Apple has released a few updates for its iLife suite of media creation and management applications, which have been made available through the Mac App Store as well as by conventional means. Because of the way Mac App Store applications are coded, the version of iLife distributed through it will require updates issued through the store instead of via Software Update. Therefore, if you have installed iLife via the Mac App Store, you will need to use the Store to update your applications, and if you have iLife installed by conventional means then you can download updates either via Software Update or through standalone installers.

iLife updates are available through Software Update and the Mac App Store.

The updates contain the following fixes, according to Apple:

iMovie 9.0.4 (76,20MB)

• [Adds] support for opening projects imported from iMovie for iOS.
• Fixes an issue where some audio adjustments were not preserved.
• Addresses a performance issue when using large quantities of video clips with keywords.
• Resolves [an] issue with slow application launch when working with large iPhoto libraries.

iPhoto 9.1.5 (220.83MB)

• The date range of each event now updates correctly to reflect changes made to photos using the Adjust Time and Date command.
• Addresses an issue that could cause the Photos view to scroll incorrectly when Event Titles are displayed.
• Fixes a problem that could prevent Ken Burns animations from being applied correctly on photos in a saved Classic slideshow.
• Deleting photos from a Web album using the contextual menu now removes them from the album without deleting them from the library.
• Addresses an issue that could prevent a crop adjustment from being removed from a photo when using the Revert to Original command.

GarageBand 6.0.4 (53.72MB)

• Corrects issues with some tempo-based effects falling out of sync with main song tempo.
• Resolves problems with Magic GarageBand project files not opening up correctly in the GarageBand tracks view.

iWeb 3.0.4 (178.57MB)

• This update improves overall stability and addresses a number of minor issues.

iDVD 7.1.2 (36.12MB)

• Improves reliability of imported iPhoto '11 slideshows.
• Fixes a problem that could prevent files from relinking.
• Addresses an issue that could cause iDVD to use all available CPU capacity when searching for missing files.

As always, be sure to back up your systems before installing the updates. (If you have Time Machine running in its default configuration then you should be good to go.)

Questions? Comments? Have a fix? Post them below or e-mail us!
Be sure to check us out on Twitter and the CNET Mac forums.

Originally posted at MacFixIt

(Credit: EA Mobile)

The official Scrabble Free app hit the Android Market today, giving word nerds yet another platform for earning triple letter scores. With alternatives like Wordfeud and Words With Friends already gathering dust on the Android Market shelves, it was only a matter of time until the granddaddy of them all became available as well.

Scrabble Free for Android lets you engage in word warfare across platforms, against players on Macs and other iOS devices or on PCs (via Facebook). It also includes an in-game Official Scrabble Dictionary, and a really neat Teacher Feature that helps you sharpen your Scrabble chops as you play.

Open-Xchange's tool for helping people reconstruct their Facebook contact list on Google+ has fallen victim to Facebook's revocation of its privileges.

Open-Xchange, a maker of open-source e-mail and collaboration software, last week launched a tool that used the company's Social OX technology to help people assemble a list of their friends. It used connections to a combination of services such as LinkedIn and e-mail accounts to create a single "magic address book."

The tool didn't actually copy e-mail addresses from Facebook--only first and last names. It then matched those names to other e-mail records in the user's accounts. But Facebook disabled the API (application programming interface) key that the software used to read the names, Open-Xchange Chief Executive Rafael Laguna said.

Facebook gave two reasons for the move and underscored the seriousness of its decision with a warning about the repercussions:

1. You cannot use a user's friend list outside of your application, even if a user consents to such use, but you can use connections between users who have both connected to your application. (FPP II.11)

2. A user's friends' data can only be used in the context of the user's experience on your application. (FPP II.4)

Please note that we will not re-enable violating applications and any policy issues in your existing and future apps will result in further enforcement actions.

Facebook told CNET its actions are an aspect of protecting Facebook users. "The people who use Facebook expect us to protect the information they share with us, and prevent phishing, malware and scraping whenever possible. We are always reviewing and revising our user protections, and when we come across Platform integrations which violate our terms, we respond quickly," the company said in a statement.

Laguna objected to Facebook's decision. In a letter to the company, he said:

We are not aware of violating anything. We are using your API to extract the last name and first name fields. We are not parsing or scraping the e-mail address. That same data is available at your site under "Account->Account Settings->Download Your Information" in the resulting friends.html file.

This isn't the first time Facebook has made such a move. Facebook also blocked use of a Chrome extension by programmer Mohamed Mansour that would export not only names but also e-mail addresses, phone numbers, birthdays, and more. Mansour's tool didn't use Facebook's API for programmatic access, but instead extracted the data from the pages themselves.

The matter of who owns address-book data in Facebook was a hot one last year, when Google and Facebook got into a spat about contact-data sharing. But now that Google finally has what appears to be a viable competitor to Facebok, the issue is a really hot one.

