By Scott M. Fulton, III, Betanews
If you have a completely new search engine -- in other words, one that's not a renamed version of Windows Live Search -- you need to give it a niche that somehow emphasizes the quality of its results compared to those from Google. Wolfram Alpha's niche of choice is the intelligence of its results, in an effort to wring the educational power out of the verbal sponge that is the Internet. So that slot's taken for now.
Enter Collecta, the product of former AOL search chief Gerry Campbell, and an indicator of what AOL could have accomplished had its previous leadership chosen to invest in ingenuity. Launched last Thursday in public beta, the ideal of Collecta is that it searches content that tends to be updated quickly and frequently, and that it conducts those searches on the fly -- it's truly searching for what you've asked it to search for, rather than look up results from a massive index.
Although you have to wait several seconds for a response from Collecta (a lifetime for users stuck in their parents' basements), the results that come up from Collecta are merely seconds old. In a spur-of-the-moment Betanews test this morning, about an hour after Apple CEO Steve Jobs was cited in a statement regarding iPhone 3G S sales, we wanted to see how soon Collecta would yield results from folks noticing Jobs' apparent return to some degree of health.
Our query to both search engines was "Steve Jobs" statement. After a countdown of 22 seconds (Collecta provides you with its own clock to let you know its ticker is as healthy as Steve Jobs'), Collecta retrieved a complete list of entries pertaining to the hour-old statement, and kept adding to that list over successive minutes until we had about three dozen in a stockpile. By comparison, a search of Google News claimed to have 1,208 pattern match results, though a check of the list itself revealed zero stories actually dealing with the hour-old statement.
So there's something to this Collecta, evidently. But if Google can inflate its own assessment of accuracy by an infinite percentage, what could Collecta get away with? It took about 90 minutes for press accounts of the story to make Yahoo News. Prior to that time, checking the initial list of Collecta's results, those which floated to the top were mostly Twitter feeds. And quite a few of them, by virtue of the fact that Twitter can often become an echo chamber (talk about infinite inflation of accuracy) were actually the same tweet repeated a handful of times.
But those tweets did include a link that led to the actual story, so anyone using Collecta to verify what she may have thought to be a rumor about Jobs' health, would have had that rumor verified in under a minute. The results themselves, however, speak to Collecta's choice of "news" providers.
Conceivably, Collecta should have been able to locate the page of information which all those tweeters were echoing to one another, prior to, or at least in tandem with, the tweeters having located it. So how old was the news, according to Collecta's own dating, by the time it was being repeated? Timestamps show those tweets came one hour after Apple released its statement at 8:30 am EDT this morning, though it was on PR Newswire at 8:30 on the dot. (Anyone familiar with Apple would know that if it's going to release a statement to the press, it will be at 8:30 -- not 8:00, not 9:00.)
Going down the list of "Older Results" in Collecta's middle pane, we didn't find any links whatsoever to PR Newswire's posts, which are carried on places like CNNMoney.com and Betanews. So while Collecta had the early scoop on the second-hand information, neither it nor Google led their users directly to the first-hand news.
Getting first-hand news is getting to be a big problem for folks who'd like to become journalists when they grow up (as noted by blogger John Gruber this morning, one of the first to realize The Wall Street Journal's "scoop" story over the weekend on Jobs' liver transplant was not sourced.) As journalists this week give more credit to Twitter for becoming "the network of record," to borrow an old phrase from Bernard Shaw, for the latest Iranian Revolution, perhaps Collecta could become the search engine of choice for folks looking for real-time news from the Middle East. We decided to test that theory.
Next: Calling CQ Tehran...
With the Iran government locking down principal sources of information emerging from the country's borders, the unofficial channels are becoming flooded with information from everyday citizens everywhere chronicling the changes to their country from their perspective. So a search anywhere for Iran election could provide you with reports or correspondence having to do with any number of events triggered by the election itself.
Suppose you're specifically interested in the status or well-being of the opposition leader, Mir Hossein Mousavi, the subject of our next test query. This is a touchy subject, especially for Iran where knowing where a popular candidate is located is not always a good thing.
In a chain of events that is literally unfolding by the minute, it is astounding to find that Google News' latest entry on Mousavi is dated last Tuesday -- six days ago. There's some good biographical information in there for folks who are just tuning into the story, but nothing for someone who would like an update on whether the fellow is still alive and safe. In fact, what you'd see here is no newer than what you might find in your dentist office's waiting room.
By comparison, while Collecta's first dozen or so tweets on the subject were the same one -- moreover, a link to a biographical piece from several days ago -- down the list was a blog post that linked to an Associated Press video dated today, released on the AP's YouTube channel. There's Mousavi in public, standing before a crowd, promising protesters of the certified election results that he will stand by them.
So on the one hand, Collecta beat Google in pointing Internet users toward the source of the information they're looking for. On the other hand, it's a very roundabout way of getting this information: a link to a personal blog whose publisher has some very outspoken viewpoints on some subjects, so not exactly a "fair and balanced" source. But it included the video.
However, as was the case with the Steve Jobs news, anyone who truly knew where to find such news would have already located it. The AP video channel is right there on YouTube (ironically, owned by Google), and it's providing the latest videos from wherever professional and amateur journalists have been able to grab them, with camcorders or cell phones or sketch pads. What's more, there are hundreds of Twitter feeds currently operating now with information of some sort emerging from Tehran and elsewhere, none of which were featured on either Collecta or Google. In fairness to Google, a search of Google Blogs did pull up posts dated a few hours ago about Mousavi, but those posts appeared to be links to other sources (e.g., Fox News) whose video was already more than a week old.
Right now, if there is any battle between Collecta and Google, it's in the bluster department: Google touts numbers like "4,906 results" that appear to be drawn at random, while Collecta provides a dozen or so very recent results whose quality or accuracy or reliability is not what it could be.
Copyright Betanews, Inc. 2009