Webware

This WebGL demo shows 3D Collada files--in this case a Spore video game creature.

(Credit: Screenshot by Stephen Shankland/CNET)

If you want to see the scale of browser makers' ambition to remake not just the Web but computing itself, look no farther than a new 3D technology called WebGL.

The WebGL vision is simple. You're running around in a video game universe, blasting radioactive aliens--but you got there by visiting a Web site, not by installing the game on your PC.

This sort of computationally demanding chore contrasts sharply to with today's Web, whose top-notch programmers strain to reproduce bare-bones versions of the rich capabilities open to applications running natively on a computer.

WebGL, while only a nascent attempt to catch up, is real. WebGL now is a draft standard for bringing hardware-accelerated 3D graphics to the Web. It got its start with Firefox backer Mozilla and the Khronos Group, which oversees the OpenGL graphics interface, but now the programmers behind browsers from Apple, Google, and Opera Software are also involved.

Perhaps more significant than formal standards work, though, is WebGL support in three precursors of today's browsers--Minefield for Mozilla's Firefox, WebKit for Apple's Safari, and Chromium for Google's Chrome. Opera has started implementing WebGL, too, said Tim Johansson, Opera's lead graphics developer.

With a little tinkering--check the instructions and caveats below--you can give it a whirl, too. Overall, I was favorably impressed with the technology.

Its performance certainly isn't enough for a competitive first-person shooter, but it's approaching utility for casual gaming. And because of how WebGL elements can be integrated with the rest of a Web site's code, it's got some advantages.

What is WebGL?
WebGL is one of a handful of efforts under way to boost the processing power available to Web applications. It marries two existing technologies.

First is JavaScript, the programming language widely used to give Web pages intelligence and interactivity. Although JavaScript performance is improving relatively quickly these days in many browsers, programs written in the language are relatively pokey and limited compared with those that run natively on a computer.

Mozilla has updated its Firefox browser to patch three critical security holes.

Firefox 3.5.6 and 3.0.16 both fix earlier memory corruption issues. "We presume that with enough effort at least some of these could be exploited to run arbitrary code," the security advisory said.

In addition, the earlier version of Firefox 3.5 had two critical vulnerabilities in its technology for playing Ogg-format media, one with the liboggplay media library and one with the libtheora video library.

The patches are among 62 fixes in the new Firefox, software that's translated into dozens of languages and runs on multiple operating systems. Users of the OS/2 operating system will be delighted to know that problems with Firefox's full-screen mode and with print preview have been resolved.

"We strongly recommend that all Firefox users upgrade to this latest release," Mozilla said in a blog posting. By default, Firefox downloads updates automatically then prompts users to restart when it's ready; updates also can be retrieved through the "check for updates" menu option.

Mozilla plans to cease supporting Firefox 3.0 in January. Meanwhile, a significant update, Firefox 3.6, is due by the end of the year.

Correction 1:23 p.m. PST December 17: This story was corrected to note that it was the earlier versions of Firefox that suffered the vulnerabilities.

There's no reason to take the Web as it comes. Not when there are Firefox add-ons that turn Web pages into putty that you can shape as you wish. These three--Zotero, MashLogic, and RSVP Reader--let you gather and store all or parts of Web pages, open a customizable info box for whatever topics you encounter, and convert a block of text into a string of phrases that flash in a box at a speed you control.

Turn your favorite pages into collections
A few days ago, I wrote about three add-ons that go bookmarks one better by letting you customize the Web pages you save. Zotero is like bookmarks cubed. Not only can you save text, images, or entire pages, you can annotate and categorize the information for easy retrieval.

My only complaint is that the Zotero window takes up half the screen and can't be resized. Fortunately, it's easy to close the window to get a full view of your browser. To reopen the window, click the Zotero button in the bottom-right corner of the screen.

Save all or parts of Web pages and categorize the content with the Zotero Firefox add-on.

(Credit: Zotero)

Zotero's capabilities go far beyond collecting and tagging Web pages. It's designed for researchers and lets them attach files and notes to items, take a snapshot of the page, and add bibliographic references. All entries are time- and date-stamped, and you can even open a mini-text-editing window. I sure wish I had one of these when I was a student.

Add-on lets custom search tag along
If you can get past the tiny blue dots the MashLogic add-on places below text and links, the add-on comes in handy. Hover over the dotted item and a small window pops up with information about the item from the sources you specify.

The MashLogic Firefox add-on opens an info box with customizable content related to the item.

