Internet

The Mozilla organization has shipped Firefox 6, eight weeks after the release of Firefox 5. Just as with Firefox 5, not a lot has visibly changed. The domain name in the address bar is now highlighted, to make phishing more apparent—mimicing a similar feature already found in Internet Explorer—sites with "extended verification" certificates appear slightly differently in the address bar, and Mozilla is claiming that there's some speed improvement. And that's about the extent of it. More substantial improvements are in the pipeline for Firefox 7—most notably a JavaScript engine that uses much less memory—but nothing so substantial is evident in version 6.

This smaller release—bug fixes, behind-the-scenes improvements, but little user-visible difference—is likely to be the norm for future Firefox versions. Bigger features will still arrive from time to time, but for the most part, users will just experience a continuous improvement. Firefox updates should be automatic and essentially invisible. Even articles such as this one, which attach some significance to the new release, are probably not what Mozilla wants—press coverage should focus on features, not version numbers. Mozilla—as with Google—wants developers to cease targeting specific browser versions, and instead target standards; the regular releases are one step towards achieving that goal.

Google has released a new service entitled Google Related, a "browser assistant" Chrome extension intended to direct users to webpages on the same topic as the one they're currently viewing. While some applications of the service, like getting extra info during a restaurant search, are useful, some others produce unhelpful suggestions in a framework that should be more trainable than it is.

Once you have the Google Related extension installed, a bar will begin appearing along the bottom of certain types of pages, such as news, shopping, or restaurant websites. Various tabs allow you access to content related to that page—visiting a restaurant's website may produce a tab with a Google map of the restaurant's address, a second tab with reviews, and a third tab of related locations (as identified by Google Maps).

The restaurant website suggestions are the most coherent, as the previous list nails exactly what I'm looking for when I look up a restaurant. But some of the tabs are too selective and Google-centric (unsurprisingly), as when the Reviews tab produces Google Places reviews and links to the Urban Spoon page, but not to Yelp.

A Google Related tab produced from a news story concerning an HTC vs. Apple patent spat.

Visiting a page with a news story will produce a dropdown (or more accurately, a shoot-up) of culled news stories on the same topic from other sources. The displayed stories seem limited to the most recent updates you might find at the top of a Google News search, a format better for the rarer breaking stories than authoritative ones getting picked up over and over in brief by multiple news outlets. The pullquote in the HTC vs. Apple-produced tab above is a nice feature, but the content is only barely related to the story.

What the extension lacks the most is the ability to train it. Links offered from the Related bar are +1-able, but if you click the "View More Articles" link from the story above, you get a get a long list of stories from various outlets that can't be +1'd. This strikes us as a prime opportunity to teach Google Related which sources you trust or would like to see in your related news tab when you visit a news story. Still, true to Google form, Google is collecting statistics on the project, so we may be training it more than we know.

Given Google's recent "more wood behind fewer arrows" declaration, the only-partially-useful Related is a mystifying addition to the company's product slate in its current state. The extension is available today for all Chrome users.

Read the comments on this post

Google has issued a beta release of Chrome 14 that includes its Native Client (NaCl) framework. The feature was previously only available through a special browser flag, but will finally be enabled by default in the next major version of the Web browser.

Google first introduced NaCl as an experiment in 2008. It allows developers to compile C or C++ code into a platform-neutral binary that is executed by a browser-integrated runtime, which uses sophisticated sandboxing techniques to avoid the historical security pitfalls of Microsoft's much-reviled ActiveX. NaCl also provides a messaging mechanism so that functions in compiled NaCl binaries can be called from JavaScript.

NaCl makes it possible for Web applications to use high-performance native code instead of JavaScript for computationally-intensive operations. This will open the door for more sophisticated games and software to operate within the Web browser. NaCl is particularly significant for Google's Chrome OS platform, which relies solely on browser-based software.

The original implementation of NaCl suffered from some major technical problems that seemed difficult to overcome. In particular, the sandboxing mechanism relied on certain characteristics of the x86 architecture. That issue has since been addressed; it now has 64-bit support and experimental ARM compatibility.

Although NaCl has matured considerably, it hasn't seen much developer adoption due to the fact that it was only available through a special about:flags option in Chrome. Now that Google is flipping the switch and planning to roll it out to users in Chrome 14, we could start to see some adoption.

It's unlikely that NaCl will ever truly become mainstream, however. Google has opened the source code and is encouraging other browser vendors to support the technology, but none have expressed much interest. Mozilla doesn't intend to implement NaCl in Firefox for a variety of technical and philosophical reasons. Google could potentially ship it as a plugin for other browsers if they decide that they want it to reach a broader audience, but such a move would likely be viewed negatively by the Web standards community.

In addition to NaCl enablement, Chrome 14 will also bring support for the Web Audio API, which enables Web applications to process and synthesize audio. For more details about the release, you can refer to the entry in Google's official Chrome blog. The beta is available for download from Google's website.

Read the comments on this post

Amazon has finally launched a Web-based version of its Kindle reader, allowing users to read their Kindle purchases on any device with a Web browser, without having to download special software. The Web-based reader, called Kindle Cloud Reader, sports both an online and offline mode and can even sync your last page read (among other things) across Kindle devices. But while Kindle Cloud Reader could be used by almost anyone, it's clear that the design was largely made for the iPad, allowing Amazon to sidestep Apple's in-app content constraints and even offer the iPad-friendly Kindle Store it couldn't implement in its native iOS app.

Since I'm an avid Kindle book reader on my iPad, I gave the new HTML5-based reader a spin on that platform to see how it differs from the native Kindle app. Upon logging in via read.amazon.com (which can be accessed from a browser, iPad or not), Amazon warns you that you'll have to authorize the Web app to expand its local database size up to 50MB so that you can download and store books for offline reading. If you're worried about the space, you don't have to store books offline—you can read them directly from the "cloud," so long as you have an Internet connection. However, as most Kindle users can attest, there are plenty of times in which you are likely to find yourself in the mood to read but without a connection—the train, the beach, during a power outage—so the offline feature is a definite plus.

