« Previous | Main | Next »

Let's experiment

Post categories:

Dan Gluckman - Product Lead | 15:00 UK time, Thursday, 27 August 2009

The 'Digital Revolution' project involves a lot of talk about the web. But can we do something more practical? Can we use the web to run experiments that test some of the big themes of the series? How the web is impacting on the nation state, for example, or our own privacy?

By experiments, I mean anything that tests how the web works. Following the twitchhiker's journey around the world gave insights into the real-life impact of the social web. Wired have an experiment on the go at the moment, to see how easy it is to vanish in the digital age.

Professor David Nicholas and the CIBER group at University College, London recently published a paper on the 'Google Generation', which analysed online searches to assess the critical and analytical skills of the 'researcher of the future'. We're planning an experiment with Professor Nicholas and his team at the moment to build on this, and find out more about the impact of the web on the way we think.

We'd also like to run some experiments to cover some of the other themes in the series:

  • privacy on the web
  • the impact of the web on the nation state
  • censorship
  • the value of our clickstreams
A few weeks ago, I was at a fascinating brainstorming session at the School of Electronics and Computer Science at Southampton University. (Professor Nigel Shadbolt and Dame Wendy Hall at Southampton, along with Tim Berners-Lee, have called for a new 'Web Science' that brings together people from a wide range of academic disciplines to fully understand the web and its impact). Topics ranged from the decline of old media, to the dynamics of power on the web.

So we're thinking of ways and means of testing the web and its effects on us, and we're open to suggestions. If you had a chance to conduct an experiment on the web and its users, what would you test?

Comments

  • Comment number 1.

    Here's a thought about how easily people can be taken in by what they read on the internet and a possible experiment idea.

    How about putting a fib/lie out in different sites on the internet/web and on various types of sites all saying the same fib (eg. blog, social networking site, microblogging site, forum, etc) and see how far it goes around the internet? Could you test which users/groups of the various sites were most likely to accept the lie and follow the chain to see how the fib spread around the internet? ie It starts out on a blog, someone reads the blog and mentions it on twitter, one of their twitter followers sees it and talks about it on a forum, another user picks it up and mentions it on another related forum.

    Could you also test how successful the lie was based on who started it off? eg. People may believe a lie about cricket if a respected cricket player starts it off on their blog, but the same fib/lie may not even get off the ground if your average armchair commentator is the one to start it off.

  • Comment number 2.

    How much information can you find out about someone via the Web. (I’m not thinking of remote access, key logging etc at this stage, but more generally.)

    If a newspaper, private investigator, switched on employer or police etc set to work on a member of Jo Public how much information could they track down and access about that person via the ‘ordinary’ Web? How much of a life-trail/click trail do we leave?

    Following on from that; if someone set out to track someone else’s life how far could they get? I’m thinking of geo-location software/apps on laptops, mobiles etc that can be used to pinpoint their location, remotely accessing text and email messages etc.

  • Comment number 3.

    I like Gary GSCC's suggestion, but would take it further than just seeing how far something goes around the web. Is it possible to change the accepted wisdom on something? ie can we propogate a lie to the extent that it becomes the dominant belief? The web is winner-takes-all, after all.

    It probably needs to be a mix of google-bombing, disinformation, spam, outright lies, and wilful attempts to mislead the mainstream media too. Remember the old chain letters about US gov't plans to tax email? That sort of thing, but spruced up to reflect the modern web.

    You'll need a theme that is important enough to generate interest, but which the general public are not well-enough informed about to question. The environment? Near-Earth objects? Invent a new terrorist group/ideaology? K-e-s-s-l-e-r syndrome? Miracle cures or hitherto undiscovered health risks? I'm sure a few smart people could come up with something believable enough to not only mislead the general public, but also to embarass a few politicians and scientists.

    http://www.dhmo.org/facts.html

    If you're going to do it then it should be planned and organised secretly, not discussed on a public forum on a highly-ranked website where it will show up on search engines.

  • Comment number 4.

    To all of the above, it's a case of great minds ;) Although, we have to temper ourselves from creating too much mischief and misinformation!

    We like the detective work suggested by @shefftim. In an opposite to the journalist who has attempted to disappear completely and challenged people to find him via any residual digital footprints he may have left; one could instead source an entire person from one spark of original online data. How much could we learn from a tweet?

    @TaiwanChallenges 'If you're going to do it then it should be planned and organised secretly, not discussed on a public forum on a highly-ranked website where it will show up on search engines.'

    Hmmm. You may have a point there...!

  • Comment number 5.

    Is the Web changing how we think? Following on from the left hand / right hand side of the brain discussion TaiwanChallenges raised - is the web making us 'think' in a more holistic way? (Or changing how we 'think' in any way?)