(Credit: MashLogic)

Click the MashLogic icon that appears to the left of the address bar to select the sources supplying the add-on's information. Your choices include Wikipedia, New York Times, Twitter, Yelp, and Guardian UK, as well as such categories as movies, books, music, shopping, and news and feeds. You can also suspend the dots for all sites or disable them for the site you're currently on.

Convert a page's text into a video stream
I was hoping to report how much faster I plowed through Web text with the RSVP Reader add-on, but I just couldn't get used to reading words as they flashed in a small box one, two, and three at a time. I still get a kick out of the novelty of a page's text appearing in bits and pieces.

RSVP Reader appears as a toolbar with buttons for making the text larger or smaller, and positioning the text in the box. In addition to the standard Play, Pause, Stop, and Rewind, buttons, you get buttons to speed up or slow down the text playback.

See a page's text by the word or phrase at your choice of playback speed with the RSVP Reader Firefox add-on.

(Credit: RSVP Reader)

I tried reading several text-heavy pages with different types of content (news, literature, even poetry) with RSVP Reader and the old-fashioned way, and even after experimenting with different text-playback rates, it didn't feel like I was going through the material faster the flashing-text-box way. I was disappointed that I couldn't reposition the text box, which is at the far right of the toolbar. But the add-on does offer a totally different way to browse.

3D graphics became ordinary first in games, then in operating systems, and on Thursday, it took a significant step toward being built into Web browsers as well.

The Khronos Group, which oversees the OpenGL graphics interface, announced that its work with Mozilla to bring hardware-accelerated 3D graphics to the Web has reached draft standard form. The standard, called WebGL, lets programmers who use the Web's JavaScript language take advantage of the fact that video cards can handle 3D graphics with aplomb.

The group now wants commentary from Web developers and others who might be involved with WebGL so it can be finalized. "I anticipate us moving toward a spec that is not provisional, not merely a draft, in early 2010, the first quarter," said Arun Ranganathan, chairman of the WebGL working group and standards evangelist at Mozilla.

Internet Explorer remains the dominant browser in terms of usage, but all four of its main challengers--Mozilla's Firefox, Apple's Safari, Google's Chrome, and Opera Software's Opera--are working hard, sometimes in an informal alliance, to get ahead by advancing the Web state of the art.

WebGL fits into that effort, and not just academically. All four of those browser makers have endorsed WebGL, and developer test versions of Firefox, Safari, and Chrome have it built in. Microsoft declined to comment for this story beyond reiterating its general support for standards.

Ultimately, building 3D support into the Web could advance user interfaces of Web applications--including games, the popularity of which can be a powerful incentive for upgrading to the latest technology.

It's not clear exactly how it will play out, though, Ranganathan said. The arrival of Canvas, an advanced 2D interface for browsers, has led to a blossoming of graphics work, and he expects a similar change with 3D graphics.

But don't hold your breath for Web-based first-person shooters that rival native applications. First, even if 3D is accelerated, there are plenty of other processing and user interface constraints on Web applications. Second, even after WebGL is standardized, it must be built into browsers, people must upgrade to those new versions, and programmers must learn how to support the technology.

WebGL isn't the only 3D Web work under way. Google has its own O3D project, which currently is a browser plug-in but that the company also is building directly into Chrome.

O3D is a higher-level interface, though, not a direct competitor. Details are technical, but O3D uses a retained mode approach to WebGL's immediate mode interfaces.

And of course, a decade ago there was VRML--virtual reality modeling language, a file format rather than interface. A VRML successor called X3D, though, can actually make use of WebGL, and indeed a project called X3dom aims to do just that.

Mozilla is wrapping up work on its first version of Firefox for mobile phones, an important step in bringing the second most popular PC browser to an area where a rival project holds more influence.

"Our goal is to have a release candidate next week," said Jay Sullivan, Mozilla's vice president of mobile. "If things go smoothly, we'll have a (final) version out in the next few weeks," with the debut planned for this year, he added.

Mozilla has been a leader in advancing the Web state of the art. But when it comes to the mobile phones, where the power of a new generation of hardware has transformed browsers from primitive afterthoughts to useful tools, Firefox has been missing in action.

Instead, an open-source project called WebKit powers the browser on the higher-end mobile phones du jour--Apple's iPhone, the Palm Pre, and Motorola's Droid and other models running Google's Android operating system, with BlackBerry headed that direction, too.

In contrast, the first mobile Firefox version will run on Nokia's powerful but relatively obscure new N900, a $569 hybrid computer and mobile phone that uses Nokia's Linux-based Maemo operating system. A Windows Mobile version of Firefox is set to arrive next year, and Mozilla has begun working on an Android version now that Google released a native developer kit.

