Why not link to sources?

A few minutes ago I saw Stephen Hawking trending on Twitter, clicked on the link, and found myself on the Twitter Search page, where the two top tweets from news organizations were these:

hawking search

HuffPo’s link goes to a brief story with no links to any sources. I see there’s a tiny AP symbol next to the dateline. Does this mean it’s an AP story? I guess so, but the AP symbol is not linked to anything. So I go to the AP site, look it up, and sure enough: it is an AP story. Here’s the second paragraph:

In an interview published Monday in The Guardian newspaper, the 69-year-old says the human brain is a like a computer that will stop working when its components fail.

No link to the Guardian story there, either. Or to anything.

So I go to the CBS News tweet, and find the shortlink leads to this story, where the second paragraph reads,

In an interview published in the Guardian, Hawking – author of the bestselling “A Brief History of Time” – said that when the brain ceases to function, that’s it.

Kudos to CBS for linking to a source, and especially for breaking ranks with other news outfits that only (or mostly) link to their own stuff. The NYTimes and the Washington Post are two familiar offenders, but not-linking and self-linking are the norms. (Less so for Guardian, which has always been much farther ahead of the curve than other major papers. Blogs at the papers, such as , link generously. But these are exceptions to the rules that seem to govern the paper’s ink-based sections.)

On the whole, mainstream media have had a passive-aggressive approach to the Web ever since they were first challenged by it, in the mid-’90s. Even now, in 2011, they’re still trying to shove the Web’s genie back in the old ink bottle. They do it with paywalls, with schemes to drag your eyes past pages and pages of advertising, and (perhaps worst of all) by leaving out hyperlinks. Never mind that the hyperlink is a perfect way to practice one of journalism’s prime responsibilities: citing sources. Or, by another verb, attriibuting.

Maybe they take too seriously ‘s “Hyperlinks subvert hierarchy” thesis (#7) in , and want to stay on (or crawl to the) top of whatever heaps they occupy.

The reasons I’ve usually heard for not linking, or for only linking to internal pages, is that the journal’s site “needs” to be “sticky,” to “drive traffic” past ads, and to maximize the time spent by readers on the site. (Nobody defends the tracking of readers.) Whatever the rationale, not-linking compromises an online journal’s editorial mission — especially if not-linking is policy and not just habit. (I think, for example, that with Fast Company it’s policy. For example, all the links in this story go to other Fast Company stories.)

So now I’m wondering if anybody has researched, or would be interested in researching, the practice of linking to sources by online journals — especially by mainstream news sites. Would this be a job for , I wonder? (I’ll bring it up with my friends there at the when I’m there tomorrow.)

[Later... ] see C.W. Anderson‘s comment below, which points to this Niemann piece by Jonathan Stray and this book by Joseph Turow and i (a colleague who will receive the Gene Burd Urban Journalism Research Prize for the Best Dissertation in Journalism Studies here in Boston on the 27th of this month).

Also see what Kevin Anderson writes here, and the comments below. Excellent conversation, all around.

40 comments

  1. Dave Winer’s avatar

    Actually I’ve had Times people link to stuff I’ve written, on Twitter, on a fairly regular basis, which seems like progress.

    Also — “People come back to places that send them away.”

    http://r2.ly/crj6

  2. Gavin Carr’s avatar

    In the hacking-a-workaround category, I’ve often thought this would make a great crowdsourcing project. Whoever first gets annoyed and motivated enough at the non-linkage tracks down the Guardian link and creates a special bookmark in some delicious-like system that has the offending page, the text to link (or an xpath or something), and the link target. On the user side a browser plugin (greasemonkey etc.) looks up the pages we’re visiting and inserts the missing links.

    You’d have to deal with trust issues and spam etc., but seems reasonably straightforward in theory. Anyone interested?

  3. Doc Searls’s avatar

    Hey, Dave. I’ve noticed a few outbound links now and then from the Times. agreed about sending away. Hey, it’s always worked for us, no?

    Gavin, sounds like a fun idea. I’m interested, but my plate is also beyond full. Still, write it up in a blog post or something and I’ll send people to it. That would be a better way of getting and measuring interest.

  4. Hanan Cohen’s avatar

    “Google/Bing/Yahoo news search subvert non linking”

    We don’t need to demand good behavior from news sites.

    We can find the sources without them.

  5. F’s avatar

    What about Google News when there are articles about something with no link to the original source and the original source is not included in Google News?

  6. Brent Logan’s avatar

    I used to like to link to news stories on my blog. Unfortunately, news stories are the links most likely to rot. Not only are news stories poor at linking to sources, they’re lousy at providing permalinks. YMMV.