    I’ve heard that most people stick to just six websites and a majority struggle with basic IT tasks, so are they more likely to use the web in ways that reinforce current ways of 'thinking', rather than develop new ones?

  • Comment number 6.

    I would test how much people think about where the information they share online will end up.
    How much do they implicitly trust the people they share it with?
    Do they think about how long it will be around for?

  • Comment number 7.

    Thanks for all this input. As Dan Biddle said, we did think about the idea of seeding a lie or misinformation - but there are clear issues about the BBC doing this, which would have to be very much outweighed by the public benefit. In this case that isn't obvious, especially when the experiment is already happening!

    As TaiwanChallenges says, there might be some experiments we plan that we can't publicise very openly, but there are others that we can discuss. Something along the lines of SheffTim's first suggestion would definitely be interesting.

    On the subject of the web's impact on how we think, watch the blog next week, when we'll be posting thoughts from Susan Greenfield and Nicholas Carr among others. Hopefully that will spark some more conversation about the sort of experiment we can do to test these ideas.

  • Comment number 8.

    “can we propogate a lie to the extent that it becomes the dominant belief?”

    It could be argued that this is already happening; David Icke’s shape-shifting-aliens-posing-as-world-leaders and-eating-human-flesh is an example, but you could pick on any other conspiracy theory.

    The Maurice Jarre example that Dan gave is also a good one, as it also involved Wikipedia, a source many use and seem to trust implicitly.

    Some of the Web’s April fool hoaxes (Google is fond of them) could also serve.
    This year’s prank took me in for a while:
    http://www.google.com/intl/en/landing/cadie/index.html

    It brings us back to the key topic of how we filter and verify information? What sources do we trust and why? Is the web dangerous by appearing to give equal weight to any and every idea or belief that people write a page about?

    There are dangers in deliberately ‘propagating a lie’, not least that once spread it is impossible to eradicate all traces or convince all those that read it that it was a hoax all along. If originating from a BBC page it could also inadvertently damage the BBC’s reputation.

  • Comment number 9.

    One of my friends suggested the following:

    I'd like to see whether personal branding on the interweb can be demonstrated to increase credibility, and ultimately average earnings. If we can make ourselves into respected experts in our field through effective branding, can we thus charge more for our services than our less well known peer group? I.E. is peronal branding worth the effort?

    This is not exactly propogating a lie, but it's certainly about influencing what people think. The reverse side of the same coin is SheffTim's idea of finding out as much as you can about someone. They both come back to the issue of what information is available about us and how reliable it is.

    Excuse me while my mind wanders, but I've been dreaming this up all afternoon....

    Let's say you start with a concept like picking a random individual on a public forum. I'm fairly active on one that uses php, which is a popular platform, so let's stick with that.

    You can pull some personal data from my profile, and also follow the link to my website. You could also read all my 5000-odd posts and build up quite a good picture. Between just these sources you would start to know a lot about me. If you start getting smart and running searches on my username, email addresses, URLs etc I'm sure you can find other sites where I've been active, and of course you're also going to stumble across links I've created to other places where I've done stuff.

    But this all takes a lot of time and energy. It's appropriate for a stalker, but commercially-motivated organisations (legal or not) are going to be looking for ways to automate this and do it for large numbers of people.

    In the sixth-sense demo I linked to earlier, the presenter mentioned that their software searched profiles and key websites for keywords telling us about people's interests. You may remember the projector shining the word "student" onto the test subject.

    This word-searching reminded me of the British National Corpus. (http://www.natcorp.ox.ac.uk/)

    The BNC is a project to analyse samples of English from different categories, totalling 100,000,000 words, to see what we can learn about how the English language is used. You can look up any word or phrase and see how often it is used in business, informal language, etc.

    So we can count how often particular words appear in any given sample. You could find that I mention sailing quite a lot, but not football. This gives us an indication of our subject's interests, maybe. But by analysing the range of vocabulary, average number of words per post, sentence length, etc. we can get a pretty good idea of someone's education level.

    Language also tells us a lot more about individuals. Older people tend to use different words than youngsters. Certain slang is particular to specific regions or cultural sub-groups. Some words indicate that you systemise a lot, others that you're very social or holistic or you see the world in black and white. And if you know which websites or online groups people are frequenting - and how much time they spend there - the data gets even richer.

    I'm pretty sure that without resorting to use of the human eyeball it's theoretically possible to categorise people by interests, income, background, area of origin, religious affiliation, technical prowess, and a whole lot more. This sort of data must be worth a fortune to marketers.

