Back to the future

Has it occurred to anyone other than me that the Web is built backwards?

When you read a web page, you generally have the option to click on what are generally referred to as “forward links”. Yet in what sense do those links move you “forward”? Clearly the author of the page you are currently reading could only have incorporated links to pages that had already existed, before the one you are reading.

So in reality the next page you jump to is, to a first approximation, older than the page you were just reading. From the time you did the Google search that got you started, you are almost always traveling back in time, from newer pages to older pages. And yet we all have the feeling, while engaged in this process, that we are somehow moving “forward” – looking through our virtual windshield at the cyber-landscape rushing past, as we drive bravely into the future, click by intrepid click.

Whereas the actual process is more akin to skipping to the reference section at the end of a book, looking up the source for something, then going back to the library shelf to pull down an older book, skipping to the reference section of that book, and so on. Or, to continue the vehicle metaphor, it’s as though we are driving furiously backwards, while knowledge of the future recedes ever further away with each passing moment.

What would an truly forward-linking web look like? Is such a thing even imaginable? Not since Ted Nelson’s Xanadu has there been a major push even to support such a thing. What comes to mind when I think about it is a hypothetical reverse use of the Google indexing engine: Every time you post a page, web crawlers scour your content, looking for something relevant, and remember what they find for later. Then as people add more pages, sometime in the day or decade to come, your page is modified – a helpful link is added, pointing to the future knowledge that later flowered from your humble seed of thought.

Such a structure might have a completely different sociology from the one we now know. People might get into the habit of following arguments from their beginnings to their conclusions. Perhaps people would become more motivated to plant their thoughts in the fertile loam of cyberspace, in hopes that something beautiful would emerge as their ideas were joined by the answering ideas of others, building into crystalline structures of evidence and inference that spiraled upward into cathedral spires of progressive thought, lasting monuments to a renaissance of shared intellectual enlightenment.

But of course it could never catch on. You have merely to look around you to realize that people are much happier when they can start an argument from its conclusion and work their way backward to whatever ideas they already had in their head. Perhaps, when all has been said and done, we got the Web we deserved.

Sigh.

6 Responses to “Back to the future”

  1. manooh says:

    Don’t we have something like true forward-linking on wikipedia? You write an article, and add links to articles that you would like to exist (but don’t exist yet), hoping that somebody else would write them for you..

  2. admin says:

    Yes, I agree there is a similarity. Although Wikipedia is kind of a hybrid between the web as it is now and my fanciful scenario. As I described it, you wouldn’t need to specify links to prospective future articles (although you could), since those links would be added by the system’s A.I. as new knowledge appeared on the Web.

    Also, right now there is no reason to visually distinguish links as to whether they point into the past or into the future, since by convention all links point into the past. Once the system itself starts providing links into the future, it might become useful for browsers to provide look-and-feel difference between links to past or future, to help people know which way they are going.

  3. manooh says:

    well, I don’t know if the scenario you described would be the most “fanciful” for me as an author. When I write an article, I want to be very specific about which kind of information I want to link, and from where in my article, using what words. What I link to is a major part of what I am trying to express. I don’t think I’d want somebody else change that. As with wikipedia, I am able to specify that. Also, I want to make sure that my article doesn’t end up to be a mere collection of links, which might make it less focused on or less readable even.

    I don’t think the distinction between links into the future and into the past is that important to me (it isn’t on wikipedia, at least). The distinction between autor-made and AI-created ones is much more important.

  4. admin says:

    I completely agree! I was indeed referring to the distinction between author-made and AI-made links. The AI-made links are to things that the author couldn’t have known about at the time of writing the page, because they hadn’t been thought of yet.

    I certainly don’t think of the concept I was describing as a replacement for links. They are, after all, references to previous work, which is the basis of all scholarship – but rather for the addition of forward references, merged in over time by bots in the ecosystem, showing how what you’ve written has led to what came after.

    Both are useful, and they are different beasts – which is why it’s important for browsers to visually distinguish between the two types of reference.

  5. gerald says:

    Let’s imagine a similar system to Google Ads and let’s call it Google Articles. This system will dynamically add simialr articles looking for certain keywords, etc. So if you wrote an article about something in May 2008, and you visit it now, you may find links to similar articles that were written in August 2008 for example. You are traveling forward in “link time”.. :)

  6. sally says:

    ugh. let’s not imagine a “google” anything. i’m fed up with google.

Leave a Reply