Walmart is pulling the plug on its MP3 downloads store, but it will continue to support DRMed tracks that it sold before the store went DRM-free. The news comes via a leaked memo to Digital Music News—later confirmed by a Walmart representative—which told music licensing partners that the store would close on August 28, 2011.

"After eight years in business, the Walmart Music Downloads Store located at mp3.walmart.com will close on August 28, 2011. All content in the Store will be disabled and no longer available for download from the store," reads the memo. The also company reassured its partners that the sale of physical media through Walmart would be unaffected.

Walmart's music store started out as a DRMed venture in 2004, but went DRM-free in late 2007 (as it was the fashion at the time). Though Walmart was one of the first major names to get into selling DRM-free music, it never really caught on—Apple soon became the number one music retailer in the world, and Amazon has managed to push its way to number two thanks in part to its own MP3 music store. Both Amazon and iTunes offere a better user experience too, with software for multiple platforms and a large built-in user base. Let's face it: for mainstream music, why would anyone choose Walmart over the competition when it comes to music downloads?

Alas, Walmart's memo to its partners doesn't go into detail as to why the company decided to shut down, but it seems obvious that Walmart just couldn't compete without something more to offer. The company did say, however, that DRMed music files that were purchased before the store went DRM-free would continue to function even after the MP3 store shut-down. The company had attempted to kill its DRM servers in 2008—leaving its previous customers out in the cold if they were to need re-authorization—but later decided to leave the DRM servers online after receiving bad customer feedback.

Read the comments on this post

The Khronos group unveiled on Monday the OpenGL 4.2 specification, the newest update to the cross-platform 3D graphics API.

The new spec includes a handful of new features designed to improve the performance of common operations in order to allow developers to more easily create high-performance 3D software. These features include fine-grained modification of textures, more efficient data sharing between shader programs, and the ability to re-use shapes that the GPU has already tessellated without having to tessellate them again.

In common with other OpenGL 4.x point releases, the new features should be supported by existing OpenGL 4/Direct3D 11-class hardware, requiring only driver updates to provide the new functionality. NVIDIA released OpenGL 4.2 drivers today, while AMD intends to release a driver with beta support shortly.

OpenGL is Khronos' best-known specification, but the group has broader goals. It has already released the WebGL specification, a 3D API for HTML5; to that the group announced plans to add a new WebCL specification. Just as WebGL provides OpenGL-like graphics to Web developers, WebCL will provide heterogeneous OpenCL-like computation to Web developers, giving them access to GPU-accelerated computation where available, with the ability to fall back to the CPU where necessary. This will allow, for example, HTML5 games to include physics engines, and for complex image and video processing.

WebCL will be developed by the WebGL working group, and prototype implementations from Nokia, which spearheaded the work, and Samsung are already available.

WebGL came under criticism from Microsoft in June, with the company claiming that the technology couldn't be implemented in a secure way. Khronos representatives dispute this claim, pointing to security features that OpenGL has added to prevent WebGL denial-of-service attacks and the adoption of the cross-origin resource sharing (CORS) specification to prevent information leakage between domains. The group acknowledges that driver bugs can still be an issue, but argues that browser whitelists and blacklists can safeguard users. Khronos says adoption of the technology will spur driver developers to fix their software.

Khronos also announced plans to create a general sensor API. This is in order to provide consistent high-level access to a wealth of sensors that computers and smartphones are increasingly being equipped with—accelerometers, gyroscopes, compasses, cameras (including depth-tracking cameras such as those found in Kinect), microphone arrays, and more. There's no generic API for accessing all these kinds of data, nor for ensuring it's all kept synchronized. Khronos has created a new working group called StreamInput to work on this problem.

Read the comments on this post

If you're a stranger who follows me on Google+, you might think I rarely use the service. That's because the majority of my posts have been limited to the seven circles I created for friends, acquaintances, family, Ars staffers, and other people I like to expose to various aspects of my personality. You had no idea? That's exactly the point.

After one month with Google+, it's clear to me that this—sending updates to certain groups of people and not to others—is the main appeal of the service. I was one of the first people to loudly declare that you can do the same thing on Facebook, but so few people know this that it's basically a nonexistent feature; that's the problem with Facebook. With Google+, sending out certain updates to some people and other updates to other people is right at the forefront of the experience. You are always asked to make a conscious decision about your social circles and about which circles get to see which posts.

Notorious spam-monger Sanford Wallace has been brought up on federal fraud charges for allegedly using phished Facebook accounts to send 27 million spam messages in 2008 and 2009. If the charges hold, Wallace could serve more than 16 years in prison.

Since forming spam company Cyber Promotions in 1995, Wallace, a proud defender of sending unwanted commercially oriented e-mail, has faced civil lawsuits from entities including MySpace in 2007 and Facebook itself in 2009. The judgments from those cases demanded from Wallace sums of $230 million and $711 million in damages, respectively. This marks first time Wallace is facing criminal charges.

In his latest spam crusade, Wallace allegedly made money by creating posts on victims' Facebook walls to drive traffic to affiliate marketing companies. He was banned from Facebook as part of the earlier Facebook judgment, but created a profile in January of this year under the name David Sinful-Saturdays Wallace. Part of the criminal indictment this time around accuses Wallace of contempt of court for logging into this profile in April while on a Virgin Airways flight from Las Vegas to New York.

Wallace was released Thursday on a $100,000 bond, and is set to appear in the US Disctrict Court for the Northern District of California in San Jose on August 22.

Read the comments on this post