Tag Archives | Search Engines

Also New in Google Labs: Google News Timeline

Google News TimelineSoftware legend Andy Hertzfeld–one of the architects of the original Mac OS–works at Google now, and today the company unveiled one of his efforts: Google News Timeline, a Google Labs service for searching news and other information sources by timeframe. It’s still clearly an experimental work in progress, not a core Google service.  But being able to search the Web in a chronological fashion ranks high on the list of useful ideas that haven’t been fully implemented yet, and Hertzfeld’s work is a tantalizing step forward.

Timeline Search’s sources include the Google News Archive, newspapers such as the St. Petersburg Times and the Pittsburgh Post-Gazette; magazines such as  TIME (covers only), Popular Science, Vegetarian Times, and Baseball Digest; blogs, Wikipedia, sports scores, and more.  As the service’s name tells you, results are displayed in a timeline format, and you can scroll back and forth in time and view information by day, year, week, month, or decade. Here’s a search I did by decade for Iran, which gave me a revealing 10,000-foot view of news coverage of that country in ten-year chunks:

Google Timeline

Timeline Search also works for more specialized research. Here are results that show the Beatles’ music in the 1960s year-by-year, and some Red Sox scores from one particular week in 1980:

Google Timeline Search--Beatles

Google Timeline Search--Red Sox

In this first version, Timeline Search feels a bit like a search engine that requires you to use advanced mode to get much out of it–it’s taken me some fiddling and experimenting to make it work. For instance, the search field at the top is used for search keywords in some instances, and to specify news sources in others, and you can’t figure out what newspapers and magazines are available without searching for them by name. I hope that Google forges ahead both on refining the interface and adding more sources–all those books, magazines, and newspapers that the company is scanning are ideal fodder for time-oriented search. And I’m glad they have Andy Hertzfeld on the job. Here’s his own blog post on his brainchild.

4 comments

New in Google Labs: "Similar Images" Search

Google Similar Images Google’s Google Labs experimental playground has added a service I know I’ll use quite a bit: Google Similar Images, which lets you drill down into image search results to find images that resemble each other.

The service looks a lot like standard Google Image Search, except that results have a “Similar Images” link below them.

Google Image Search

Click on any of those links, and you get more images that resemble the one in question. Here, for instance, are results for three distinctlly different plane images: a childlike black-and-white line drawing, a photo of a big plane, and a photo of a little one:

Google Similar Images

Google Similar Images

Google Similar Images

Google says that it hasn’t created the similar-image connections for all the photos in its index, so they won’t show up for all searches. And it’s going to refine this capability more before it rolls it into standard Image Search. But it looks pretty darn cool to me already. Here’s an official Google blog post with some more info.

No comments

The Return of Jeeves

The Return of JeevesI’ve written before that Ask.com has been a search engine that’s skittered from one advertising message to another for years. Now it’s trying yet another approach: Its original one! Barry Schwartz of Search Engine Land brought to my attention the entertaining fact that Ask’s UK version has gone back to its first name, Ask Jeeves, and has brought back P.G. Wodehouse’s famous manservant (or a rough approximation thereof) as its spokesmascot.

Here’s what you now get at uk.Ask.com:

Ask Jeeves UK

Schwartz’s story quotes Ask.com as saying the sort of things you’d expect it to say: that this is a new-and-improved Jeeves (he’s computer-rendered!) representing a new-and-improved search engine, that Jeeves’ name recognition is extremely high in the UK, and that it has no plans to to bring him back in the states. All of which makes sense. But I kind of hope that he makes his return here, too–the original positioning of the site was the only one that was memorable and made intuitive sense, and I remain steadfast in the belief that it’s almost always a mistake to try and rename anything. If VW can bring back the Rabbit after all these years, there would be no shame in Jeeves getting his old gig back.

(Ask spokesman Nicholas Graham did tell Schwartz that going to AskJeeves.com gets you Ask.com with Jeeves hanging out on the home page–if I’d known that, I’d forgotten about it. Betcha the number of people who do that every day is much higher than zero.)

Also, is it just me, or does this latest version of Jeeves bear an uncanny resemblance to his ultimate boss, IAC chairman and CEO Barry Diller?

Jeeves and Barry Diller

3 comments

Microsoft's Search Engine Way Behind Google? Must be the Name!

