Cutting Edge

Though only three weeks old, Wolfram Alpha is already showing growing spurts.

On Monday, the project unleashed a variety of updates to its computational search engine, according to the latest blog from the Wolfram Alpha team. The updates include 1,850 changes to its code base and 1.1 million updates to its data.

In one sense, Wolfram Alpha is in constant update mode, since new data is flowing into the system all the time. But this is the company's first major release of so many core updates in one shot.

The blog post lists about 20 of the many updates. Some of the descriptions are abstract, such as "Additional linguistic forms for many types of data and questions." But several of the updates intrigued me, so I took them for a test drive to see how they fared with my own questions.

Here's a list of the updates I tried, along with my results:

More "self-aware" questions answered (e.g. "how old are you")
When I asked Wolfram Alpha its age, it told me it was 24 days, 10 hours, 21 minutes, and 38 seconds old. When I asked how old President Obama is, it said he was 47 years, 10 months, and 4 days old. Simple enough.

Improved linguistic handling for many foods (e.g. "love apple")
I typed "Big Mac," and Wolfram Alpha displayed a nutritional label, breaking down the amount of fat, cholesterol, sodium, and other nasty items in the burger. Definitely makes me want to stick with a salad.

Big Mac Facts from Wolfram Alpha

Combined time series plots of different quantities (e.g. "germany gdp vs population")
I entered "US population vs. oil prices." The site displayed a single chart tracking and comparing the rise in the number of people in the U.S. with the jump in the price of oil since 1970. I can see this as a useful capability to compare two distinct but possibly related items.

Additional support for stock prices with explicit dates
I typed "IBM June 1 1980," and Wolfram Alpha popped up a chart tracking IBM's stock price for the four-year period from 1979 to 1982, with the June 1, 1980, price highlighted. But Wolfram Alpha's stock data only goes back so far. When I entered a date for IBM stock earlier than June 1970, no chart or data was found. Other stocks also came up empty for data prior to the '70s.

Support for planet-to-planet distances and "nearest planet", etc.
Typing "Earth Saturn" brought up detailed facts and figures for both planets, including mass, radius, number of moons, distance from the sun, and distance from each other. This proved a quick, easy way to get all those facts in one shot.

More comparisons of composite properties (e.g. "US military vs. UK")
I entered "US unemployment vs Spain." The site told me the U.S. jobless rate is 7.2 percent, but Spain's is 13.9 percent. To double-check, I ran Google searches for the two countries' jobless rates. I found the current U.S. rate is around 9 percent (sourced to the U.S. Bureau of Labor Statistics, and I double-checked on the BLS site), while Spain's is 18 percent. Checking further, I discovered that Wolfram Alpha was showing me data from the last three months of 2008 rather than the latest figures.

When I searched Wolfram Alpha for "US employment" and "Spain unemployment" by themselves, I got the same 2008 figures, but this time Wolfram Alpha told me the numbers were for 2008, a fact omitted in my first search results. Finally, I searched for "US unemployment 2009," but the site simply told me no data was available.

Based on my limited searches, my impression of Wolfram Alpha was mixed. I was intrigued by some of the obscure data it conjured up. I also liked the convenience of finding all relevant facts on one page.

But I was disappointed by the gaps in its knowledge. Why no stock information prior to 1970? (I'd like to check stock prices from the crash of 1929). Why no unemployment data later than 2008? And why the lack of consistency in telling me the jobless numbers were from last year? I can more easily Google "unemployment rates" to grab the latest results.

Granted, Wolfram Alpha isn't like a traditional search engine, serving up links to other sites. It relies on its own research and database to deliver results. The service has even said it's trying something different here. And at only three weeks old, the site is still experiencing growing pains.

The latest updates can help track down certain types of data. But Wolfram Alpha may have a way to go before it offers truly complete and up-to-date information.

This post was updated at 9:10 p.m. PDT to note that Wolfram Alpha is now up and running.

Wolfram Research founder Stephen Wolfram (blue shirt, center) convenes a meeting Friday night live on the Internet to discuss launch problems with Wolfram Alpha.

(Credit: Screenshot by Tom Krazit/CNET)

Wolfram Alpha struggled to get up and rolling Friday evening under difficult conditions, as the company scaled back expectations for its performance this weekend.

The new search engine attempted to make its debut literally in the middle of the perfect storm: a tornado watch had engineers on edge in Champaign, Ill., where Wolfram Research attempted to bring the service online. However, networking and database problems also prevented the engine from launching as of 6 p.m. PDT, an hour after the company said it would go live to the world.