    I don't know if this sort of thing could be done by a bot, or whether you would need to have a plug-in installed into participating websites, or a combination of these approaches. But if it's doable, then why not ask for volunteers willing to have their palms read by a virtual fortune teller? Telling the public "we can find out where you live" is one thing, telling them "you believe in Jesus, have good people skills, like animals, but you're stressed and a bit depressed" may be something else entirely.

  • Comment number 10.

    TaiwanChallenges. #9

    Good post.
    I'm sure it's what Amazon, Google and Microsoft (and many others) are attempting to do when they ask if they ask if they can (anonymously) track your Web use? (Or just do it anyway.)
    As for what 'Big Brother' might make of this is anyone's guess.
    If the Internet had been invented n 1900, and you and I lived in 1930's Germany, I would be feeling very nervous!

    As for "I'd like to see whether personal branding on the interweb can be demonstrated to increase credibility, and ultimately average earnings?"
    I'm sure it could. Success is more about appearance and 'believability' than reality. As many con-artists, gurus, religious and political leaders have realised.

  • Comment number 11.

    I am more interested in mobile web. I guess by the time the programm will be broadcasted some of the issues we are discussing here will have changed anyway. So why not concentrate on something that is still evolving and has not reached it´s peak yet?
    Mobile web means having access to everything everywhere at any time. I hink it will spur the expectation of immediate fulfillment even further. What I also observe is that you start seeing the world from above. When I grew up I experienced my surrundings differently. All I could know was what I could see from this normal frog perspective. I took ages to discover what was outside the nearest neighbourhood. Today I always feel like looking on a huge map. But it´s not a map that has streets and towns on it. It has nearly everything on it. So we felt like small subjects creeping around on this big object (the world) whereas today we are subjects following around all the different somehow segmented objects (that, put together, create the world). I have absolutely no idea how to verify these thoughts in an experiment. Or at least a short term experiment. Maybe you just need to give a group of people mobile devices and tell them to try and solve every single problem that comes up in daily life using mobile web. Track what they are doing and do semi-standartized interviews every second week including issues of general perception, interest, self regulation, self effectiveness (? I don´t the English word for that).

  • Comment number 12.

    The #NOHTTP Experiment

    Introduction

    #NOHTTP – a twitter hashtag
    bit.ly – a URL shortener service used on twitter as it is the shortest URL service – 6 characters
    tiny.url – the original URL shortening service – 8 characters
    j.mp – bit.ly’s new service domain – 4 characters
    twitter – a 140 character microblogging service
    http:// - the standard prefix to a URI – Internet web address

    Experiment

    So basically on Friday (4th Sept) I came across a headline in mashable’s rss feed that bit.ly had just launched it’s new domain for the service on j.mp saving tweeters 2 chars

    About 30 minutes ago it occurred to me that if twitter used some simple parsing they could strip out http:// from all URL shortener domains. It would not be difficult. On page generation just create a normal A HREF link, with no http:// displayed.

    Now whether Biz Stone and the crew think that the microblog should be limited to 140 characters or not is important, because it is their number. However they seem to have hit a magic number that just works with 140 characters. So this would be a way to bend that rule, not break it. bit.ly just gave us another 2 characters with j.mp and twitter stripping the http:// would give users 7 more characters. Yes 7 in one simple step :)

    It’s a simple idea I know. I went to twitter to suggest it but could not find “make a suggestion”. There is a community site that they monitor, but I will do that a little later. For now perhaps that would make a good little techie Internet experiment.

    A simple idea, here is the hashtag #NOHTTP

    Maybe it would be possible to seed an idea using a hashtag…

    I have a quid that says it gets nowhere ;)

    However, first Firefox dropped the need for http:// and then IE followed suit… there is a precedence… so who knows.

    However that may impact all twitter app providers as well… hmmm. A very small idea could result in a mountain of work. However, we are generally backwards compatible so maybe it will be a small race.

    Now for a sanity check? Is it quite simple? Is it is possible?… Yes I think so.

    And post…..

    http://www.of-networks.co.uk/blog/%2523NOHTTP-experiment

  • Comment number 13.

    I tried posting the URL on twitter without the http:// on the short URL and it did not do it. I thought it may have but alas no. Fitting it is the first post at #NOHTTP

    http://twitter.com/earthgecko/status/3796268287

  • Comment number 14.

    @taiwanchallenges I agree with @SheffTim re #9 - interesting: especially 'This word-searching reminded me of the British National Corpus. (http://www.natcorp.ox.ac.uk/)' - Reminds me of Calvino's 'If on a Winter's Night A Traveller': the woman who has invented the computer that reviews and 'reads' boks according to the frequency of words appearing in the text (I love that novel). I suppose that's not a million miles from a Search Engine's algorithm ascertaining relevance from the content of a page...

