Making Wikipedia better, Part II

Making Wikipedia better, Part II

Summary: Making Wikipedia more transparent for readers, reflecting multiple views and adding more hooks that make monitoring articles convenient will add to readers' confidence.

TOPICS: Browser

This is Part Two of my three postings about how Wikipedia can be improved. Ross Mayfield, a friend and frequent interlocutor with me on this subject told me the first installment made him want to strangle me, which I find encouraging. Tough love makes us better.

I wish Wikipedia would forget about being an encyclopedia, because the form is broken. As a reference work, the encyclopedia strives for a single version of reality, even thought there is little doubt that on many issues there are many viable interpretations of events and facts. Because Wikipedia emphasizes the strengths of its community, the members of that community make some egregious mistakes of self-indulgence that a greater attentiveness to the reader would ameliorate. I wish Wikipedia would forget about being an encyclopedia, because the form is broken. As a medium for dialogue, which is essential to democratic society, Wikipedia suffers from a lack of convenient tools for understanding who is changing articles, preventing readers from making informed judgments about the value of information they find in the encyclopedia.

Let's take this summary of the faults of Wikipedia, which I think is can be an invaluable addition to the world of knowledge, and focus on how to make it better. Better than Encyclopedia Britannica. Better than an encyclopedia. Better as a community.

Forget one reality, objectively delivered. It just ain't so, unless one is talking about how to perform a mathematical calculation. Even in matters of science, there are contending interpretations of the meaning of data and they are critical to revolutions in perception. The fossil record is reinterpreted by new findings that place species in different epochs than they once were. Quantum Mechanics undermines some of Einstein's findings, which overturned Newtonian physics. Capturing the flow of change is more important to human understanding than freezing one version of reality. This is why I agree with Ross that a "stable version" of the Wikipedia, which Jon Udell called for in a thoughtful article yesterday, need not be a goal of the site, though it may be useful for a printed version—but, I'd ask, why print the thing? It needs to continue to change and fork.

The greatest problem with the Wikipedia, in my opinion is that it inherited from the encyclopedia the notion of a single entry about a given topic. Wikipedia must encompass differences of opinion much better than it does. Rather than have a single entry about controversial topics, such as George W. Bush or sociobiology, the most productive approach would be to allow the reader to see contending views-—more than two of them, as life isn't binary—presented as a series of articles under the same heading. It would yield a much richer view of facts. For example, the Bush entry doesn't deal with campaign issues he ran on, putting them in another canonical article (one article about each campaign) that, amazingly, summarizes the campaign issues of 2004 as focused primarily on non-issues, such as allegations about President Bush and Senator Kerry's military records, without any mention of domestic policies that were prominent or substantive debate about Iraq policy.

Giving full range to controversy by allowing different opinions eliminates the tendency of the encyclopedic form to take a stance of infallibility, which is the greatest weakness of existing media. Multiple articles would allow controversy to flourish and, where someone has made a decision about what they want to consider, readers can choose for themselves to ignore and pay attention to what they want. If I read about evolution, I will not give intelligent design equal weight and would likely ignore any entries about it, but  someone else may want to consider it.

Freedom of thought requires that even stupid ideas get their due. The evolution entry could, ultimately, dismiss intelligent design, but the fork to the intelligent design article could remain as an option when displaying results on searches for evolution entries. Let all sides be heard rather than try to summarize all views into an "objective" article.

An author needn't write the whole of each article, but could use other authors' text to fill in areas where they agree, while focusing on other issues of dispute. A reader could then view some summary of what is agreed upon across all versions of an entry in Wikipedia to understand what is less controversial information. 

Fernanda Viega Iraq Entry StudyBetter representation of authorship and changes. I wrote last month about how difficult it is to identify the source of one part of an article. Readers should be able to see a list of contributors. Visualization of contributions would help understand who contributed the bulk of an article or repeatedly changed an article to assert a particular view of the topic. Fernanda Viega has an excellent visualization of Wikipedia contributions that should be adopted by the site. Her analysis of the Chocolate entry, for example, shows how a small group of editors erased correct information until the contributor who had inserted it just gave up. In the visualization above, Viega's study of the Iraq entry in Wikipedia shows that the whole article was deleted (the black strip that breaks the color bands) and replaced with a rewritten version at one point.

Knowing who made changes and being able to investigate contributors' conflicts is key to informed readership. 

Identity can be pseudonymous. It doesn't matter if the contributor is "Zyclops" or "Robert J. Knowitall, Ph.D.," as long as the reader can look into the contributors' expertise, what they contribute to and how their contributions are accepted by other editors. It helps them make a judgment about whether to take the information with a small grain or large rock of salt. Contrary to the assertion that all articles should be taken with a grain of salt, some need to be viewed with deepest skepticism, so we have to arm the reader to make those decisions for themselves.