And to top it all off, uplink problems with the Justin.tv service prevented Wolfram from explaining exactly what was going on for nearly half an hour, while commenters in the chat room mercilessly heckled the company with the 21st century Bronx cheer: EPIC FAIL. Eventually, around 5:30 p.m. PDT, Wolfram founder Stephen Wolfram appeared on camera to explain that glitches were holding up the launch, as claps of thunder sounded in the background.

After initially claiming that the service would go live Friday evening, Wolfram lowered expectations by only promising a test launch over the weekend, with full service expected by Monday. Searches could not be conducted through the main home page as of 6 p.m., but it was possible to get in through a back door posted on Twitter and in the Justin.tv chat room and start searches, although performance was spotty.

Wolfram Research was forced to show videos showing off the data center servers and power-redundancy systems taped earlier in the day while Stephen Wolfram convened an emergency meeting to figure out what was going on. The broadcast was somewhat less-than-polished, with audio engineers talking over Wolfram's initial address to the audience and rendering much of his speech incoherent.

Please let us know if you're able to get into the search engine Friday night, and we'll update our coverage over the weekend if Wolfram is able to bring the service online.

Update at 9:10 p.m. PDT: Wolfram Alpha is now up and running and seems to be working just fine. A spokesman for Wolfram Research called at 7:30 p.m. PDT to say it was about to launch, but its exact launch time is unclear.

(Credit: Screenshot by Michelle Meyers/CNET)

Screenshots: A tour of Wolfram Alpha

View the full gallery

This post was updated at 4:08 p.m. PDT with information from a Wolfram Alpha blog post and again at 5 p.m. PDT with info from a Wolfram spokesman.

Wolfram Alpha, the new "computational knowledge engine" set to debut publicly Friday, has hit a technical snag that could delay its launch, a spokesman for Wolfram Research confirmed.

The online tool--which some say could give Google a run for its money--supplies answers to factual, data-intensive questions but also does math in the process. It was set to go live to the public at 5 p.m. PDT.

However, in an interview with the Los Angeles Times (and confirmed by Wolfram public relations director John Ekizian), Stephen Wolfram said a large-scale traffic simulation test had failed. "We ran into a small snag, which hopefully won't turn into a big snag," Wolfram said.

Here's more on how Wolfram explained it to the Times:

We have several supercomputer-class compute clusters. One of our tests was to use one cluster to simulate traffic and run it against the other cluster. And when we did that last night, we found that the through-put we got degraded horribly when we increased the amount of traffic that we were pushing from one cluster to the other.

Ekizian said, for now, he could only confirm the accuracy of Times' story. But stay tuned for expected updates related to the planned 5 p.m. launch.

Update at 4:08 p.m. PDT: A Friday afternoon post on the Wolfram Alpha blog about the countdown to launch doesn't offer any specifics on time frame. Rather, it says the team has "been switching on more and more compute capacity, with the expectation of having full capacity available on Monday."

As for the snag, it only says the team is in "full-court-press resolving network infrastructure, database, and all sorts of other challenges, made particularly interesting by the sheer scale and complexity of launching two supercomputer-class clusters along with three other locations."

The post concluded with a reference to a tornado watch in Illinois where Wolfram Research is based: "It will be an interesting evening!"

Update at 5 p.m. PDT: Ekizian confirmed that Wolfram expects Wolfram Alpha to go live between 5 p.m. and 8 p.m. PDT Friday, "for some time." He warned not to be surprised if the service is up and down over the weekend, but it should in "full throttle" by Monday morning.

One of my concerns with the public launch of Wolfram Alpha later this month is withstanding the crushing load the Internet can impose. But Wolfram Research revealed Tuesday it's building the service on the world's 66th-fastest supercomputer.

The machine, built out of Dell hardware by a company called R Systems, can sustain performance of 39.6 trillion mathematical operations per second, according to the November 2008 list of the top 500 supercomputers. That muscle will come in handy for Alpha, which I think of as a combination of a graphing calculator, search engine, and reference library that not only supplies some answers to factual, data-intensive questions but also does math in the process.

"There is no way to know exactly how much traffic to expect, especially during the initial period immediately following our launch, but we're working hard to put reasonable capacity in place. Will we have enough computing power to provide computable knowledge for everyone who visits? We hope so," Wolfram Research said on its Wolfram Alpha blog Tuesday

The system, called R Smarr, has 4,608 processor cores using 576 quad-core "Harpertown" Xeon machines, 65,536GB of memory, and high-speed InfiniBand data-transfer connections, according to the Top500 site and a Dell case study on the system (PDF). It also uses both the Red Hat Enterprise Linux and Microsoft Windows HPC Server operating systems, according to the Dell paper.