  7. Juliette’s avatar

    Linking practices by news websites, that is precisely the topic of my PhD!
    As far as I know, the need for the journal to be sticky and not send readers away is becoming quite old-fashioned – at least in the Belgian media outlets I’ve visited so far.
    But still, external links to sources are scarce. So I’m currently trying to understand why by carrying out ethnographic observations in newsrooms, and I believe there are tons of different factors involved (e.g. technical, cultural, occupational, etc.)

  8. Doc Searls’s avatar

    F, I’ve been frustrated by the same thing with Google News, even though I think it’s the best at what it does.

    Brent, you’re right about link rot. AP used to be one of the worst offenders. So was Yahoo. I haven’t checked lately, though. Speaking of which…

    Juliette, that’s a perfect PhD research topic. I’m interested in hearing more about your observations. And I’d like to see some numbers (if anybody has them) shedding light on changing practices around linking.

  9. C.W. Anderson’s avatar

    Hi Doc,

    Wonderful thoughts on this topic, which periodically erupts on Twitter.

    In addition to Juliette, (whom I was lucky to meet two years ago in Boston and who is doing path-breaking work on this):

    I have written a paper on the topic, called “Textual Tunnel-Hops and Narrative Chutes-and-Ladders: The HTML Link as an Uncertain Object of Journalistic Evidence” … I actually presented it last year at the Harvard-Yale-MIT cyberscholars working group, which met at Berkman, and I would be happy to send you a copy.

    Mark Coddington (at UT-Austin and author of Nieman Lab’s wonderful weekly journalism news roundup) is working on a major paper on links and journalism.

    Jonathan Stray has also written about this and has also included numbers:
    http://www.niemanlab.org/2010/06/linking-by-the-numbers-how-news-organizations-are-using-links-or-not/

    Finally, there is a (by now, in the world of the internet) book, “The Hyperlinked Society,” which I believe your Berkman colleage Dave Weinberger contributed to.
    http://www.digitalculture.org/books/hyperlinked-society

    Anyway, I think Juliette and Mark (young, rising scholarly stars) will be doing the definitive work on this (though, in the world of academia, it may take a little while to reach the finished stage). I’m interested in the topic mostly just to help them, and to provide some larger context to their research. Happy to talk to you about any of this stuff.

    All the best, Chris

  10. Brian Boyer’s avatar

    The answer is mundane, at least for some news organizations (like ours).

    At the Chicago Tribune, workflows and CMSs are print-centric. In our newsroom, a reporter writes in Microsoft Word that’s got some fancy hooks to a publishing workflow. It goes to an editor, then copy, etc., and finally to the pagination system for flowing into the paper.

    Only after that process is complete does a web producer see the content. They’ve got so many things to wrangle that it would be unfair to expect the producer to read and grok each and every story published to the web to add links.

    When I got here a couple years ago, a fresh-faced web native, I assumed many of the similar ideas proposed above. “Why don’t they link?? It’s so *easy* to link!”

    I’m not saying this isn’t broken. It is terribly broken, but it’s the way things are. Until newspapers adopt web-first systems, we’re stuck.

  11. Terry Heaton’s avatar

    Doc, I run into this ALL the time with clients. They view it as an unnecessary step in their very busy world, because they can see no real benefit to themselves, which is what it’s all about. The only clients who practice it are those who do what we call “Continuous News” in blog format, because links allow brevity. Occasionally, I’ll run across an individual that I can encourage, but as an industry, media doesn’t get reciprocity.

  12. Mathew Ingram’s avatar

    To be fair to the New York Times and some other major papers, their CMS systems don’t make it easy for reporters or editors to include links — which means that someone else has to go back and add them later, which quite often just doesn’t happen. I agree that there’s a passive-aggressive attitude problem as well though.

  13. Michael Morisy’s avatar

    Hey Doc,

    I’ve noticed the same thing. I run a site that hosts a ton of primary source documents, MuckRock.com, and we have a lot of journalists who use us. They’ve slowly gotten better over the past year about linking, and I don’t think it’s out of malice but out of:

    * The unfamiliarity and slight fear of even basic HTML.
    * Bad culture, in that they’re just not used to Internet thinking.
    * Busy, busy, busy.

    The best way to change that culture is make it as dead easy as possible (I have a love/hate relationship with the sites that subvert copy and paste to include a URL) and provide incentives (we started automatically highlighting news coverage of “good sites” that link to us: http://bit.ly/lIEByS).

    I think that latter carrot approach also helps get them over the “we’re losing readers” since they know they’ll be also getting readers.

  14. C.W. Anderson’s avatar

    Brian,

    (apologies in advance; you’re going to be the recipient of a larger criticism I have never really fleshed out, because I’ve mostly had these conversations on Twitter. But this comment is not really about you or the Trib. per se … you just made the argument).