Addresses should be displayed, too, so that someone using the same address for different accounts can't game the system by appearing to be different people endorsing the same idea. 

Anonymity may sound great, but we're not talking about state secrets. Wikipedia is a public record of human knowledge, so be on the record (even if it is pseudonymously). It feels pretty good to be recognized, under whatever name you choose, for meaningful contributions, which is why David Weinberger's suggestion that "arrogance, individual heroism, accountability and discipline" are elements of the institutions of knowledge propagation that can be abandoned in favor of "modesty" is misguided. Society happens on the record, even when the table is set privately. The downside—and there has to be one—is the writers can be ridiculed for dumb mistakes. Welcome to my world. All you can do is demand rational critiques from others and ignore or mock invective.

Solicit input, don't wait for it. The Wikipedia needs to reach out much more actively to experts in many fields. They should be invited to edit and contribute, because they will also increase the reach Wikipedia has in specialized communities that reflect deep study and conflicting opinions that will make articles richer. I read several academic history mailing lists where experts in, for example, Gilded Age economic history, have never been invited to participate and are reluctant to because of the lack of transparent governance and editing practices I discussed in the previous posting about Wikipedia.

In a world where publishing is important to the dissemination of ideas, academics still seek new channels for their articles and would, I believe, be happy to have clear guidelines and the confidence to participate in Wikipedia.

Wikipedia administrators, in addition to taking responsibility for topics, need to adopt the practice of reaching out to people with complimentary expertise.

Remember, it's a resource, not just a community. Wikipedians must remember that, even if they work hard on the Wikipedia, the users of the information are the readers. The readers, today and in the future, are your customers. The poorest entries in Wikipedia reflect the self-absorption of a poor writer as much as, or more than, the want of an editor.

The vaunted "social process" of editing Wikipedia is not very different from previous forms of editing. Instead of an editor calling and author and saying "I don't think you got this across as clearly as you liked" someone can just change the content. But, like all human processes of creation, both modes of editing involve give-and-take. Showing how that process of compromise happens in Wikipedia, rather than keeping it hidden as it is in traditional publishing, would be a great step forward in transparency.

Disclaimers, disclaimers, disclaimers. I agree with David Weinberger and some Wikipedians that there should be a disclaimer on every page of the site saying that students shouldn't cite the articles without doing their own research. The point is not to diminish the value of the Wikipedia, but to encourage skepticism and critical reading by children and adults.

The other element of disclaimer has to do with how Wikipedia plans to license its content to other sites. If Wikipedia intends, as it seems to, to become the encyclopedia of record for the Internet with versions of its content displayed on third-party sites, that content has to be accompanied by these disclaimers, as well. It may make the text less attractive to partners, but that's the cost of being upfront about your principles. 

Dedicate resources to ease of monitoring. Jon Udell also calls for RSS feeds of Wikipedia articles. This is critical to allowing people with an interest in the representation of a particular topic to monitor changes. People should not have to visit Wikipedia to keep up on changes. Alerts should also include authorship information, so that coordinated efforts to control content are more easily identified. PR people are sure to use Wikipedia to influence market views of issues related to their clients' products and services, so if all that is available about an editor is an IP address, provide address resolution to show readers where the changes are coming from (e.g., if addresses are changing Windows- or Linux-related articles, for example).

Next, I'll turn to the question of relevance and information overload. The infinite storage and bandwidth of the Net is no excuse for mindless errata becoming part of an encyclopedic canon.

Topic: Browser

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Sure, like I have that much time...

    Mr. Ratcliffe makes a huge assumption in his article, which is that anyone has the time to actually research these things. Let's look at some of his ideas:

    * The reader should investigate the change history of an article. Even with the nifty charts he links to, I simply do not have time for this.

    * The reader should investigate the contributors' history, other articles, etc. to see if their input is likely to be useful. Sorry. I. Don't. Have. Time.

    It is not my job to investigate if a source is trustworthy on an item-by-item basis. I can either take the time to investigate the source as a whole, or not bother at all. For example, I trust The Economist to be extremely correct on a factual level, because that is their reputation and I have not seen anything to the contrary. Wikipedia, as a whole, is not trustworthy. For the amount of time I would need to invest looking into the factuality of each article, I could go read two or three entire books about the topic which come with a higher trust level right off the bat.

    Don't even get me started on "community policing" or weighting contributors, or rankings by users, or whatever. The same spambot that I envision to flood Wikipedia with total garbage could just as easily pump rating up too.

