While today's new social features are limited to the Spotlight section, it adds another way in which Google News can personalize content for logged-in users using their social data. Google is rolling out these kinds of Google+ features across all its Web properties.

Google Web search has treated public G+ posts as search results since soon after the social network launched. Google is insisting upon making its new social layer a pervasive, personalized filter for the whole Google experience.

It's all part of an effort to redefine relevance in the way Google crawls the Web. Instead of brute rankings of the Web's content, Google has decided that personal, real-time recommendations are more relevant to us. Do you agree?

In addition to the algorithm itself, Google has changed features of the search interface recently. It eliminated the "Timeline" view of results to organize them by date range, and it has integrated Google+ social content in a variety of ways.

Google continues to improve rich snippets and learn how to pull relevant page content into search results. Recent updates make Google smarter about pulling page body content, rather than header or menu content. Others improve page titles by de-duplicating anchor text in links, and improve details in rich snippets for applications.

Google has also retired a signal for image search that looked for images referred to by multiple documents around the Web. Another change improves detection of which pages are "official" for a topic or brand.

Changes To Time-Sensitive Results

Google is making a concerted effort to shift from basic chronological results to real-time search. Its recent updates to the Caffeine search infrastructure semantically determine when a user would want recent, "fresh" results instead of the all-time ranked pages. For example, it may determine that users searching for "olympics" are more likely to want results about the upcoming 2012 Summer Olympics than the Wikipedia page for the Olympic Games.

By eliminating the "Timeline" view and applying "freshness" adjustments to queries with specified date ranges, Google is pushing timeliness as a new priority in how it determines relevance.

Google is also improving "freshness" signals by incorporating Google+ activity into overall search. After its real-time search deal with Twitter expired this year, it needs new signals for what's currently trending. Google+ offers just such an opportunity, and Google is trying out real-time search within the social network.

Other Search Changes

Several recent tweaks to Google search improve cross-language results, using Google's powerful translation to retrieve content for searches in languages that have limited Web content. Another improves query auto-completion in Russian, which used to produce some arbitrary and unhelpful predictions.

To see the rest of Google's bullet-point search improvements, visit the Inside Search blog.

Do you think Google search is improving over time?

Timeline view is still available in Google News, but it only searches certain archived publications instead of all Web results. Google community managers have suggested the normal date range filter as an alternative, but this isn't a browsable feature like Timeline was. Just as it has done with Google Reader in recent weeks, Google has killed off a feature used by a small but dedicated set of its users.

In response to user complaints about the disappearance of Timeline, Google search community manager Kelly Fee suggests using google.com/trends or google.com/insights/search for graphs of search results over time, but those tools only go back to 2004, and they aren't a part of Web search. She also suggests the Google Books ngram viewer, but that's only for book searches.

Google's Going Real-Time

The end of Timeline coincides with its implementation of new real-time search algorithms that privilege recent results over old ones by assuming when users want current information. It's also experimenting with real-time search on Google+, and it's surfacing recent posts from the social network in Web search. The removal of Timeline pushes users of Google search away from historical content and toward real-time results.

Google can do whatever it wants with its free services, and it doesn't have to explain itself to anyone. And pushing around smaller products like Google Reader in the interest of Google+ is a sensible business move. But taking away useful features of Google search raises a more core issue.

As Vic Gundotra has made clear, the + part of Google is integral to Google itself. Is it starting to change the company's priorities? Google's mission has always been to organize the Web's information. Is its new social experiment in real-time trends compromising that?

Timeline search screenshot credit: Digital Marketing Rucksack

Sites that benefited from the update tend to be content sites and brand sites with frequent updates. Many are news sites, but others are travel sites or other consumer sites. The sites that lost in this update are more of a grab bag. Many of them are government websites or less time-sensitive news organizations.

Searchmetrics measures SEO visibility of sites across the Web. Its results are adjusted to eliminate outliers in either relative or absolute terms. In the Caffeine update, the biggest winner was last.fm, which shares the music its users are listening to right now. It gained in SEO visibility by 12.56%. Comcast.net lost the most ground, falling 23.02%.

Caffeine winners:

Caffeine losers:

See the Searchmetrics blog for the full list.

Google Search Going Real-Time

Google's Caffeine update is its latest effort to gather its own real-time search signals after its 2009 deal with Twitter expired. With social data from Google+, Google now has a range of ways to determine timeliness and relevance of Web content. It offers true real-time search within Google+, and it surfaces Google+ posts in Web search results.