Kumo LogoThe Wall Street Journal’s Nick Wingfield has a story up in which Microsoft’s Yusef Mehdi says that blind taste tests the company conducts show that consumers find the results from its Live Search search engine to be indistinguishable from Google results. But when Microsoft slaps the Google logo on Live Search results–in a sort of reversal of its Mojave Project prank–consumers like them better than when they know they’re from Microsoft.

Could be–although I use Live Search enough to know that I don’t find its results indistinguishable from Google’s, or as good. It’s unfair to judge a search engine based on one result, but here’s one that matters to me: If you google search Live Search for “technologizer”, its first result is our terms of service, and the Technologizer home page isn’t even on the first page of links. Every other major search engine manages to figure out that the single best result for “technologizer” is unquestionably www.technologizer.com.

That’s kind of emblematic of my experiences with Live Search, and it’s part of why I don’t go out of my way to use it. And that’s the challenge for Microsoft or any other company that wants to take on Google in search: They have to figure out how to convince consumers to go out of their way to use them.

Microsoft has a long history of changing the names of underperforming products (and sometimes of products that are doing just fine). I can’t remember an instance of the change being a clear improvement, and changing the names of things never makes them better. But in this case, a switch might be in order: Microsoft’s whole “Live” branding initiative has little traction, and it’s just confusing. The Journal’s story says that Microsoft plans to spend a lot of money promoting its search engine in the future, but has no revelations about a name switch to Kumo or Bing or anything else.

But even though either name might be an improvement on Live Search, I already have visions of consumers liking Kumo or Bing search results better when Microsoft tells them that the results are from Google…

3 comments

More Live Search Name Drama!

ZDnet’s Mary-Jo Foley is speculating that the new name for Microsoft’s Live Search might not be Kumo after all. Maybe it’ll be…Bing!

Bing

Of course, as Mary-Jo reminds us, rumor had it in the past that Microsoft was also considering a third new moniker, Hook. That could work, too…

Captain Hook

3 comments

Google Gets a Little More Refined

Google LogoThe two tweaks to Google Web search that the Office Google Blog announced this morning aren’t game-changers. Actually, you might not even notice ’em unless you’re looking. But they’re both worthwhile refinements that should make some searches go faster.

Tweak #1 is the more interesting of the two: Google’s related searches feature–which suggests queries related to the one you entered–now has a deeper semantic understanding of some concepts, and can therefore suggest additional searches that go beyond slight variations in wording.

Here are a couple of examples:

Google Results

googleresults-art

It’s not magic–in both of the examples above, the suggestions feel a bit random, and other queries I performed didn’t seem to benefit from the tweak at all. But it makes related searches more valuable–and it whets my appetite for the day when Google and other search engines are really smart about figuring out your queries. (What if a search for “famous scientists” could give you a neatly-organized grid with queries for dozens of distinguished scientific folk?)

As Google’s blog post points out, related searches appear at the bottom of search results (when they appear at all) and sometimes, but not always, at the top as well. Google is clearly a big believer in weaving together results in different ways depending on the query, and while the logic behind some of its decisions is obvious–like putting maps at the top of searches with a strong geographic angle–I’m still not sure if I understand the reasoning behind the inconsistent position of the related searches.

Tweak #2 is really subtle: When you enter long, wordy search queries, the snippets of text that show up in Google’s results are now longer in some cases, to provide more information and show more of the words you searched for in context. Such as in this example:

Google Results

It’s good to see Google continuing to polish up its most important service–even minor improvements to Web search will probably make more people more productive and happy than any number of Google Livelies, Mail Goggles, and pseudo-Undo features ever could. (I’m reminded of the famous story about Steve Jobs arguing that shaving ten seconds off the original Mac’s boot time, times fifty million users, was the equivalent of saving a dozen lives.)

And while I don’t think anyone would ever accuse Google of resting on its laurels, having those laurels does give it the luxury of worrying about small things as well as big ones. Other search engines, from Microsoft’s Live Search to Ask.com to upstarts like Cuil, will only steal meaningful numbers of Google dans if they offer obvious advantages over Google–ideally in the form of breakthrough features and great leaps forward that are very, very hard to come by. Only Google doesn’t have to worry about Google: If it just makes all the users it’s already got a bit happier, it’ll do just fine. And small steps like the ones it took today should help.