Alpha requests will be served from five co-location facilities, Wolfram Research said. There actually are two supercomputers in the project, with nearly 10,000 processor cores total and hundreds of terabytes of hard drives.

R Smarr doesn't use ordinary Dell servers. Instead, custom-made machines were ordered through Dell's Data Center Solutions division. "We evaluated the standard Dell PowerEdge servers, but at the time, those systems did not offer a server board that could deliver the high memory bandwidth necessary for our client," said Brian Kucic, R Systems' vice president of business development, in the Dell paper. Kucic.

Some of what goes on behind the Alpha covers is use of Wolfram Research's Mathematica software, which can perform a wide variety of mathematical and graphical operations. A massive number-crunching utility freely available over the Web sounds like a recipe for disaster, but in my tour of a preview version of Wolfram Alpha, I encountered a timeout limit of about 5 seconds in searching all occurrences of a particular sequence in the human genome, so it looks to me like Wolfram has the ability to throttle usage.

Wolfram Alpha is like a cross between a research library, a graphing calculator, and a search engine. But does Wolfram Research's "computational knowledge engine," set to debut publicly later this month, live up to its hype as a Web site that Google needs to be afraid of?

Wolfram Alpha creator Stephen Wolfram on Tuesday gave a demo of the service to a crowd of online reporters. Few have access to the private test version of the service itself, but we got access Monday night. We found it compelling, if limited.

We're eager to see this site develop. It does things with online information that Google does not. Here are our impressions of the current version of Wolfram Alpha.

Who's it for?

CNET reporter Stephen Shankland: Today at least, Wolfram Alpha is for the tech crowd--the kind of people who want to dig into the data. It's a great exploration tool to find out whether somebody who's 5 feet 5 inches and 160 pounds is overweight, the chemical properties of boron, and whether you're going to get a full moon during the evening of September 4 in Buenos Aires when you want to propose to your fiancee.

Wolfram Alpha will show you when the next eclipse will occur over San Francisco. Click above for a gallery of screenshots.

(Credit: Screenshot by Stephen Shankland/CNET)

It'll tell you the family, genus, species, and caloric value of an apple, and it'll forecast Apple's stock price, but it won't give you apple pie recipes. It'll tell you the box office take of the first "Star Trek" movie, but it won't tell you the theater where you can see the newest "Star Trek" movie.

But a technical audience is still big. This could unlock a lot of data that students, research assistants, lawyers, marketing managers, financial analysts, and scientists might not have readily available. And those folks are important, too--just the kind of influential folks people with Web sites like to reach.

CNET Editor Rafe Needleman: I wouldn't dream of pointing my parents at this. It's too picky about syntax and not intuitive to get into. When I saw Stephen Wolfram give a demo of the system I was blown away. He ran through dozens of demos from weather to genetics to calculus to finance, each resulting in beautiful and informative results. But when I tried the service I'd say maybe only 10 or 20 percent of my queries actually worked.

Shankland: On the other hand, my dad has a Ph.D. and I most definitely will point him at it. He bought Wolfram Research's all-purpose computation software, Mathematica, though, so for him Wolfram Alpha is like preaching to the choir.

Needleman: My dad has a Ph.D., too, but in philosophy. There is no Wolfram Alpha for that.

Shankland: Yet. Alpha handles numeric data well, but loosey-goosey stuff like art or philosophy is tough. But maybe in some glorious future Alpha will be able to chart the trains of thought from the Enlightenment to the present.

Is it easy to use?

Needleman: You need a clear mind to take advantage of this service. Again, it's picky about syntax, and in the pre-release version we tried, if you got a query wrong--if it didn't return what you were looking for--it wouldn't offer you much in the way of help to refine the query. I kept trying to figure out how to correlate weather with earthquakes in San Francisco. I can get the data for weather. I can get it for earthquakes. So I know that Alpha has the information. But I can't figure out how to show them together.

Curious about how closely NetApp's stock price has correlated with EMC's? Wolfram Alpha will tell you. Click above for a gallery of screenshots.

