long before this goes the way of the search engines with people abusing this to promote their own links?
This looks a lot like Everything2's automatic links. I wonder if people won't start using it to express their dislike in an anonymous manner (like, outlinking to "pieceofcrap.com" if they don't like the page)...
Excluding mutually authenticated ssl sessions, how can I trust that the document I'm reading is the document I tried to download? The tangle service is already modifying the page to add its navigation links, so why not change the content too ( e.g. remove content that users might find offensive, replace ads on popular pages with ads that you've sold, change links to documents you host, etc. )? The same really goes for any proxy or cache service, and I'm not accusing these good people of doing this, but how do we protect ourselves from services that would as more of them appear?
It sounds cool, but might prove to be useless... the phenomenon will happen that popular sites will be the ones getting the most hits and just perpetuating that way just because they are popular. More useful but less popular sites will be overlooked because they haven't been looked at much.
Comment removed based on user account deletion
If this caught on, I can imagine that it might be possible that people would tend to depend on it. It seems that information would become stagnate and new information ignored since nowone would have exited to it initally. Then again, maybe not. Just a thought.
Im not here now... Im out KILLING pepperoni
Well, we know exactly what the first entry link at NineNine's and autopr0n's sites will be.
bytesmythe
Hypocrisy is the resin that holds the plywood of society together.
-- Scott Meyer
Responstimes are close to a minute right now on the linked proxy. How would it stack up, if you ran a local entanglement proxy? Would response times still be high, due to negotiations with other nodes?
We do not live in the 21st century. We live in the 20 second century.
OK - not really like it, but if they start letting people leave comments, it'll be like thirdvoice (man, I feel OLD in internet time and thirdvoice wasn't even all that long ago!)
creation science book
I'm currently trying to figure out why people visit /. most often after visiting this link, which the entangle system tells me is a popular entry link:
/. to see if a similar article was posted. Weird.
CNN: Iraq Weighs U.N. Resolution
I can only guess that a lot of people rushed over to
Cogito ergo sum in Slashdot.
You should be using this (http://zip.cse.ucsc.edu:8080/request?inform_about _proxy=&link_from_page_title=&link_from_page_url=h ttp://slashdot.org/&link_to_page_url=http://www.gn u.org/ for those who don't trust me) link instead so the referrer will be Slashdot, so the referrer will be correct.
--j
put links on your web pages based on what the web page is about
http://www.vanillaafro.com - take me seriously and I will shoot you
This appears to use the same idea as referer-links on weblogs. Here's the progression from idea to uselessness:
- Obtain data from visitors as they browse.
- Post data obtained form visitors on the same site.
- Watch as three new internet startups market a tool to spam pr0n links on all the pages that use (1) and (2), above.
Only let your users post shit on your site if you want it to all be pr0n spam or goatse links.....Spam links start appearing?
Another question... When does Alexa get involved in doing "web page entanglement"... It would sort of complement their existing spyware infested "toolbar".
Hey, I just checked the entangled version of the Microsoft.com site, and all the entry and exist links seem to go to Slashdot, Free Software Foundation, or other places that Microsoft stands against. Looks like Slashdot has done its job. Pretty funny.
This appears as an exit link:
"anarax.net - easier to use than a virgin on prom night"
Not very tasteful for a professional site.
Don't get me wrong though, this is a very creative and useful thing. For example, this would be extremely useful for searching through technical support knowledge bases or for a large company's document archive system. I would just rather they leave my web surfing alone. ;)
I just added something along those lines to my website. I agree it's a cool idea. Of course mine is way more simplistic ::P
-- taking over the world, we are.
Is anyone working on a personal P2P portal? Seems like an extension of what you're talking about. What I see is software which works like a webserver but is local and accessed P2P. Instead of DNS you use the P2P model to direct traffick and search for content, whether it is files or html/web media. All you'd need is a renderer (think gecko) hooked in to parse html, etc. to the peer who is browsing your site. This of course could also serve up blogs or calendars or whatever other types of web services you wanted to offer to your peers.
