The content map and corrections

I think there is an elegantly simple solution to the problem of attaching corrections to earlier errors in news: It’s the link, the tag, and the content map.

There has been a great deal of discussion, following NY Times ombudsman Clark Hoyt’s column on errors, regarding what to do about old, incorrect articles on a subject that come up higher in search results than newer, corrected articles. Suggested solutions range from killing the old articles, which Hoyt considers, to correcting them to relying on the web and search. I agree most with that last solution, which comes from taguru David Weinberger. Nicholas Carr gets it exactly backwards when he says that search-engine optimization of article archives manipulates history and so old articles should be killed to make the web forget; that would be the criminal manipulation of history. [See correction below – ed] Weinberger says, for example, that if the Times killed all of Judy Miller’s incorrect stories we would be left without an understanding of the paper’s role in the Iraq invasion. I would follow the ethic of the correction I have learned in the blog world, a standard that requires openness and transparency (that is, admitting our errors as we correct them — quickly).

I say we can use the architecture of the web to fix errors and follow the ethic of the open correction, using those existing tools I listed above. Consider the case of a Times reporter writing an article that follows up on and corrects an earlier article. You can bet that the reporter writing the later story looked up the prior art; we are all trained to check the clip file. So there is likely to be knowledge of the conflict. In this case, here’s how the two can be connected:

* The reporter or editor can link to the old, incorrect article. The web site can then sense any internal links to the original article and display those links on it. If you find the wrong article in a search, you can see that there is a follow-up. Indeed, that follow-up could be labeled “correction” to make it apparent. And the Times site could display anything with the “correction” tag separately and prominently.

* Even if the two are not explicitly linked, they can be connected with tags. If reporters and editors both tag their stories about the subjects, they can be connected.

* Say they aren’t tagged. Their shared topicality can still be sensed. I don’t mean this to be a plug for Daylife, but finding such connections is turning out to be one of the great values of analyzing the body of news, inside one site or across all.

* Now let’s say the correction does not come from the paper that reported the error but from without. Let’s say that here, on Buzzmachine, I write a correction about a Times article. I could link to it and use the tag “correction” and that would then be discoverable (“‘show me all links to this article tagged ‘correction'”). I’d argue that the Times should display such links. But if they don’t, I’ll suggest that Craig Silverman could make a service of this at Regret the Error.

* And let’s say this isn’t about an explicit correction but instead about followups and more information. This is why I want to see the map of content and all its interrelations.

* Now if you want to get really ambitious, it’d be great if I could subscribe to old articles I’d read or written about so I could be alerted if there are any corrections, an idea I talked about last year. I could easily see becoming inundated with corrections but I think there’s a way to prioritize them.

But now pull back to the simplest level: If the Times linked to and tagged articles and exposed the links among them, many of the problems Hoyt et al wrote about would be fixed.

: LATER: I spoke with Times reporter Abby Goodnough at length about this and more for her Week in Review piece today about rumors that do and don’t get traction in media and blogs. It’s also possible that this content map could affect stories as they develop, linking half-baked reports with later reporting and then complete stories and then followups.

: CORRECTION: Nicholas Carr in the comments corrects me: He did not call for killing articles. I got that wrong and apologize. We still disagree about who’s manipulating history. But we don’t disagree about maintaining history. Sorry. This is what Carr said:

So if we are programming the Web to remember, should we also be programming it to forget – not by expunging information, but by encouraging certain information to drift, so to speak, to the back of the Web’s mind?

Though he explicitly said that information should not be expunged, I misinterpreted — and actually still don’t understand — what he means about letting information drift. Expunge or hide, I’d still argue that linking is best.

  • Tom Foremski

    Corrections, and more, could be handled by a “right to respond” which should be a fundamental right of the internet. A right to respond is a link next to a search result, sometimes activated sometimes not.

  • Simon Dickson

    I’m not sure tags are the model to adopt: the notion of showing up links between articles is almost a perfect fit for blogging’s Trackback concept. Although it’s been largely discredited due to spam, there’s no reason it couldn’t be successfully managed within an internal operation, especially with some kind of ‘local trackbacks only’ filter. Since the Trackback code is ‘out there’, it should be relatively easy to implement… or alternatively, it’s another good reason to base your news website on a blog platform.

    Of course, it’s some way off your vision of a content map, Jeff: but I’m gradually abandoning my natural purism in favour of a ‘near enough is good enough’ strategy. A guaranteed 80% fit today is better than a potential 100% fit next year.

  • Charles

    No mention of the Guardian’s system, which puts any corrections prominent at the *top* of the story that’s being corrected?

    This has two benefits:
    -original story is preserved, in all its wrongness.
    -original story is updated, in situ
    -web server will tell crawlers from search engines that there’s been an update on story, so indexes will get new correction
    -stuff that points to the original story pointing out its wrongness now points to a story that admits its wrongness.

    No need for fancy-schmancy spam-liable trackbacks or anything. Nor hard-to-define tags. (Does Judith Miller’s stuff now get tagged “misinformed” or “plain old wrong” or “thought it was right at the time”, which could apply to almost anything, including the weather forecast.)

    Sometimes when you need to put a nail in, a hammer does it best – not a 5’x3′ Perkin-Elmer nail insertion platform with added bunnies. Which is what everyone seems to be suggesting.

  • Pingback: Make Them Accountable / Media

  • Nick Carr

    Re: “Nicholas Carr gets it exactly backwards when he says that search-engine optimization of article archives manipulates history and so old articles should be killed to make the web forget; that would be the criminal manipulation of history.”