    I guess this is my issue with your above response: why not– instead of spending a ton of time and money coming up with paywall plans (which I acknowledge and understand the Chicago papers have not done yet), building lots of fancy data visualizations, and investing in computational journalism- haven’t legacy newspapers invested some of that money and time in making their workflows link friendly?

    At this point “linking”- as a technology, a style, and a web culture- have been around for nearly twenty years (!!). The fact that this is STILL an issue is almost worse than if newspapers papers said, “screw you, we don’t believe in this linking nonsense and we won’t do it!”

    But in the end, that *is* what papers are saying; “it’s just not our priority.” And that *so* violates the basic culture and architecture of the web … well, do you see where I’m going with this?

  15. Brian Boyer’s avatar

    @C.W. Agreed.

    My point is that the doers at newspapers are powerless. This is not some “we shouldn’t link out because then people won’t come back” problem, or even a “what’s a link?” problem.

    It’s definitely a priorities problem. but it’s also a big company problem. I’m not particularly confident that any big company in this industry will survive long unless they’re A) magic, see New York Times or B) willing to basically reboot their whole operation, see JRC.

  16. Erik Kubitschek’s avatar

    I’ll just toss this frustrating technique used by MSN on a regular basis, which is to link out to their own search engine (Bing). For example, a catchy title grabs your attention and the link takes you to the top search results for the particular keyword/phrase.
    I’m also interested in thoughts regarding advertising on video advertising. At first blush, it’s no big deal, just 15 seconds of potentially entertaining or useful advertising, but I’ve noticed the advertising being placed on video being streamed from YouTube, Vimeo etc…
    The websites are making money by selling advertising on the publicly posted content.

  17. Chris Amico’s avatar

    I’ll offer a third option to Brian’s: Go beyond the mothership.

    I don’t expect chicagotribune.com to be running a web-first, web-friendly CMS any time soon, but triblocal.com is running dozens of blogs on WordPress. The same is true of the New York Times’ blogs (which are more tied to the mothership) or NPR’s Argo network (which, in disclosure, I work on/with).

    The big, clunky news sites were always going to be hard to turn around. The workflow at most newspapers (think of all the small town dailies out there) basically hasn’t changed since reporters gained the ability to cut and paste in MS Word.

    But spinning off new properties may give us all a few more lifeboats.

  18. Alex Byers’s avatar

    Ditto Brian.

    It’s the reporters who are reading and citing the original material – but few reporters file using software that plays nicely with their organization’s web CMS. If there’s no good mechanism to get the data (in this case, a URL) from reporter to web producer, it’s tough for the producer to add that attribution or context.

    What we need is more newsrooms getting reporters to work in applications that integrate well with their websites.

  19. John Zhu’s avatar

    Definitely agree with Brian that the doers at newspapers are often powerless, and that this is more a CMS problem (and, to me, staffing problem), stemming from links not occupying a place of priority at the very top level of a news company, than a “we shouldn’t link out” mentality problem at the doer level.

    Here’s my experience as someone who worked at a small-to-medium-sized paper, in the early part of the 2000s before the bottom really fell out of most newspaper staffs: The CMS was centered around managing content in a print workflow, and the web only came in at the end of the night, when the stories were piped over to a web site queue (manually at first, and later automated), and it was the design/editing desk’s responsibility to assign the stories to various pages on the site. Imagine how motivated you’d be to do that, much less go through every story and add links to the appropriate sources, at the end of an 8-hour shift where you just banged out a couple editions on deadline. Fast forward 10 years, and a desk staff about 1/5 the size of the one I was on is now putting out not just their own newspaper, but also a sister paper, and doing the same web duties at the end of the night, oh and make sure you don’t go over your 8 hours b/c we can’t pay you OT. Those things don’t justify not linking out, but it certainly explains part of the problem — top decision makers in the company not prioritizing the web, thus making staffing and technological decisions that handcuff the staff.

    Like Brian, I think many big news companies need to reboot their whole operation to fix this problem it is intertwined with many other aspects of the company. Do news companies really have to deteriorate to the point where they literally have nothing to lose before trying such a radical reboot? Well, let’s not forget that JRC went bankrupt and was about as low as a company can be before it was willing to take the chances that it’s taking now.

  20. Anna Tarkov’s avatar

    As Brian (and many others have said) it really does come down to having a web-first CMS. Most news orgs, even ones very open to the web and a web ethos, just don’t have it. So they leave the linking up to web producers and like Brian said, many are too time-starved to be able to do that.

    This entire conversation reminds me of the crux of the issue: print and online are still seen as separate entities with separate staffs, separate priorities, separate resources, separate content and so much more. I wrote about this not too long ago: http://www.annatarkov.com/its-their-stuff-were-just-putting-it-online I think that until news orgs really commit to being web-first (which, it’s important to note, doesn’t mean “print last”), we’ll continue to see poor linking, poor design, poor U/I, etc., etc. on news sites.