    Also akin to the We Feel Fine project which created a mood map of the web by pulling through any Tweets / posts that used the words 'love' 'hate' etc. Other apps have done similar - how long before Twitter-bots send you messages of concern for your mental state based on your 'dark' tweets? (And offer you Prozac into the bargain...)

    @vanboy74 - I agree mobile web is an exciting and developing medium (though friends of mine mobile would argue 'It's HERE, man! It's been here for years!'). And the web is mobile.

    Sort of. Setting up Twitter on some of the production team's 'less smart' phones showed me how much I take data access and applications for granted owning a certain brand of touch phone.

    But, in terms of the documentary about the web, we'll be touching on the mobile connectivity - especially useful in developing nations - but it's not the hub of our series.

    @earthgecko - I'll take that bet ;) You shouldn't think Biz et al at Twitter would do anything to upset a plan to get more in, as I believe the 140 character limit wasn't an arbitrary 'Heinz 57' number, but a direct reflection of the telecoms standard of 140 SMS text limits, Twitter originally being a web browser and SMS-based service (remember those days? They seem so long ago...!)

    Considering the business that URL shortening has become, a trick to shorten that process further would possibly mean losing your £1 bet but making a small fortune in the process :)

    I'd be interested to see whether #NOHTTP takes off.

  • Comment number 15.


    That is the thing Dan sometimes a good idea just has to be put out there as you cannot build everything, but someone else might ;) I would be surprised if it took the dev team at twitter longer that 30 minutes to put the #NOHTTP stripping into the entry form and into the page generation.

    ON the post:

    if grep -c http:// var_POST -nt 0 , then
    URL_SHORTENER=grep -c var_URL_SHORTENER_DOMAINS
    if var_URL_SHORTENER -gt 0 ; then
    if echo var_URL pipe_thru grep -c var_URL_SHORTENER_DOMAINS -gt 0 ; then
    NEW_URL=echo var_URL pipe_thru sed -e s/http:\/\///g
    URL=var_NEW_URL
    fi
    fi
    fi


    On the page generation: (note an a nref was used to prevent HTML linking an a href)

    if echo var_URL pipe_thru grep -c var_URL_SHORTENER_DOMAINS -gt 0 ; then
    NEW_URL=echo a nref=http://var_URL var_URL
    URL=var_NEW_URL
    fi


    That is a very simple shell interpretation of the concept (expect the dollar char was replaced with var_ to stop the post breaking and square brackets were removed, and pipes and backticks, and all quotes). That could be translated into ruby quite simply. And a lot better than that :)

    And yes!!! How we do forget where twitter started out :)

    I still think I will win that bet ;)

    However they will surely do it at some point. A WHOLE 7 characters :)

  • Comment number 16.

    Better yet, seeing as it is a link shortener and short link that has not other data other than the link provider, why does twitter not just convert the entire short URL to a link?

    (A HREF="short_url")L(/a)

  • Comment number 17.

    Dan, posting on a Sunday is a sign of addiction. I've managed to keep away for three whole days, just to prove I'm still in control.

    Can't offer any opinions on Twitter as I'm not a user. Ouch! Sounds like heroine.

    Computers to read books for us. VCR's to watch TV for us. Sounds like Douglas Adams' electric monk, which believes things for people who are too busy to spare the mental bandwidth.

    Internet. Belief. Information processing. Can't help feeling there's a link there. Not least with your comment about Prozac.

    Google reads your mail to target ads at you - albeit very badly. They could read your tweets too, and could they read your blog or posts like these? I'm sure that 'they' could, whoever they may be.

    How long will it be before, instead of Prozac, we get policemen coming to our doors after being automatically notified that we sound discontent or potentially revolutionary?

    The semantic web doesn't just make it easier for individuals to find what they want, it could also make it possible to build up a very deep picture of who that individual is.

  • Comment number 18.

    @earthgecko 'Better yet, seeing as it is a link shortener and short link that has not other data other than the link provider, why does twitter not just convert the entire short URL to a link?

    (A HREF="short_url")L(/a)


    You'll get no argument from me there. There may already be apps that do support that concept - I'm sure I've seen tweets where the text was an active link without having to slavishly surrender characters to a URL, shortened or no. But I couldn't swear to that...

  • Comment number 19.

    Another hashtag experiment for you this one is a bit time limited:

    #yourvote4nobel #Tim_Berners_Lee

    After all there must be better than Obama.....

 

BBC iD

Sign in

BBC navigation

BBC © 2014 The BBC is not responsible for the content of external sites. Read more.

This page is best viewed in an up-to-date web browser with style sheets (CSS) enabled. While you will be able to view the content of this page in your current browser, you will not be able to get the full visual experience. Please consider upgrading your browser software or enabling style sheets (CSS) if you are able to do so.