    Mr. Ratcliffe just needs to own up to the truth and stop trying to polish the turd known as "Wikipedia".

    Justin James
    • That's the world you live in today

      J.Ja?I'm not trying to polish Wikipedia per se, but the whole
      information gathering process. If I am now a "defender of
      Wikipedia" after being lambasted by Wikipedians as a critic, then
      I must be doing something right.

      Whether you like it or not, you already choose to trust some
      sources today. You do it by either: a.) Doing research on a
      source until you are satisfied it is trustworthy, usually based on
      your own or other people's experience with the source, in other
      words you trust a brand like the Economist, or b.) You trust until
      the source disappoints. Both are wrong, since the basic idea of
      an informed society is one in which sources are challenged, not
      accepted. Even the Economist makes mistakes in every issue, so
      keep your bullshit detectors on, that's all I am saying.

      What makes you think a book is more accurate than a Web page?
      You either trust the source or you don't. Plenty of books are full
      of spin and lies.

      Rather than suggesting any of the community policing or
      ratings, I am urging Wikipedia to recognize that the editorial
      processes of the past can help improve something that has, at
      least, shown that groups of people can collaborate to make the
      authorship of an encyclopedia more egalitarian. So, the truth,
      which I've gone into here and in other postings, is that all
      information publishing needs a lot of work. If these suggestions
      aren't a positive step, I'll be surprised.

      Wait until the last installment, then everyone can think I am a
      Mitch Ratcliffe
      • I would like to agree with you, but I can't

        "I'm not trying to polish Wikipedia per se, but the whole information gathering process."

        That is completely understandable. Everything I have said is actually pretty generic as well; it can be applied to virutally any type of content that does is not produced by people paid to produce it, with a staff of editors and publishers to oversee the process (wikis in general, blogs, personal home pages, etc.).

        "Both are wrong, since the basic idea of an informed society is one in which sources are challenged, not accepted. Even the Economist makes mistakes in every issue, so keep your ******** detectors on, that's all I am saying."

        The difference between a professionally source, whether it be online or offline, and amatuer hour, is the final authority. For example, with a magazine such as The Economist, or a website such as Slate, my trust is not in the individual authors per se, but the entire team: authors, editors, publishers, fact checkers, etc. Once a source has been verified to have an acceptable level of accuracy, unless there is a major change in personnell or process, it can be counted on. With a "community" based source, such as Wikipedia or my local 'zines printed Xeroxed and found at the record store, the lack of a team to support the authors results in a situation where the accuracy of every individual piece must be verified. To make matter worse, lies, misinformation, and mistakes replicate infinitely faster online (at the speed of cut and paste!) than they do offline. Look at recent events in the newspapers and TV news programs. They are rushing stories out the door without proper fact checking in order to keep up with bloggers. As a result, their credibility is sinking to that of bloggers, which is to say none. At this point (thanks to Jayson Blair), the New York Times' articles are considered barely more credible than op-ed pieces. The National Guard flap last year sent TV news' credibility to zero as well. Both of these were caused by a failure in process. Correcting the process will correct the problem. Wikis, blogs, et al have no process for fact checking. This is their fundamental flaw. They arrive at a community consensus of truth. If 95% of the people on the planet think the world is flat, that doesn't make it flat.

        "What makes you think a book is more accurate than a Web page? You either trust the source or you don't. Plenty of books are full of spin and lies."

        I have already addressed this issue a number of times, please see the following links:

        "Rather than suggesting any of the community policing or ratings, I am urging Wikipedia to recognize that the editorial processes of the past can help improve something that has, at
        least, shown that groups of people can collaborate to make the authorship of an encyclopedia more egalitarian."

        The only thing that can truly correct the problem is having editors edit what they have no emotional investment in. That is the beauty of professionally published works. Often, no one working on a piece actually cares about the subject. They are objective as a result (like Plato's mythical philosopher kings who rule because they are suited for it, not because they want to). Anything produced by a "community" is going to, by its very nature, be created and controlled by people who have a strong interest, and therefore emotional investment in a particular viewpoint. Look at Slashdot. You will NEVER see anything objective about Linux on there. I trust them when they talk about astronomy.

        "Wait until the last installment, then everyone can think I am a jerk...."

        Eagerly anticipating it! I know I'm arguing against you here, but I do think your posts are a breath of fresh air on ZD Net!

        Justin James
        • Welcome to social constructivism

          "Wikis, blogs, et al have no process for fact checking. This is their fundamental flaw. They arrive at a community consensus of truth. If 95% of the people on the planet think the world is flat, that doesn't make it flat."