Do you find Google's adjustments to the timeliness of search results helpful, or do you prefer absolute page rank?

With Google+ indexed in Web search and providing real-time search data, Google now has strong signals for timeliness and relevance. By tweaking search algorithms on one side and gathering social data on the other, Google is working towards a clearer of picture of what's happening on the Web this instant.

• Recent events or hot topics like breaking news - stories that might be only minutes old.
• Regularly recurring events like elections, conferences or sports games.
• Frequently updated info like product reviews.

Google Going Real-Time

As illustrated by the above graphic from the 2010 announcement of Caffeine, Google wants to get away from the boring stack vision of page ranking and smartly determine what kinds of results are relevant for each query. That requires Google to index in real time. Caffeine accomplishes this on the crawling side, and Google is also working on publishing systems and networking protocols to alert the search index of new content faster.

But social networks lead users to do the real-time publishing and alerting work themselves, making Google, Twitter and Facebook's jobs of surfacing relevant content easier.

Google+ And Real-Time Search

That's why Google partnered with Twitter in 2009 for real-time search. But that deal - along with real-time search - was shut down this summer. It wasn't beneficial for either party anymore. Twitter is now acquiring search companies for its own purposes, and Google has started its own social network.

Public Google+ posts are now incorporated into search results, and Google is practicing with home-grown real-time search within the social network. By collecting social signals, Google can further improve the "freshness" of its search results with a human touch.

Are your Google searches smart about returning timely results? Share your experiences in the comments.

Ever since Google launched its instantaneous search product last fall, I've noticed something - everywhere I go on the Web, I enter search terms and and pause for a second, only to remember that not everything has yet gone instant.

Today, Yahoo has announced that it too has gone instant, but with a slight twist on Google's solution to providing an ever faster real-time Web. Yahoo says it will offer "answers, not links".

So how does Yahoo's offering differ from Google? According to the company, with Search Direct "Yahoo! content is combined with information from the Web to provide rich answers, not just links, and to give people the option to immediately engage or continue to a traditional search results page."

As you can see with the above screenshot, Yahoo isn't about to offer its users a link to The Weather Channel when they search for "weather". Instead, the search engine will provide just that. The same goes for a variety of other categories. The initial release of Search Direct is currently in beta and available only within the U.S., but offers immediate "answers, not links" for topics including top trending searches, movies, TV, sports teams and players, weather, local, travel, stocks, and shopping categories.

While it's yet unclear whether or not the move will do anything to increase Yahoo's share of the search engine market, it should do one thing - ensure that it doesn't bleed out users as they begin to become accustomed to real-time results in other places on the Web.

Microsoft, it looks to be your turn. I, for one, don't have much patience for that "enter" key any more.

Convore is a real-time communication tool, but unlike many other apps that offer group chat, Convore runs in your browser and doesn't require a download or chat client. "Basically, it's a contemporary version of IRC," says co-founder Leah Culver.

Convore supports both public and private groups. Currently, based on the activities of the alpha-testers, the topics of the public groups veer towards tech, movies, and "who's up late at night coding," but users are able to create their own.

Within each topic are a number of threads, and within these are conversations, updated in real-time. This makes Convore feel like a blend of forum and chat. When you log into the site, you can see the conversations that have occurred while you were offline, and - a great feature - when you step back in to chat, Convore scrolls to the point where you left off, so you can read to catch up.

Culver is joined by co-founders Eric Florenzano and Eric Maguire. The three have built the real-time technology upon which Convore runs. It uses what Florenzano describes as "long polling, which means that on every page, the browser opens a connection that talks back to some proprietary technology we've built to speak to a cluster of Redis servers. We have some technology that parses each message for links, for embedded images, and constructs some structured data about the message, and then sends it out to everyone in the group via Redis."

You can sign up for Convore now and take it for a spin.

Khris Loux and his company, Echo, have always had a tentative relationship with the lowly blog comment. Echo launched in 2009, described as a "blog commenting platform" much like Disqus. Right from the beginning, however, Echo went beyond the comment and aggregated all sorts of real-time data from around the Web to pull into the section normally reserved for comments. Now, Echo has gone beyond declaring the death of the comment and re-launched as a "real-time as a service" platform.

At its re-launch event today, the company brought out the big guns to show off just how useful it thinks its new incarnation will be. We got a chance to talk with them beforehand to go beneath a bit of the flashiness and we got a glimpse of a service that adds a new building block onto the Web and could bring the real-time Web to previously untouched corners of the Internet.