A fool throws a stone into a well and a thousand sages can not remove it.
yeah, that's slashdot, modding it up as insightful instead of funny ;)
Is it just me or did the proxies listed on the site already get /. ed?
"Not knowing when the dawn will come, I open every door." - Emily Dickinson
Mods, this isn't redundant, it's true... and old news since Everything2 is already around.
Of course the problem they've experienced on Everything2 is that some cool or sexy sounding link is irresistible to click on, causing these links to rise to the top regardless of their relevance. Thus, it decreases the usefulness of the "entanglement".
Sex memes really are the most pernicious out there... can you honestly tell me you could resist clicking on "The Screensavers - Nude Episode"? The cost (clicking) to possible benefit (grrrrrrrrrrrrrrr) ratio is just too small not to expend the click.
Pop-up hell might increase cost, thereby disciplining hormonal clickers, but even then. The Onion used to have an ad called "Naked Scottish Weathergirls" -- one of the most clicked on on the web. It led to a messageboard eventually where people posted digitized women in Scotland -- so many people must have arrived there and posted messages asking about the naked women it was unreal.
I'm reminded of the idea of leaving your campus grounds unpaved, and then waiting for the "natural" grooves to appear in the ground where people walk, and then paving over those to make the sidewalks. You've probably seen an example of where there's a sidewalk connecting two points but then there's a worn-out groove nearby that's better, or connects from a more popular location.
Some people think it's rude or immature for people to create these grooves by not walking on the sidewalk, but I see it as an example of an arrogant designer who thinks he knows the best way simply by studying a piece of paper. It's amazing sometimes, the groove just appears almost magically in an optimal place, given the layout of buildings and traffic patterns.
This applies to web pages too. But, unlike sidewalks and buildings, you can't see your other destinations when you're sitting on a web page, so how do you know where to go next? This seems like it will just constantly reinforce the previous set of links, whatever they are.
I didn't fully read the documents (/. strikes again) but what I saw says you move from page to page either by 1) following an existing link or 2) using a search function. #1 is not going to create fresh paths.
It seems to me, a better idea would be to present a user with all possible links, or a subset of possible links, the first few times they visit. Then as they click through the site, add their arcs to the database.
After the first few visits, you can stop showing all links, and show them the "most popular" links. If you just show the popular links up front, new paths may not be discovered.
So perhaps this technique could be seen as a way to remove unpopular links, to trim the fat from a page. Then again, it might not be good to change a page after a person has gotten used to it.
It's very interesting though. As the web matures, you'll see more of this sort of analysis to move beyond static web pages.
1. Server load.
2. Limited feedback. Would be much more interesting as a tool for discovery if users could grade their findings. Presumably annotation would allow memos to be posted.
3a. Privacy concerns, i.e. this would seem to provide more transparency to crowds. And Slashdotters might become more predictable. (Nah!)
3b. Privacy concerns II. By announcing statistics of aggregate use it might be possible for a repressive regime (China, Scientology) to gain ammunition against individual websites by being able to prove how many visitors they had and (by purchasing an advertisement on an associated server like yahoo) what their IP addresses and demographic profile are (as impled by 3a above). ActiveX or Javascript exploits may also target heavy traffic streams with relatively little effort.
4. Confusing intent. Adding visible backlinks seems quite valuable. However the client still cannot look more than one ply above its current location in what is still an undirected tangle. Is the tangle team (nice name by the way) aware of the large body of work already accomplished in annotation, syntactic web, Xanadu, etc.? What pressures exist to get people to take the less-travelled routes, or is the purpose to increase the traffic of popular sites? In that case are annotations superfluous? More docs please.
5. (?) a bug in slash they note.