          A few years ago I returned to college, thinking about getting an advanced degree in Technical Writing (my prior degrees are in Chemistry). Those plans changed when I ran into one of the major foundations of modern Technical Writing (in academics, at least): social constructivism. I couldn't get past that point.

          I deal with "facts" every day - the boiling point of a liquid, for example. While the number assigned to the boiling point is arbitrary, the underlying fact that the liquid boils at a certain kinetic energy content is indisputable. It is a "fact" not dependent on any "community consensus". Yet social constructivists would have us believe that liquids only boil because everyone agrees that they boil, and that the Earth only revolves around the Sun because everyone agrees that it does. Kind of sounds like the ultimate Humanism to me - all of reality only exists because we Humans agree that it does. Talk about making ourselves the center of Creation! Even religion acknowledges some authority higher than humans.

          Wikipedia is just the next step in the "democratization" of knowledge - the idea being that truth should only be what people agree upon. I guess it's considered 'elitist' to try and claim that there are truths outside of human acknowledgement.

          Carl Rapson
        • We almost agree, but not quite

          Having been a "professional" for a long time, I know that there is
          nothing really objective about edited work, which is simply the
          product of a smaller community. That's not bad, but believing it
          is objective is. At the same time, the groupthink you're pointing
          to in communities is very real, which is why my big suggestion
          in the last piece was to have many articles about one thing
          rather than one. This, at least, is an acknowledgement of the
          difficulty in trusting any one version of the truth.

          But, I do belief that if people who decide to do something like
          Wikipedia does?create a collaborative document?embrace what
          has been learned by professionals and academicians seeking to
          be as factual as possible, their "products" would be much
          improved. A degree, however, isn't required to be self-critical or
          to invite criticism rather than simply battle critics.

          Mitch Ratcliffe
        • Consider the following

          "True objectivity" is unachievable. And if you believe that the authoritative editors of Scientific American or Nature or any other lettered academic journal do not have emotional investment in what they edit, you are mistaken. No process will ever completely eliminate bias or create total "objective" accuracy, but the problem of precision is what is best addressed by the Wikipedia model.

          In other words - if 95% of the world really believes that the world is flat, that bias is going to be reflected in the Flat World Wikipedia article. But it's what 95% of the world believes anyway, so it's essentially the consensual reality that everyone agrees upon, right or wrong (objectively).

          However, the observations of 700,000 editors and authors who are vested in the Flat World article will create an emergent system that portrays a precise statistical normalization of those 700,000 views and gets continuously edited until either observations change or consensual reality shifts.

          What Wikipedia is, is a conversation. It's not a book - it's a continuous cycle of information improvement through constant communication.
          • The problem with consensual reality

            You are right, it is a conversation, but it is also a document
            because it is collected and retained for future reference. The
            conversation doesn't make it right and the demand people are
            offering to Wikipedia is that it reflect what is right; my
            suggestion is that it not default to a consensual reality but
            reflect contending views with unique entries, because that
            ensures minority and heretical views will be heard. John Stuart
            Mill explains the danger of a discussion the eliminates
            inconvenient or detested information through enforcement of
            consensual reality this way in On Liberty:

            "Every truth which men of narrow capacity are in earnest about,
            is sure to be asserted, inculcated, and in many ways even acted
            upon, as if no other truth existed in the world, or at all events
            none that could limit the first. I acknowledge that the tendency
            of all opinions to become sectarian is not cured by the freest
            discussion, but is often heightened and exacerbated thereby;
            the truth which ought to have been, but was not seen, being
            rejected all the more violently because proclaimed by persons
            regarded as opponents. But it is not on the impassioned
            partisan, it is on the calmer and more disinterested bystander,
            that this collision of opinions makes its salutory effect,. Not the
            violent conflict between parts of the truth, but the quiet
            suppression of half of it, is the formidable evil; there is always
            hope when people are forced to listen to both sides; it is when
            they attend to only one that errors harden into prejudices, and
            truth itself ceases to have the effect of truth, by being
            exaggerated into falsehood."

            So, simply saying "Wikipedia is a conversation, not a document
            (which is incorrect)" does not excuse Wikipedia from the
            responsibility as a thing that calls itself open and free and an
            encyclopedia?in addition to being a conversation?to see
            beyond the consensus and offer reader, Mill's disinterested
            bystander, access to all views, even those obscured by
            Mitch Ratcliffe
          • It's not really a conversation

            If one person can delete or modify another's comments (which is my understanding of how Wiki works), then it's not a conversation, any more than it would be if ZDNet users could modify or delete each others posts.

