News org wikis or contextual archives?
January 31, 2009 § 10 Comments
This idea is not new. But it has been bubbling in my head for a few months: how to incorporate a wiki or wikis into our online content. We’ve tried a couple of wikis in the past couple years as part of our in-depth series but we struggled with how to allow contributions. We ended up only allowing comments/threads. Not what most people would consider a wiki.
About 2 years ago Amy Grahan wrote a column over at Poynter about the idea of news org wikis. She wrote:
Most notably, wikis can transcend the short attention span and fragmented view of issues and events inherent in traditional story-format reporting. With a wiki, no topic ever really “scrolls off the home page.” Wiki pages are forever active — even if they lie fallow for long stretches of time. And interested people can continue to watch and edit these pages indefinitely.”
In her post Amy asks for examples of wikis implemented by news orgs. As far as I can tell there were no examples given of implementation as I envision it.
Also in 2007, Paul Bradshaw of the Online Journalism Blog also wrote a piece on new org use of wikis. He mentioned a couple of attempts by mainstream media:
A number of experiments with wikis have already shown its potential to both reach out to a readership – and to fall flat on its face. An example of the latter was the LA Times ‘wikitorial’ – an editorial piece on the Iraq war which the newspaper allowed readers to edit. After only a day the newspaper had to pull the feature due to readers flooding the site with inappropriate material.
On the positive side, however, was Wired’s experiment with the form late last year, when they allowed readers to whip an unedited article about (yes) wiki technology into shape. Over 300 users made edits, with one interviewing a Harvard expert, and another suggesting a contact – and when one user complained about some quotes from an interviewee, the original journalist, Ryan Singel, posted his interview notes so that users could pick a better one.
Paul also posted a more in-depth examination of possible uses of wikis in news orgs. In that post he wrote:
Internally, wikis also allow news operations to coordinate and manage a complex story which involves a number of contributors. News organisations interested in transparency might also publish the wiki ‘live’ as it develops, so readers can view as it develops, and look at previous versions, while the discussion space which accompanies each entry also has the potential to create a productive dialogue with users.
This refers to the idea of using a wiki to develop a story online.
Typically a story is posted on a news org website which occasionally generates a list of related stories. Perhaps a photo or two are posted as well. If there is a video, that might be embedded into the story. And usually that is the end of context.
My wiki/archive idea would work like this:
A story posted on a new org website would also have a link to its wiki/archive page. The wiki/archive page would display all related content in a chronological thread (maybe utilizing some nifty AJAX coding so that you don’t have an endlessly scrolling experience.)
You’d get:
- related stories including any comments posted by the public
- video(s)
- multimedia including slideshows, etc.
- maps
- related tags
- a timeline
- annotations by editors and reporters
Some concerns:
- Online vandalism. Shane Richmond wrote a piece on this for the Telegraph UK. Shane wrote:
The trouble is Wikipedia, as ‘the public face of wikis’ is frequently derided for inaccuracies and vandalism. Will the mainstream media be able to surmount those problems?
…will the wiki dream be killed off through the fear of cyber vandals treating our news websites as virgin walls for virtual graffiti?
- The NY Times recently reported on Wikipedia incorporating a new system to keep those types of edits off. We could have a stringent registration process but I’m not sure that it would be a wiki at that point.
- Too much transparency. In my opinion this isn’t a reasonable concern. Strike-throughs and footnotes should be included in stories on news org websistes to alert the reader to errors or edits. Many news orgs are already doing this.
- Information overload. This feature may not be useful for many stories or even for many news org customers. But with a well designed interface it would be a great resource for community members to follow stories that have developed over months or years.
- Not many users know what wikis are. Paul Bradshaw sites some statistics:
Finally, one of the biggest disadvantages may be readers’ lack of awareness of what a wiki even is: only 2% of Internet users even know what a wiki is, although similar statistics were once applicable to blogs.