Does anyone else remember the What's Related feature that was in Netscape? It's still in Mozilla, but as a sidebar - pop open the sidebar, and there should be a tab labeled "What's Related." It's a list of links between the current page and webpages that people most frequently either leave from the site or use to arrive at the site (I think). Sounds very similar, but since it's already been Slashdotted, I can't compare the two. An interesting idea, but based on having played with What's Related, it isn't really all that useful - you wind up with a common set of sites, and the less well-known sites just get lost in the flood of popular ones.
You are in a maze of twisty little relative jumps, all alike.
So what happens when someone adds a line in their hosts file for gnu.org that points to a local server, adds a link to a modified version of the site with a link of their choice and clicks it?
What if your brower doesn't ever send referer headers? How does the system cope with that? Or do simply pass through without voting?
We're going to make information free Mr. Anderson, whether you like it, or not.
A note for the paranoid:
Though tangle keeps track of web usage patterns, the focus is not on tracking the habits of individual users, but on tracking the trends of an entire community of users. tangle is GPL'd open source [source here], so you can see for yourself...
Yes, but since this runs on the server, how do I know you're really running the source that's available?.
Or maybe I'm worrying too much, and the check really is in the mail, my information really won't be sold to 3rd parties, that really does happen to all guys at one time or another, and it's not me, it's you.
Ever notice that most comments starting with "Excellent" or "Brilliant", etc tend to be trolls? I almost overlooked this one because of it.
I still see a problem with the described methods though. That being, I don't think that the second-best search page selling product X would want a link running to the next-up competitor selling same product X.
The same is definately true for the second-best... do you really want users checking out where everybody else is looking for better deals?
If you knew that your prices beat the competition it would be a no-brainer, but otherwise it would be in some ways virtual suicide.
Which mod's mission is it to mod everything about everything2 redundant? It seems perfectly relevant...
No, (s)he's right. Turns out the article specifically mentioned E2's softlinks. My bad.
Stupid like a fox!
People will naturally click on the top-ranked link(s) on a page in the hope that they're useful. If they're not, you've just voted for them, making them even higher ranked.
Google has a much better method for this - it looks to see how many links there are on the web at large to a page. People don't tend to link to stuff unless they like it. Although it's open to some abuse, it's a much better solution.
I am typing this in my French cybercafe, which has 10 linux terminals on a broadband connection and an ageing Minitel (1200/75 baud, 9" monochrome screen, Cornflakes packet keyboard...). Scary thing is, to find a specific (and reliable) bit of information, it is often faster to use the Minitel. One of the main reasons is that the Minitel is structured in a way that is relatively intuitive for most people.
Tracking which paths people follow is very clever, but I can't help thinking that it would be better if website designers put more effort into their navigation aids, link pages, and - gasp - maybe listened to their visitors a bit more.
The real genius of the Minitel is that it got thin client technology into millions of French homes long before anyone in France or the USA had heard of the Internet, because it is as easy to use as a telephone. The Internet has a long way to go on that score, and I don't think being able to see how everyone else gets lost is going to help in this respect.
Virtually serving coffee
Content providers don't have control of what happens to their content after it leaves their server (other than not publishing it to the web in the first place). A link between two similar products is to the benefit of the visitor. They can do comparisions between products, and make a better educated decision. This benefits the visitor - the people who make the Web a thriving community.
If a company doesn't want a link on "their" page to a competitors better product, then they can catch a wake-up and improve their product, instead of rallying against freedom of information (in this case links) and the freedom of user choice.
A company has no problems with being indexed by Google and ranked lower than their competition - so they should have no problem with this method of ranking.
The original article can be found here
mt
Softlinks have escaped from E2 to the rest of the Web! No one is safe!
This was five years ago (+/-), and "grooves from walkers on a campus" was given as an analogy by Brewster as he showed off the alpha version. I recall that people's choices were only one of six factors going into the calculations, so popularity wouldn't create a positive feedback loop of overly deep grooveness (my paraphrase).