  21. Hanan Cohen’s avatar

    Here’s an idea for a solution that came to my mind after thinking about this post.

    It solves the problems of silos and linkrot.

    We should have a new SEARCH: URI scheme.

    http://en.wikipedia.org/wiki/URI_scheme

    It works just like MAILTO:

    SEARCH:”doc searls” will open a browser windows with the default search engine of the browser searching for this string.

    SEARCH:”doc searls”@harvard.edu will do the same but narrow the search to this domain

    The creation of a new URI scheme involves writing an RFC and starting a process which I have no idea how to do.

    Maybe the Berkman Center will want to take this upon itself?

  22. steve white’s avatar

    you don’t necessarily need to go web first, they have these things invented before the web called footnotes and bibliographies, its not about linking its about newspaper upping their game with due respect to the reader.

  23. Doc Searls’s avatar

    Clever idea, Hanan. I’ll tweet it and bring it up on the ProjectVRM list. As for Berkman, what happens there tends to be projects that researchers take on — and most of them already know what they want to do (or are doing). I’ll think about who might be interested, but to get somebody to bite I think you’ll need to cast the net wider.

    What do others reading this think?

  24. KD’s avatar

    Perhaps I have a technologist’s blinkers on, but I think there is a rather simple technology fix for this, and I’m not so brash as to believe that nobody else would think of it, so my conclusion is that the newspapers actively oppose including links in their web content.

    Consider the workflow at a high level. I expect that there is one or just a few points at which the contents are formated for the print page. Similarly, I expect there is one or just a few points at which the contents are formated for the web. If that is wrong, then ignore the rest of this comment.

    Links could be included in the original text written by the reporters, tagged in some simple way. The reporter would submit the articles, they would be edited and such, with the links included. Then when the text is formated for print, the links would either disappear or be replaced by some alternate text. (I’m not sure which would be the better choice.) When the text is formatted for the web, the links are put into proper HTML form.

    I think that approach would require only small changes to the systems that newspapers use to manage their operations, not a drastic changeover to “web first” CMS, that several folks above say is necessary.

    Let me repeat that my assumption is that there are only a small number of points in the workflow where the text gets formatted for print and a similarly small number of points where it gets formatted for the web. If that is true, minor tweaking at those points should be all that is necessary to let reporters include links that don’t disturb the print edition, but survive to appear on the web edition. If that assumption is false, then I’m all wet and forget I said anything.

  25. Kyle McCluskey’s avatar

    While I believe Brian’s current take on the situation to be accurate, there are interim solutions we can undertake to bridge the time period between now and newsrooms being ‘web-first’.

    (In the interests of full disclosure, I should state that I am the architect of the Tribune CMS that B-squared references above).

    At Tribune, we’ve implemented an automated text-mining solution that generates links for our content. It is fair to point out that the links only point to our internal topic galleries and don’t yet provide as rich an edifying experience as a web producer (or even the journalist him/her self) linking to a related document. It does however hint at the technological approaches that could be available to us (a la the SEARCH: URI) to augment our content creation process until it follows a web-first philosophy.

    And I believe that time is approaching faster than many of you think, one of the first things that we had to do once we rolled out this auto-linking feature was to make it aware of our producer added links and not overwrite them.
    The hand-rolled links were ample. At Tribune, we’ve also worked hard to make it easy for producers and journalists to link their content (to wherever they want) and they are using it. I believe that they will continue to do so, in greater and greater numbers.

    These autogenerated links won’t replace human-curated content. As they become more sophisticated, however, they can provide an enriching supplemental experience. They can also have the side effect of increasing producer/newsroom excitement around linking and hopefully spurring adoption of the next generation of linking tools to be created.

    Web first newsroom, here we come!

  26. @boris_tweets’s avatar

    Doc, thanks so much for bringing this issue up. I too would be very interested in getting data on news organizations’ linking practices (Juliette, on compte sur toi!).

    I also very much agree with KD’s comments, and would love to see a new SEARCH: URI scheme!

  27. Eddie Gear’s avatar

    This only goes to show that even top level blogs and companies violate copyrights. There should be a system that due credit has to be given to the source of information. Their are several occasions where I’ve had bloggers steal my content pretty much as is. I hope that there is soon a system that would stop all this. I hope all the bloggers get to read about this and make sure they do not steal other content and if they do give credit as they should.

  28. Dominik Bjegović’s avatar

    Is there a way you can track on Twitter great authors and their educational posts (I am asking about some filter that would allow you to quickly sort through all the mess and get into the juice).

    Best,
    Dominik Bjegović

Comments are now closed.