When we spoke to Echo CEO Khris Loux the other day, he laid it out for us quite simply.

"In the ways that print gave way to TV, static pages will give way," said Loux. "The challenge for the rest of the publishers on the Internet is that they're running static websites. The revenue has moved on from those sites."

He went on to explain that the primary form of real-time interaction on most sites is through Facebook comments or "likes" or Facebook Connect and, while those are good and valid tools that publishers should still use, they are not enough. According to Loux, the main problem there is that "Facebook and Twitter still control the experience and ultimately control the revenue."

The answer to all of this, of course, is the new version of Echo, which acts as a "real-time as a service" platform. It can help aggregate all manner of real-time data - from Facebook posts to Tweets to comments to blog posts on your own site - and help you and your users to interact with the content.

"Real-Time As A Service"

"We are putting forth the notion of 'real-time as a service.' Just like a start-up would no more build their own data center, a publisher or start-up should no longer build real-time. You could build Plancast with this. You could build Yammer," said Loux. "The Web is becoming designing blocks and Echo, real-time as a service, is the new block in town."

In many ways, Echo has done this all along - it has allowed publishers to pull in and aggregate real-time content to display, in real-time, on their site. The big difference now is that the service is acting more as a real-time platform and less as a simple service to display real-time content. Once it pulls in the data, it stores it and lets you work with it. It "socializes" it. It lets you interact with the data in ways that are based on what users do with it. They can vote content up or down, comment on it, share it socially, and based on these actions the publisher can display it differently. And, of course, it's all in real time. As Loux put it, Echo "doesn't care" what type of content you're dealing with, once it's in the system it can be treated the same as any other.

What Can Echo Do?

If you want to see the new Echo in action, you can take a look at the Sports Illustrated World Cup site from last year, which uses the new service. It aggregates on-site content, tweets and photos to create a real-time site about the World Cup. Another site, set up for teen idol Greyson Chance, shows off the ability to pull in content and interact with it in different ways. In this case, the stream of pictures is created from Tweets using the #WOTL hashtag. But, as Loux pointed out, it would only take another step to turn the entire page into a contest, letting users vote on each others' pictures and displaying them according to votes. That's the type of interactivity that echo is trying to enable with its new service.

Will it work? The company is coming out today with a number of big names, from NBC to Sports Illustrated to Reuters and Newsweek. But these are all companies that could certainly afford to build their own real-time components. The real question is whether or not small companies will use Echo as the real-time building block that Loux envisions. If Echo can bring the real-time Web to publishers big and small alike, in a way that they can interact with and own the content, it could make a big splash. Stay tuned below to watch today's launch event, live from SF MOMA.

Unlike other Twitter widgets that simply stream tweets in a timeline, Topsy's new module let publishers display only the most relevant and brand-safe tweets. The content will be filtered - profanity-free and language-specific.

Topsy touts the increased engagement rates from publishers using social content powered by these modules. As the content is filtered based on relevancy, Topsy argues that it will provide readers with a better experience (leading in turn to more time spent on the site, hopefully).

As Vipul Ved Prakash, Topsy co-founder and CEO says, "The launch of Topsy Social Modules furthers our position as the fastest growing and leading supplier of relevant realtime search results from the social web."

Indeed, Topsy points to the importance in not just the social web, but in how it dovetails with search. Rather than PageRank as being the sole marker of what's important, links from Twitter are becoming an increasingly important signal of relevance. And as Topsy powers the largest searchable index of Twitter data in the world, it makes sense that publishers would want to turn to it to power widgets on their sites.

And while the world may not need another widget, this new feature from Topsy does seem to provide one with content that's more useful - for both readers and publishers.

It's a development of historic significance: few events have changed the experience of life on earth as much as last century's proliferation of hundreds of millions of automobiles. The automobile was a revolution in personal autonomy, but it came with great costs. Now we've entered an era when that personal autonomy will become automated and some of the automobile's costs could be mitigated as a result. As a technologist, I find it helpful to understand the emergence of the fabled self-driving car as a convergence of three trends: the Internet of Things, Big Data and Real-Time Technology. Those trends are poised to go far beyond a self-driving car.

The Internet of Things

The personal automobile was invented near the end of the 19th century and popularized throughout the 20th - approximately 600 million now traverse a world of roads. There is 1 car for every 11 people on earth, but sometimes it's hard to remember that - they can seem so ubiquitous.