One comment

More About Microsoft's Kumo Search Engine, Sort of

Kumo LogoLiveSide.net has a screenshot of the home page of “Kumo,” the next-generation Microsoft search engine which is currently in internal use at the company, and which may or may not be called Kumo when it goes public. It looks a lot like the current Live Search home page, which is dominated by a big striking photograph with hotspots that take you to search results relating to the image. The LiveSide image jibes with my only personal exposure to Kumo, which happened last week when I met with a Microsoft exec who had it loaded up in his browser; there, too, it had the Live Search-style photo teaser.

We still don’t know much about Kumo, though–and with search engines even more than most things in the world of tech, it’s hard to form even preliminary impressions without a fair amount of hands-on time. The bottom line with Kumo or any other would-be Googlekiller will ultimately be whether it helps you find relevant information more quickly than Google. Which would require that it make dramatic strides over Live Search, which usually leaves me less than completely satisfied when I use it. (I’m willing to confront the possibility that I’m so comfortable with Google that its style of results, and my understanding of how to form queries that will get me what I want, influence my impression of other search engines–but even taking that into account, Live Search results usually feel less far smart and refined than Google ones. To me, at least.)

Can we all agree that everyone involved would be best served if the weird tango between Microsoft and Yahoo ended soon–either with a breakup or with marriage? Kumo may be a Japanese word for cloud, but until the question of whether Microsoft and Yahoo will work together on search is resolved, there’s a little gloomy raincloud lurking above Microsoft’s homegrown search efforts, such as Kumo. Or whatever it ends up being called.

3 comments

Yahoo Aims to Make Web Research Easier With Search Pad

YahooConfession time: I do almost all of my Web searching these days at one site, and its name isn’t Yahoo. But I’m intrigued by Yahoo’s Search Pad, a new feature that’s entering testing today. It looks a bit like the soon-to-be-defunct Google Notebook and other existing services, except Search Pad is designed to be smart enough to notice you’re engaging in Web research, whereupon it offers to help you collate sites, snippets of info, and notes.

Only certain Yahoo users are seeing the feature so far, not including me. So most of what I know about it, I know from this video produced by Yahoo:

[vodpod id=ExternalVideo.779285&w=425&h=350&fv=id%3D11859907%26vid%3D4423517%E2%8C%A9%3Den-us%26intl%3Dus%26thumbUrl%3Dhttp%253A%2F%2Fus.i1.yimg.com%2Fus.yimg.com%2Fp%2Fi%2Fbcst%2Fvideosearch%2F2888%2F79567711.jpeg%26embed%3D1]

I’m instinctively skeptical about software and services that think they’re smart enough to figure out what you’re doing and help–call it the “Clippy thinks I’m writing a letter” syndrome–and I’m not clear on whether Search Pad can be invoked manually. (Seems like it should: When I’m researching something on the Web, I’m usually keenly aware that I’m researching something on the Web.) But the video has whetted my appetite sufficiently that I’ll check back to see when Search Pad shows up for me…

No comments

Yahoo’s Inquisitor: Better Searching With Fewer Clicks

We users of Firefox (and the Firefox-based Flock) are spoiled: It’s easy to slip lazily into the assumption that every cool browser tool premieres as a Firefox add-on. So I managed to remain ignorant of Inquisitor, an interesting Safari plug-in that brings features for speeding up Web searching directly to the browser’s search box. But as of today, Inquisitor is also available in beta versions for Firefox and Internet Explorer. And if you use either of those browsers, it’s worth a gander.

Continue Reading →

No comments

Ask.com: A Search Engine in Search of Character

Observant readers may have noticed that my look at the new version of Ask.com contained no mentions of the fact that Ask used to be known as Ask Jeeves, pining for that old name, or clever butler references. That was intentional. There oughta be a statute of limitations on clichéd references to things which are no longer true about technology products and services. And it’s been two and a half years since Ask.com dumped Jeeves, so I figured it deserved to be judged on its current merits rather than obsolete branding.

But…

After I finished up that post, I happened across an article in the UK’s Guardian about a new Ask.com ad campaign that coincides with the update to the search engine, and it got me thinking. Ask is far from the largest search engine, but it may be the most heavily-advertised one–for years, it’s attempted to make inroads against Google in part through multiple barrages of TV spots. But Ask, which in its Ask Jeeves days at least had a distinct personality, leaps from advertising message to advertising message with abandon, always in search of a new way to differentiate itself from the crowd but never holding onto a message for long. After the jump, a retrospective.

Continue Reading →

11 comments