(Credit: Screenshot by Stephen Shankland/CNET)

What the system does know is beautifully presented. Type in the name of a city, for example, and it will give you some fun stats on a clean and clear Web page. But from that sort of page you'll probably want to start exploring the data available: Maybe you want to know about population growth, economic information, or weather trends. Alpha doesn't give you hints as to what's available, nor a good way to drill into data. You have to take stabs at re-typing your query. I tried a variety of queries like "test scores san francisco schools" and "population of portland by year" and got, respectively, no result and a pointless result (533,429 person years: what is that?). The system that interprets Wolfram Alpha queries needs a little bit of help. It may be improved by the time the system is opened to the public, later this month, but I think that this will be the product's Achilles heel.

Updated at 3:12 p.m. PDT with further detail.

Wolfram Research founder Stephen Wolfram publicly debuted his company's forthcoming online "computational knowledge engine" Tuesday--but search Goliath Google launched a service of its own that bears significant resemblance.

Wolfram Research CEO Stephen Wolfram

(Credit: Stephen Wolfram)

The Wolfram Alpha engine is a Web service designed to process data from controlled, vetted sources of data--many not on the Web--then present the results in a way that lets people dig deeper into the subject. It's something of a cross between a graphing calculator, repositories of scientific data, and a system to interpret questions posed in human terms.

"Like interacting with an expert, it'll understand what you're talking about, do the computation, and present the results in such a way you'll be able to understand what the consequences are," Wolfram said in a talk at Harvard's Berkman Center for Internet and Society Tuesday.

For example, people can ask about the molecular weight of caffeine, about the location of a gene in the human genome, the number of people named Andrew born in a particular year, the amount of fish produced in France, the life expectancy of 40-year-olds, and the performance of Microsoft stock--and then dig into the results. The height of Mt. Everest can be expressed in terms of the length of the Golden Gate Bridge.

Wolfram has deep technical chops. He's a MacArthur "genius grant" recipient who got his Ph.D. in theoretical physics at age 20, founded Wolfram Research to commercialize mathematics software called Mathematica that can perform a wide variety of computational and graphing chores. He also spent a good portion of the 1990s writing "A New Kind of Science," a 1,200-page tome (also available online) that seeks to transform science by presenting a computational view of physics.

The Alpha site will be publicly available "in a few weeks," with free access to all users supported by sponsors and subscriptions for heavy-duty users who want the system to process their own data, Wolfram said.

Gatecrashing Google
But another similar service is available today: a Google feature that can search public data and present the results graphically.

"We just launched a new search feature that makes it easy to find and compare public data," Ola Rosling said of the service in a blog post. "The data we're including in this first launch represents just a small fraction of all the interesting public data available on the web. There are statistics for prices of cookies, CO2 emissions, asthma frequency, high school graduation rates, bakers' salaries, number of wildfires, and the list goes on."

The service is based on Google's 2007 acquisition of Trendalyzer, Rosling said.

Google now lets people search public data sets.

(Credit: Screenshot by Stephen Shankland/CNET)

One example: "When comparing Santa Clara county data to the national unemployment rate, it becomes clear not only that Santa Clara's peak during 2002-2003 was really dramatic, but also that the recent increase is a bit more drastic than the national rate," he said.

Thus far, Google's service includes data only from U.S. Bureau of Labor Statistics and the U.S. Census Bureau's Population Division.

"We hope people will find this search feature helpful, whether it's used in the classroom, the boardroom or around the kitchen table. We also hope that this will pave the way for public data to take a more central role in informed public conversations," he said.

Google didn't immediately comment about whether the timing of its launch was coincidental, and Wolfram Research didn't immediately comment on the Google product.

Alpha's underpinnings
Alpha has four main components, Wolfram said.

• Data curation. Wolfram Alpha uses public and licensed proprietary data sources, and the company uses automated processes and human choices to prepare the data. "At some point you need a human domain expert in front of it," Wolfram said.

• Algorithms. Alpha must pick the right computational processes to present its results. "Inside Wolfram Alpah are 5 million to 6 million lines of Mathematica code that implement all those methods and models," he said.

• Linguistic analysis to understand what a person typed. "I thought one of many things that could have gone wrong was that short, lazy things would (have) huge amounts of ambiguity," for example figuring out whether "50 cent" had to do with musical artists or money. "That turned out to be not nearly as much of a problem as we expected."

• Presentation. "There are tens of thousands of possible graphs. What do you want to show people?" Wolfram asked.

Wolfram hopes the tool will help researchers perform scientific chores that before were possible but not necessarily worth their time.

"What's the angle of sun at particular moment? Given 20 minutes, I could compute it and get it right, but I probably wouldn't bother," Wolfram said. "What Wolfram Alpha does is take that piece of scientific knowledge and make it immediately accessible to everybody."