Google probably only has something like a hundredth of Facebook's population. But with one Google+ user base estimate exceeding 4 million people, that's still pretty significant in absolute terms.

Some of the issue about the Facebook data is semantic. If you call it your address book stored within Facebook, it seems more like you should be able to get access to that information. After all, it's information your friends have shared with you. But if you think of it as your friends' own profile data, the ownership claim doesn't look as strong, regardless of the matter of whether they agreed to share.

Fundamentally, though, there's nothing stopping people from copying each e-mail address manually from Facebook. The problem crops up when such extraction is automated.

Which is why the Yahoo position is so peculiar. Anyone with a Yahoo Mail account can easily extract not only the names but the e-mail addresses of their Facebook friends. If you want to try this method, first open a fresh e-mail account at Yahoo. As soon as it's created, you'll get an option to import contacts from elsewhere, including Facebook. Choose that option, authorize the tool, and it'll import the contacts.

To make them useful in Google+, you'll then have to transfer them to Google by exporting a CSV (comma-separated values) spreadsheet. Gmail lets you import a CSV file by clicking Contacts, then the More actions button, then the Import menu item. Of course you can use the CSV file elsewhere, too.

Laguna asked how it's possible for Yahoo to extract the data but not others.

"Is there a way to get sanctioned or even paid access?" Laguna asked, then raised the matter of the two objections Facebook raised about Open-Xchange. "You must have some kind of arrangement with Yahoo, which even has an import capability of not only the names but also the Facebook e-mail addresses? How can Yahoo do it without violating points 1 and 2 above?"

Facebook wouldn't comment on the Yahoo export tool.

Updated 2:28 p.m. PT with Facebook declining to comment.

Originally posted at Deep Tech

Physics calculations such as this simulation of particles attracting each other can be accelerated with WebCL. Unaided JavaScript runs the simulation at 5 or 6 frames per second compared to 87 to 105 with WebCL enabled.

(Credit: Screenshot by Stephen Shankland/CNET)

Web applications such as image editors and advanced games could get a new performance boost from a graphics chip's processing power through a technology called WebCL.

Hardware acceleration is all the rage right now among browser makers: it can speed up everything from animating graphics to laying out all the elements of a Web page. Tapping directly into the hardware at a low level not only speeds things up, it saves precious battery power, too.

Now several companies are hard at work on a new interface called WebCL through the auspices of the Khronos Group standards body. WebCL is a browser-specific offshoot of a movement called GPGPU--general-purpose graphics processing unit.

Backers of WebCL aren't afraid to raise expectations.

"Providing millions of Web developers access to high-performance graphics and parallel computation will unleash a wave of creativity that will result in amazingly innovative Web applications that we haven't even imagined yet," said Neil Trevett, vice president of mobile content at graphics chipmaker Nvidia and the Khronos Group's president.

But WebCL advocates still have some persuading to do.

"We don't have plans to integrate WebCL at this point," said Mike Shaver, vice president of technical strategy at Mozilla, which has worked for years to make browsers a better foundation for more powerful Web applications. Its caution is notable given that one of its projects is WebGL, a close cousin of WebCL that's confined just to hardware-accelerated 3D graphics.

"WebGL reflects OpenGL, which is a well-understood and widely used graphics technology with a long history of interoperability," Shaver said. WebGL is based on a very mature Khronos graphics interface standard called OpenGL, but WebCL is based on the much younger OpenCL. "It's much, much earlier in OpenCL's life, and that makes it harder to say whether WebCL will let people write important new kinds of Web apps," Shaver said.

What it offers
The technology--if it comes to fruition, as appears likely for at least some browsers--won't accelerate everything under the sun. It's best suited to mathematical chores that can be broken down into independent pieces running in parallel.

Some examples of what it can offer: image processing algorithms, video game engines to simulate real-world physics, and even mining bitcoin virtual currency. Because such applications come from the Web, they arrive in a browser when you click a link--no downloading an installation file--and don't require programmers to worry about what operating system the browser is running on.

WebCL is part of a bigger trend, showing how the browser world is working hard to match what operating systems can do. That means, in short, that Web apps can be a more viable competitor to native apps. That's a particularly hot competition for developer attention when it comes to how to program apps for smartphones.

WebCL is not just an idea kicking around a standards group.

This Samsung WebCL demonstration of transparent and reflective deformed blobs in a 3D world shows the vastly faster performance of WebCL. It yielded more than 87 frames per second for WebCL vs. 1 or 2 for unaided JavaScript.

(Credit: screenshot by Stephen Shankland/CNET)

The first proposal for the work came from Nokia, which developed a WebCL add-on for Firefox.

And earlier in July, Samsung contributed WebCL features to the open-source WebKit project that provides a foundation for many browsers, notably desktop and mobile browsers from Apple and Google. It's not clear yet which WebKit browsers if any will add WebCL support, but one possibility is the browser in Samsung's Bada operating system, which also is based on WebKit.