So maybe calling it a wiki is the wrong thing to do. Maybe it would be more precise to call it a contextual archive of news stories. Although I think incorporation wiki conventions such as public input via comments and edits (after a reasonable registration to preclude trolls) should be a big part of this feature.
What do you think are the pros and cons of something like this; for journalists and consumers of information? Are there news orgs already doing this?
new media notes 
I think the type of story/archive you’re proposing is an interesting one, Carlos. We’re experimenting with a wiki (one of the first newspapers to do so on a large scale since the LA Times event, I think) based not around a story but around a series of public policy issues and getting responses — it’s at http://policywiki.theglobeandmail.com. I’ve actually been pleasantly surprised (so far) how little vandalism there has been, and how much sincere effort to create quality content and contributions.
Hey Mathew it’s great to hear that you got great contribution without a lot of vandalism. We did a series on the Cost of Community in Walla Walla and created a wiki (although it wasn’t open to outside edits only comments and threads. And it was a struggle to keep troll-like behavior at bay as most commentors chose anonymity. Though we did get a few insightful comments. When did you guys launch the wiki?
We launched it January 15
Carlos,
The idea of a contextual archive is one that has appealed to many people. The problem is that it’s extremely difficult-to-impossible to implement. To my personal and direct knowledge, both the WashingtonPost.com and NYTimes.com tried, spending a lot of $$ in the process.
The nub of the challenge is in categorizing the articles so the links to and lists of related articles can be produced. If it is going to work with the news audience, the links must be based on the actual concept behind the article, not just tags about people and places. Equally important, the categorization by concept has to be much more precise than current technology allows. (Technically, this is described as precision and recall.) If it’s not precise, then viewers will instantly lose any trust they had to start with.
All of the ideas you mention (related stories, videos, maps, etc.) all depend on that categorization. If it isn’t good enough, the rest doesn’t matter.
PS: As I think you sense, this categorization challenge has nothing whatsoever to do with whether the results are published via wiki or not.
Thanks Terry,
The actual implementation of this is what has had my non-programmer mind spinning. Is see how this goes beyond tagging something “JoeSmith, wallawalla,” etc. Have you seen Zemanta? Not sure if this would be robust enough to latch on to a news org CMS.
I think this is why open source and collaboration on top of a common platform are so critical. It’s simple distribution of labor.
Also, I think OpenCalais might be the semantic engine you need to power the topics.
Carlos,
I don’t have the time or inclination to seriously investigate Zemanta in any detail. However, from reading their ‘About’ page, I don’t see that they are dealing with the underling categorization problem.
They analyze incoming user text, presumably converting that into a query that is then used to identify (hopefully) relevant related content.
It’s not a bad idea, as far as I can see. But it certainly doesn’t say anything to suggest that they are expecting to achieve serious levels of precision and recall.
Remember, the challenge for a newspaper is to get viewers to view closely related content from the newspaper’s own archives (or at least on their own site). Otherwise, there’s no economic multiplier factor.
Best,
Terry
@Daniel
OpenCalais is indeed focused on categorization. From my quick look, it appears that they focus on extracting more concrete entities rather than concepts. I have no idea what kind of precision/recall they achieve (but if it were unusually good, I’m sure they’d make that a prominent feature). I did a search for those terms, but didn’t come up with anything too promising.
BTW, as near as I can tell, OpenCalais isn’t open source. It’s offered as a service, free for small volumes, $$ otherwise. They do have a high end product that is hosted at the clients’ premises, but I’d hate to ask the $$ for that.
The company that the NYTimes and Wapo used was Teragram (which was acquired by SAS in March of last year). Their product is top of the line, and expensive. But to my knowledge it wasn’t good enough to produce the kind of links and such needed by the newspapers.
[…] Full story at this link… […]
I can point whorunsgov.com
“WaPo wiki site to focus on US politicians and government officials” via http://blogs.journalism.co.uk/editors/2009/01/23/wapo-wiki-site-to-focus-on-us-politicians-and-government-officials/