            Making Wikipedia a conversation appears to be what Mitch is talking about, but I don't think that's really feasible either, except by allowing a set of comments on the articles, which users other than the author cannot modify.

            Wikipedia can and should be part of the conversation, but it can't really be a conversation itself and maintain its character as a reference work (too complicated). Instead, the focus should be on improving the process to insure both accuracy and fairness. Not everyone will be completely happy with what comes out, but if the vast majority find it to be a useful and reliable reference (even if few entirely agree on all the details), then the project will have succeded (and not otherwise).
            John L. Ries
  • Elvis hovers around the building

    Great series and stimulating discussion.

    I have a question though. Where is the cut off point between supplying the differing lines of thought for inspection by the disinterested bystander and peddling falsehoods?

    More importantly - who makes the call and how?

    I realise this cuts to the heart of the transperancy issue that you mentioned in the first blog but I'd like a little more detail on how to implement it and still satisfy the crieria set out in THIS blog.

    Elvis is dead, right?
    Dave F_z
    • Unfortunately, Elvis is not dead

      Elvis keeps coming to my house to borrow Cheese-Its and never
      brings a fresh box back.

      Your question is the Big One. Ideally, the falsehoods get their
      say but people seeking reference data get the most accurate
      view of information rather than buried in alternate versions.
      Thinking in terms of user experience: If you have to go to
      Wikipedia for an article, then the problem is daunting, but if
      Wikipedia were just an API and others?people dedicated to
      being filters?were able to point to the articles they trust and
      display those views to visitors on their own sites, readers could
      choose filters to trust to give them the "correct" view of the
      knowledge in Wikipedia. Yes, there would be an ACLU view and
      an NRA view, which would be used by people who want to see
      the world through those organizations' agendas, but the full
      range of contending views would still be available, which is
      better than one view that promises objectivity it can't deliver.

      We've seen a lot of news-f and blog-iltering sites, so why not a
      filter on Wikipedia? Problem is, Wikipedia's built as a destination,
      not an API.

      Wikipedia isn't very Web 2.0, in other words.
      Mitch Ratcliffe
      • Elvis is an alien

        Thats why the Cheese-Its aren't replaced, they're used to coat the anal probes.

        What then of the disinterested bystander? Is there no place anymore to learn or gain new knowledge without approaching a (potential) source through existing filters?
        Has there ever been?
        I guess Wikipedia is at the forefront of the conflict between information and knowledge, relativism and the absolute.

        I fully take on board your comments about objectivity but......

        Doesn't Wikipedia run the real risk of catering to all and being trusted by none?

        Is there a place for some entries to be classified as either (say) [b]Confirmed[/b], [b]Unconfirmed[/b] or [b]Unknown[/b]

        Maybe Web 2.0 is too scary
        Dave F_z
  • Encyclopedias

    I do notice a problem, one which does seem to be more prevalent in the US than elsewhere.
    Americans like to be told what to think.
    And they seem to like to tell the rest of the owrld what ABC and CNN tell them.
    Read the average US newspaper - international news covers a quarter page but only if US troops are involved.
    It's not like that in the UK, Frnace, Germany, Australia, Japan...
    We all have good and impartial international news.
    Not the US - I've read about golf in St. Andrews, England - I mean - if your reporters can't get that right, what hope have they of informing anyone of anything of import?
    I'm not surprised the US makes foreign policy mistakes - US citizens know nothing about the reat of the world.
    US politicians appear to have a less than complete knowledge of world geography, let alone anything more.
    And it's clearly the same in the education system - most Americans got all their knowledge of Scotland from Braveheart (and don't realise that Mel Gibson was born in Australia, naturalised in the US and the film is entirely inaccurate - like most "historical" films made in the US - it's made to agree with idealised, incorrect ideas of the rest of the World)
  • Good way to put it

    "As a reference work, the encyclopedia strives for a single version of reality, even thought there is little doubt that on many issues there are many viable interpretations of events and facts."

    Good way to put it. I like to say a reference work should have a single editorial voice and that the good researcher understands this.

    But why not have the best of both worlds? Rather than the schizophrenia of multiple views, I propose this: Keep wikipedia exactly as it is to be used by those looking to enhance their general knowledge and who can accept the risks. But in addition, have a tab that reveals an edited version of the article. The edited version would be a professional reworking of the public page (fact checking, perspective, etc) to bring it to the quality and unity of a traditional encyclopedia. General users could comment on, but not change, the edited version.

    This would mean that wikipedia would need to suck it up and choose what kind of encyclopedia they want to be. And the practicalities of maintaining a staff would mean that only a fraction of articles could have an edited version. But such a solution would retain the benefits of the current model, but also give users the piece of mind they might now be lacking.