The work so far is still a long way from widespread availability--and the standardization process isn't done yet, either. And even when WebCL arrives in browsers, programmers will have to learn how to use it. So don't expect an immediate effect. When it does, though, it'll be useful, Trevett argues.

"Typical use cases would include accelerating a physics engine for a 3D game in a browser using WebGL for the graphics. Or accelerating editing HD photos or videos--smoothly and interactively, right in the browser," Trevett said.

In the cases where WebCL is useful, he predicts a significant help to battery life.

"Using a GPU instead of the CPU for image and computationally intensive tasks that can be highly parallelized can often deliver between 5 and 100 times performance increase," Trevett said. "A GPU is typically also far more efficient at highly parallel computation than general-purpose CPUs--and can deliver the same workload with a tenth of the battery consumption."

How much of that battery-life savings will actually be visible isn't yet clear--especially because game programmers have a tendency to write software that fully exploits the hardware rather than does what yesterday's games do with less power consumption.

How it works
GPUs are designed to accelerate visual tasks, for example by running programs called shaders involved in 3D graphics or converting videos into the color scheme used by a monitor. The work typically involves many similar operations proceeding in parallel.

As graphics chips grew more powerful, chipmakers such as Nvidia and AMD jumped on the idea of letting them perform more work than just graphics.

There are complications--for example, moving data into and out of the graphics chip's separate memory into the memory used by the main processor--but the idea has gained currency in programming circles.

Early attempts to use the technology were difficult, but it's become easier as higher-level interfaces arrived: Nvidia's CUDA (Compute Unified Device Architecture), Microsoft's DirectCompute, and the Khronos Group's OpenCL (Open Computing Language), an interface available in in Mac OS X.

WebCL is an OpenCL derivative that takes the form of a JavaScript "binding," which means it offers a way to utilize OpenCL using the JavaScript language used for Web programming.

WebCL dovetails with the WebGL project for the more limited job of 3D graphics in the browser.

"The performance of JavaScript Web applications will not match highly optimized AAA [top-notch] native games--but the recent optimizations in JavaScript, and the fact that both WebGL and WebCL can be driven very efficiently form JavaScript, enable surprisingly high levels of performance," Trevett said.

WebCL, being an element of JavaScript, also can plug into another mechanism to improve Web-app programming called Web Workers. This lets programmers move some computing work to background processing tasks that don't bog down the user interface, for example, and they're becoming more important as Internet Explorer 10 joins the list of browsers with Web Workers support.

Security matters
Granting software downloaded from the Web access to low-level hardware adds new risks, though.

An illustration: security firm Context Information Security uncovered a WebGL vulnerability, and Khronos allies quickly updated the specification to disable the feature that led to the hole, requiring programmers who'd used it to rewrite their software. Mozilla backed out the WebGL feature with Firefox 5 to sidestep the problem, and Google, which prides itself on Chrome security, is doing the same with Chrome 13. Meanwhile, Microsoft declared WebGL too unsafe to use.

Security is "job one," Trevett said, and its backers are working with hardware companies to make sure their drivers--the software that enables operating systems to use their graphics hardware--are hardened against vulnerabilities. And the Web interfaces have limits not present in their native equivalents.

"WebGL has simplifications and restrictions compared to 'raw' OpenGL ES to enable secure operation. WebCL will go through the same exercise and expose a secure subset of OpenCL to JavaScript," he said.

Don't expect Microsoft to jump aboard, though, given its expressed reservations--and its competing interfaces--DirectCompute and Direct3D for Windows and the new version of its Silverlight browser plug-in. And as Mozilla and 3D expert Gregg Tavares have observed, Silverlight and the Molehill 3D interface in the upcoming version of Flash Player expose the same low-level hardware to downloaded software.

Missing Microsoft support is a problem: when Web developers can't count on a feature being present in all browsers, they must either write different versions of their applications to adjust, or omit support for a particular browser in the first place. The first option is time-consuming, the second is limiting and potentially unpleasant for users.

There's another wrinkle here: Google has its own project, called Native Client, that's geared to let Web applications run with the speed of native software. Thus far, though, Google seems content to support WebGL even though Native Client has a 3D acceleration component, too. In fact, WebCL can even run on a regular CPU, something Khronos says is faster than ordinary JavaScript. WebCL also can tap into multicore processor abilities for work that can be done in parallel tasks.

One possible scenario is that Web programmers familiar with JavaScript prefer WebCL and WebGL, while those from the C and C++ programming camp go with Native Client--if Google can encourage programmers and other browser makers to adopt it.

Sure, it's a bit of a muddle. But given how many parties are collectively developing the future of the Web as a programming platform, a little chaos is to be expected.

"Even written directly to operating systems," Mozilla's Shaver said, "general-purpose GPU computing is still in its infancy."

Originally posted at Deep Tech