Google Universal Search looks impressive and it brings them in line with recent upgrades made to AOL and A9. But perhaps the more interesting step forward will be their 12-language Cross Language Information Retrieval. I already see many referrals from Google translation services. I am guessing I'll see many more in the future as the the CLIR initiative comes online.

• Posted by: Michael Martinez
• May 16, 2007 2:51 PM

John, when you write about this, I would like to point out one thing that stands out to me, one thing you may want to keep in mind. In this "Universal Search", at least as Google presented it today, it is not just a matter of Google automatically displaying the results from the set of most relevant data genres/verticals. Google also introduced the notion of a context-aware navigational toolbar, by which users can offer explicit refinements of the search.

Thus if Google automatically determines that the best 3-4 verticals for your search are blogs, news, and images, you can disagree, and select "video" instead. Or you can narrow your search results, and select "images", thereby choosing only that particular cluster of results.

To me, what is significant about this announcement is not the idea of Universal Search, per se. It is the idea of allowing users manual, explicit refinement of the automatically-generated system results. This latter notion is a very new direction for Google, and one that they heretofore seem to have studiously avoided. "Users are too lazy to do refinements", is the mantra I have often heard.

This has now changed, officially, and I think it is a big deal. I will now also officially eat my own words, as I have been griping on this blog for a few years now about Google's failure to do things like this.

Nice job, Google. Let's keep it coming.

• Posted by: JG
• May 16, 2007 2:59 PM

Let me know when they add my history and google browser sync information as well. Then they can call it "universal". And target my ads accordingly. No more mortgage refi ads!

• Posted by: Derek Tumolo
• May 16, 2007 4:22 PM

I do not know what it becomes later, for now as I could see it on google.com, it is somewhat a copycat feature of AskX, A9, and even Live search. The last two engines already allow users to add/remove buttons for other verticles. The first engine senses the approriate verticals as Google does.

There is nothing wrong in copying and improving. But when an independent observer like you mention, proper credit should be attributed too. This so called radical change is competition in action. It can force even more radical changes than this one. Even the search refinements, I first saw on Ask.

Disclaimer: The commentator works for Microsoft, the parent company of Live search. The opinion is his personal opinion.

• Posted by: Kamal Jain
• May 17, 2007 2:45 AM

I´m not sure if I like the new universal search. The search entry result pages have already been fed up with many useless results, now it´s probably getting more confusing. I´ve also read that Google plans to use video and graphical advertisements on the SERPs - that would be harmful to me, no thanks.

• Posted by: Feinkost
• May 17, 2007 9:36 AM

John, you said "Google will, because it's the market leader, get credit for what others have already done."

Does not seem to be the case in some other industries. In fact it is quite the opposite.

Even in this case, there is only a small fraction of people giving credit to Google. I read a BBC article and even they mentioned that others have already done it. Though, it is significant that Google decided to follow.

• Posted by: Kamal Jain
• May 17, 2007 8:29 PM

Google will, because it's the market leader, get credit for what others have already done. Jim Lanzone, you now know what it feels like to be Steve Jobs when Windows came out.....

You mean how PARC felt, not how Steve Jobs felt, right John? ;-)

But sadly, I think what you say here is probably true, about Google getting credit the same way MS Windows did. But there is one crucial difference: Before Windows came out, I don't think Microsoft strenuously argued against Apple, saying that windowing operating systems took too much time to load and users did not want to wait that long, that "users were too lazy" to arrange their desktop icons, that the command-line style of a DOS interface was much cleaner and well-designed than all the messy overlapping graphical boxes. I remember Unix geeks making those arguments, but I don't think Microsoft did.

Microsoft might have been playing catchup, and not innovating internally. But at least it had the self-consistency of not being philosophically opposed to what Apple (and PARC) were doing.

[snip - I am cutting this post into two parts, because your blog is rejecting it]

• Posted by: JG
• May 17, 2007 10:42 PM

[Part Two]

Google, on the other hand, is not only playing catchup, but is changing its philosophy to do so. For many years Google's minimalist interface was purposely, almost religiously, DOSy/Unixy in its lack of GUI. Yet there are now loads of colorful, bubbly widgets on the Google personalized homepage. Google used to say that, in order to achieve its goal of organizing the world's information, it would not do chat and calendars. Now, it claims that chat and calendars are essential parts of organizing the world's information. And whereas for many years Google has argued vehemently that users are too lazy to use query refinement tools, Google is now offering things like timeline clustering of results, to help users quickly navigate through hundreds and hundreds of documents at once.

I for one welcome our new philosophically-challenged overlords, and am glad that they finally see what Vivisimo and Ask and others have seen for years. I welcome it because I think more competition is better. I pretty much stoppped using Google on any sort of regular basis about a year ago because they were not offering the things I wanted. Now that they offer things like timeline clustering, I might actually start trying their search again. I have to say: It is about time that Google has come around. I thank all the other search engines for showing that users are not lazy, that we actually yearn for many of these improvements. I hope to see more from Ask et al as well as from Google. Remember, even with Windows, MacOS is far from dead.

And let's not forget that, just like Steve Jobs was also not the first person to come up with windowed operating systems, search engines over the past ten years were not the first to come up with query refinements based on clustering. In fact, the cluster hypothesis itself, the notion that relevant documents tend to cluster, goes back to Van Rijsbergen in 1979. Sometimes innovation is only innovation if you have a short memory.

Just my $0.02.

• Posted by: JG
• May 17, 2007 10:43 PM