"Let’s Do the Time Warp, Googlenewsweek"

2008
Jul
2

I’ve gotten into the (bad) habit lately of doing much of my websurfing at the WebPortal2.0 site PopUrls. Only about half of the featured sites summarized on PopUrls’ long page are really worth watching, but it looks so cool, it makes me feel like a cooler Internetter just for using it And without Popurls, I might not have noticed the funny goings-on at Google News.


See all the shaded headlines from Newsweek? Here’s what I found when I clicked on them:
“A President in the Dock” subtitled ” History: What Rehnquist’s Take On 1868 Tells Us About What’s Ahead” and dated Jan 11, 1999. 1999? Rehnquist? It’s about the yet-to-happen impeachment trial of Bill Clinton! How did that end up among the current news?
“On the Go in His Golden Years” subtitled: “Even In Retirement, Nelson Mandela Finds Himself In The Thick Of Things, From Burundi To Bill Gates’s Jet”; looks like a nice story, but it’s dated Jan. 31, 2000. Huh?
“Sizing Up Sunscreen” dated June 28, 2007. Well, we’re getting closer.
“What the Court Should Have Said”, “Who Won The Election Remains Unresolved. But That Wasn’t What The Justices Were Asked To Decide.” dated Dec. 4, 2000. What is this? Newsweek’s Greatest Hits? Are they doing the web equivalent of a sitcom “clip show”?
“Spitzer Plans More NYSE Charges” WTF? It’s from May 20, 2004 and when he was New York Attorney General and suing the New York Stock Exchange’s Richard Grasso, a case the current NYAG Andrew Cuomo just dropped. Which helps explain why that particular old story got “re-currented”. Because it was about essentially the same topic as a genuinely current story. Let’s go to the Google News homepage and see what it shows us.

Well, here’s the Nelson Mandela story:

Those BBC and AME stories underneath are a lot more current, but the Google News algorithm picked the Newsweek oldie as the top source for “Nelson Mandela” news.

And here’s the Sunscreen story:

It was in first position when I first checked the page, but dropped to second before I got it in a screenshot.

And here’s another Newsweek story on top of the topic:

Can you guess the date of the Newsweek story? If you came within six months of May 5, 2004, give yourself a cookie!

Some of the Newsweek links on Google News had extra added oddness. Check out the third headlines for these two topics.

“Your Name Your Email Address Recipient’s Email Address”?!?
No, that was not the title or head for either article and they were both old stories related to the new stories.

So, how did all this old Newsweek content, all clearly dated where human editors can see, end up misidentified as new news by the Googlebot? Well, it all seemed to happen at about 11:00PM EDT, and that certainly could have been the time Newsweek did something to its site, either changing the permalinks for archival stories or changing enough of the format of the news pages to make Goog think it was new content, possibly both. Whatever it was, it screwed with Google News’ way of collecting news in a way that I didn’t think it was possible to screw with anymore.

Still, it was kind of fun to look back at the Greatest Hits of Newsweek for the past 10 years, and notice how a few of them might even have fooled human readers who didn’t pay attention to the dates (Sunscreen and Nelson Mandela, particularly)

Here’s a news flash: Google may not be evil, but it ain’t perfect either.

Your Turn...

You can use these HTML elements without blowing things up: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

AND NOW A WORD FROM SOMEBODY WHO MAY HAVE NEVER HEARD OF ME...