As widespread and significant as they are, though, automobiles themselves have been to date relatively unconnected to each other. They are big, dumb, fast-moving hunks of metal. Essential, in fact, to keep apart from one another as they hurtle down the highway.

"Your car should drive itself," Google CEO Eric Schmidt foreshadowed in a public statement last week, days before the Google self-driving car was unveiled. "It's amazing to me that we let humans drive cars. It's a bug that cars were invented before computers."

The instrumentation and networking of previously offline objects, like cars, buildings, roads and more, may represent the next big stage in the evolution of technology. A network-connected car in particular - primarily navigated by artificial intelligence able to leverage a world of data hosted in the cloud - promises big gains in safety, efficiency and quality of life for everyday drivers.

This product illustrates the potential for what's called the Internet of Things in a big way, but it should be understood in the context of the much larger trend. "Trying to determine the market size of the Internet of Things is like trying to calculate the market for plastics, circa 1940," says Georgetown University Communication, Culture & Technology professor Michael Nelson, the former director of Internet Technology at IBM, and the former director of Technology Policy with the Federal Communications Commission, in a February report by The Hammersmith Group. "At that time, it was difficult to imagine that plastics could be in every-thing. If you look at information processing in the same way, you begin to see the vast range of objects into which logic, processors, or actuators could be embedded."

Big Data

How is Google able to drive a car? In part because its software has access to an incredible amount of very detailed data: maps of the world, speed limit information, live video of a car's surroundings, sharp computer vision analysis of what that live video contains and big-picture data about an emerging network of cars in motion on the road.

Multiply that by not just one car, but countless cars throughout our roads and you've got a whole lot of data being processed.

Where there are network-connected devices, there are waves of data made available. Where there is big data, there are opportunities for pattern analysis, rational decisions and recommendations based on that data.

It's hard to know whether Google's self-driving cars ought to be called recommendation technology based on big data, or whether this is going beyond recommendation and into directive commands.

Big data is, in large part, about decision making. A Google self-driving car will make it easier to make decisions about the important parts of driving by automating the parts that don't require human decision-making.

As Michael E. Driscoll wrote this spring,

"...as information generating processes become more frictionless -- as humans have been excised from information read-write loops -- the velocity and volume of data in the world is increasing, and at an exponential rate...

"As the Big Data stack matures, tools that help manage the workflow from data to analytics to visualizations, and ultimately to decisions, will be critical. Someday, creating and sharing a data analysis through a web dashboard should be as easy as writing a blog post. Until that day, there's plenty of work to keep us data scientists well-employed."

Or managing the data through automobile navigation, as the case may be. It's not hard to imagine new, self-driving automobiles competing in the market based on their navigation algorithms or the ways they enable human drivers to relate to all the rich data made available by sensors and cameras on cars.

Real-Time Data

The real-time Web has been mistakenly characterized as nothing but a stream of Tweets and Facebook updates.

Just as real-time automated trading has changed the stock market forever, real-time delivery of big video and sensor data to and from the cloud and the car on street will change driving forever.

Accident avoidance is the most-obvious example, but there are others. Google spokespeople have talked about car-sharing programs that drive cars to the location of a would-be user on command, avoiding the inefficient use of valuable parking spaces in every neighborhood. Presumably those vehicles could predict where they are most likely to be needed, based on real-time data about existing behavior.

Ultimately, though, real-time will be most evident when you're in the car and it makes decisions based on surrounding circumstances as they unfold.

History

The point is, trends like the Internet of Things, big data and real-time data are poised to impact the human experience in many different ways. Their convergence in the sci-fi scenario of self-driving cars is the example that will be most evident to everyday people.

The consequences of such a technology on the psychology of autonomy, the ecology of transportation and urban planning and the million-plus lives lost every year due to human negligence behind the wheel, will be profound.

Even more profound, though, will be the spread of these same technology trends throughout our lives. Beyond our cars, automation and real-time analysis of data-rich environments could change the human experience in many, many ways in the relatively near future.

Google may have killed Wave (prematurely by some accounts), but it has added a little bit of real-time collaboration to one of its flagship offerings, Google Docs, with the addition of collaborative highlighting.

The difference here is that it is all happening in real time. As we both look at the document and go to make changes, we see one another highlight text and add commentary. It's a step-by-step process rather than a completely asynchronous communication. It's the difference between chat and email and it makes all the difference in the interaction between writer and editor, or even editor and editor if you're collaboratively editing a document.