    Speaking of corrections, may I correct this hysterical distortion of my words, even if it doesn’t quite rise to – gasp! – the criminal manipulation of history? (I need to correct it here because, alas, most links go unfollowed.) What I wrote, in part, was this:

    “With search engine optimization – or SEO, as it’s commonly known – news organizations and other companies are actively manipulating the Web’s memory. They’re programming the Web to ‘remember’ stuff that might otherwise have become obscure by becoming harder to find. So if we are programming the Web to remember, should we also be programming it to forget – not by expunging information, but by encouraging certain information to drift, so to speak, to the back of the Web’s mind?”

    You’ll note that there’s no mention of the killing of any old articles. In fact, in posing my final question, which is just a question, I specifically warn against expunging articles.

  • robertdfeinman

    If SEO works (that is the information provider can manipulate the outcome of a search) then it just means that the search engine services are not doing a good enough job.

    Perhaps they have gotten complacent since Google makes billions while providing only mediocre search results. Librarians will tell you that many people are content with 10-20 references regardless of whether they are the best (or even relevant).

    I’m sure everyone has had the experience of searching for something and being presented with a lot of irrelevant clutter. Do-it-yourself searching also means that the expertise of those who find information for others as a profession aren’t involved, so each person is left to their own level of ability.

    A good search engine would provide the original story, the corrections, and the comments, and present them in an integrated fashion. We shouldn’t have to depend upon the policies of the organization that published the original story for this to happen.

    I’ve written an essay on how the big three in the search engine business are controlling what is found and stifling competitive improvement. It’s on my web site, for those interested.

  • Mindy McAdams

    There’s a good example of the “embed the correction in the archived story” practice in a New York Times story from 2000 — in which a nonexistent person was quoted in a NYT print story (“For the New College B.M.O.C., ‘M’ Is for ‘Machine’ ” – Aug. 10, 2000 – By Lisa Guernsey). You can see it on the third page in the online-archive version of the story — if you have Times Select.

    I use a short PowerPoint (available at ) to impress this example on students. If you don’t have Times Select, you can download it.

    However, the last time I checked, the correction does not exit in the Lexis-Nexis archive version of the 2000 story.

  • Eric Gauvin

    I agree with Robert Feinman:

    “A good search engine would provide the original story, the corrections, and the comments, and present them in an integrated fashion. We shouldn’t have to depend upon the policies of the organization that published the original story for this to happen.”

    Also, I think we’ve really fallen under Google’s spell if we put that much value on and faith in how Google ends up ranking results. It’s a search engine—not a crystal ball. I often feel there’s something freakish or strange about the first few results and I view them with skepticism—in the very least wondering simply, “why is this the first result?”

    As for Jeff Jarvis’ “content map,” I still don’t understand what that is.

  • Henk Blanken

    To make things just a little more complicated: what should we do when a person, who we wrote about years ago, now asks us as journalists to make the net forget that old story. I know of a case like that in The Netherlands. Nothing wrong with the old story. The facts were ok, but this person changed over the years and doesn’t want to be bothered by his own past. In a way he has a point: when he agreed to be a part of that story years ago – the journalist did an interview with him – he could not foresee that this story would end up on the net, being the top result when you google his name. How do we decide between his private interest and the public interest of not manipulating history?

  • Tish Grier

    Jeff…to implement a blogger style ethic of correction would take staff, time, and money–three resources that, from all the hoo-ha, we know newspapers are having a bit of trouble with at the present time.

    Until newspapers have a better understanding how corrections work out here, and are willing to alocate resources to it, the problem will continue.

    That is, unless Google’s new article comments feature catches fire–then Google would simply become the Paper of Record without ever having done any original reporting. Ironic, eh?

    As for “drift”–I kind of get what Carr’s talking about (I think.) If you search your name on a regular basis, you might see certain content or blog post “drift” from the top spot to the bottom. You might also see different content under particular urls or misspellings of your name. I’ve seen this with google results for my name and blog over the past year (which has been a year of rapid change for me and my content.) So, content does indeed move around and “drift” in search–but the question of how to get it to drift down or be superseded by better, more relavent content is something of a mystery (at least I haven’t figured that out for myself just yet.) Maybe it’s a matter of super killer SEO. Then again, who really knows the mysterious mind of Google’s algorithm enough to “trick” it to put the good stuff first?

  • bert

    On a related note, what happened to the post about One step forward, two steps back” you posted? It’s still in your RSS feed, but read more leads to a 404.

    Anyway, to answer that question in that post, here’s the RSS feed for that blog: (Safari showed it automatically.)

  • Jeff Jarvis

    I posted that accidentally after I’d found the feeds. Thanks.

  • Pingback: links for 2007-09-06 | SOJo: Student of Online Journalism

  • I.F. Stoner

    From Day One Bloomberg’s correction policy is as follows
    1) the orginal erroneous story is deleted;
    2) it is replaced with a corrected version);
    3) the corrected version ALWAYS has a trashline showing what was corrected;
    4) failure to follow above can be a fireable offense.

    Its about transparancy, and that means holding yourself as accountable as those whom you cover…

  • Pingback: BuzzMachine » Blog Archive » Moore’s Law, as amended

  • Pingback: Salvatore

  • Tanika Nizova

    Actually, Jeff, you raised a very complicated issue.
    I think that killing an old content is not a suitable way out. Just tagging, links, and the date of posting on the visible place will help.
    Good luck in your quest for convenience and relevance! :)

  • rickystephen123

    I was extremely happy to locate this
    website. I wanted to thank you with regard to this terrific read!! I certainly
    appreciated every little bit of it.

    best web design company

  • Imenso Software

    Always follow the right way :)