Wave may be dead, but we hope that Docs will begin to see more real-time collaborative tools introduced in its place.

What do you think - what other uses could you see for this sort of real-time interaction? And what other bits of Wave would you like to see revived in Docs?

A lot of us here at ReadWriteWeb have been testing the preview version of TweetDeck and the only major limitation we have discovered so far is the fact that user lists in TweetDeck don't update in real time and still use the old polling mechanism. Sometimes, when you add a search very popular topic that creates more than a few updates per second, Twitter will also throttle your updates. This, of course, is not TweetDeck's fault, though you might want to keep this in mind when you are using TweetDeck to track a breaking news story.

One nifty new feature of Twitter's API is that you will also get an alert when somebody favorites one of your tweets. Until now, these actions often went unnoticed by many users. This update makes favorites a more interesting part of the Twitter ecosystem.

You can download the TweetDeck Twitter User Streams Preview here.

Alternatives

It is worth noting that other Twitter clients, including Seesmic, also offer beta versions of their applications with support for Twitter's User Streams API. Echofon, a popular Twitter client for the Mac also offers a User Stream preview. As far as we are aware, there are currently no mobile apps that support real-time updates yet.

"I wish we were launching something worth 'all this whoop-la,' but this is an update to an existing analytics product that very few people see."

However, a post on an official Twitter blog seems to give the lie to that statement.

"Our analytics, operations and infrastructure teams are working on a system that uses cassandra for large-scale real time analytics for use both internally and externally."

In other words, Twitter is in fact working on an important analytics product that will launch to the public soon. This blog post bolsters the Tweets we referenced in our earlier article.

Cassadra is an Apache "open source distributed database." A "change in strategy," as King describes it, means Twitter will not use Cassandra to store Tweets as it planned. Now it's going to be used for the real-time analytics that we reported on.

Not a big deal? Nah, it's a big deal. To paraphrase Marshal Kirkpatrick, "Twitter is launching a public analytics product for the countless business customers who will line up to use it."

So, with all due respect to the VP, we're going to stick to our guns.

Unlike Google, however, it looks like Yahoo gives time relevance a higher priority in its suggestion algorithm, separating the two suggestion services.

The implication here is that Yahoo suggests the most relevant information based on what's going on, right now, while Google suggests a search based on the number of searches, which could be hours or even days old. The fact that today's match doesn't even make the radar of Google Suggest is telling in how the two services differ.

Yahoo said that it will also offer similar near real-time suggestions in Yahoo News search, an obvious use for real-time search suggestions.

While Yahoo has increased the importance of real-time information in its search suggestions, Google recently took a similar step, but instead focused on localizaing its suggestions.

We have to wonder when all of the real-time information made available to Yahoo through a number of websites - including Facebook, Buzz, Flickr, YouTube, Delicious and more - will make
its way into its Search Assist results, if it hasn't already.

The company delivers streaming news, via widgets, APIs and its Site Search Platform, to partners. It intends to use the infusion to build out its team and establish more partnerships, according to Collecta CEO Gerry Campbell. Past partnerships have included CNET's coverage of the iPad launch and MySpace's Today on MySpace feature.

"We've got a solid team and have spent time building a scalable infrastructure," Collecta CEO Gerry Campbell told us in a phone conversation. "Now, Collecta is more than a great idea. We understand the business behavior of our users and we'll be building out a full-fledged business."

Campbell anticipates adding personnel in business development, program management and ad sales - the latter because one of the major initiatives on deck is the creation of a full-fledged advertising system.

We've mentioned in previous coverage that when we tested Collecta, the news it streamed from 15 million content sources was in fact less than a minute old, and that its widget is freely available for anyone to use in their website.

Collecta says that 60% of queries it sees come in via its partner API. OneRiot, its main competitor, says it sees 97% of its queries that way.

Collecta is notable because it forms explicit partnerships with publishers, such as WordPress, and stream their content into the index in real time.

OneRiot focuses more on P2P indexing with the use of their installed toolbar and indexing Twitter & Digg. They also recently launched an ad network.

Real-time searches also tend to shy away from many normal search parameters: big on entertainment and very low on porn.

Read more ReadWriteWeb coverage of Collecta. ReadWriteWeb's real-time web coverage is here.

Disclosure: In addition to announcing a new round of funding, Collecta is also an event sponsor of this week's Real-Time Web Summit.