Firefox has one big thing going for it, though: it's a close relative of the PC-based browser that today is used by about a quarter of people on the Web.

The link extends beyond brand familiarity. For one thing, mobile Firefox is based on the same code as the present Firefox 3.6--also a beta version due to finished by the end of 2009. For another, through a Mozilla service and browser plug-in called Weave, mobile Firefox synchronizes bookmarks, passwords, and even open tabs with the desktop version of the browser.

In addition, Firefox for the N900 can run many Firefox extensions--AdBlock Plus among the 30 or so now available. Sullivan recommends updating their interfaces for the small devices, though.

Nokia partnership
Mozilla has been working closely with Nokia to develop Firefox on its N900 handset. It already ships with a lighter-weight browser that uses the same Gecko code base as Firefox, but the full Firefox mobile version--a project code-named Fennec--is more powerful.

"You want to put the desktop experience into a pocket-sized device," said Ari Jaaksi, vice president of Maemo devices. "What do people use on the desktop? Firefox."

However, Firefox won't ship with the device, at least initially. The company is open to the idea of including it in the next version of the operating system, due in the second half of 2010, Jaaksi said. Mozilla, meanwhile, is comfortable with the idea of people having to actively download the browser, the most common way Firefox has been distributed on PCs.

The N900 is available as an unlocked device through various retail channels, but Nokia doesn't yet have any partnerships in the United States with wireless service carriers who might help bring the N900 to a broader market. It's a relatively powerful device with a 600MHz processor and 3D graphics hardware--enough oomph to run Adobe Flash on Web pages today. Its price may seem high, but bear in mind that unlocked devices don't get a subsidy by carriers that expect to see their up-front payment returned over months of subscription payments.

Nokia's N900 will be the first device that runs the mobile version of Firefox.

(Credit: Nokia)

So does Firefox require this level or horsepower?

"We need pretty high-end stuff to make the Web great," Sullivan said, but not so high-end that the N900 is the only handset to fit the bill. "Everything now on the mid- to high-end is fine."

WebKit has intercepted the newer generation of smartphones. Through the wonders of Moore's Law, new devices get steadily more processing power and memory. So aiming for today's top-end phones can mean software will work on tomorrow's mainstream models.

The N900 is at the top end of the range, but Firefox runs elsewhere, too. The Windows Mobile version of Fennec is in alpha testing now, lagging the Maemo version by about three or four months, Sullivan said. Mozilla plans to release it in final form in the first half of 2010, he said.

Firefox: like an operating system?
Mozilla has a lot of plans for mobile Firefox that, to some extent, put it in opposition with Nokia. The N900 is aimed in part at programmers who want to low-level control over a device through its Linux operating system. But Firefox--like Google's Chrome--is assuming the role of a general-purpose foundation for running programs.

"We're almost an operating system," Sullivan said.

Several features support the direction. Built into Firefox now is geolocation, which lets a Web application tap into the phone's services to figure out where a user is and, for example, show a map of the nearest pizza shops. Also included is support for orientation detection, important for games, and offline data storage, important for a variety of programming needs.

There's more on the way in 2010, Sullivan said:

• Support for multitouch displays for a more sophisticated user interface.

• Support for haptic feedback, such as the phone vibrating when a virtual keyboard key is tapped.

• The ability to control a camera.

• Support for Electrolysis, Mozilla's project to split tasks such as the user interface, tabs, and plug-ins into separate processes. That improves stability and performance, he said.

• Support for JetPack, Mozilla's next-generation extensions system.

S• Integration of the Weave synchronization software so it's no longer a plug-in.

• Support for WebGL, an interface to provide browsers with accelerated 3D graphics.

• Faster execution of JavaScript programs that are common and increasingly powerful on the Web.

Applications that run natively on a device--whether directly on the hardware as in the case of the iPhone or on the Java-derived layer called Dalvik on Android--are an important area of mobile development today. But the Palm Pre uses a browser-based application design.

"It's the right model. It's not there yet," Sullivan said of Web-based programs. Today programmers must create separate versions of applications for BlackBerry, Nokia's Symbian, Android, the iPhone, and other mobile phones.

But that profusion will be replaced by the universality of the Web, he predicted.

"In three years," Sullivan said, "80 percent of those applications are going to be Web-based."

Web pages aren't getting any smaller, but there are usually not more than a few paragraphs or a couple of images of particular interest on any given page. Firefox add-ons ICyte (also available for IE), Wired-Marker, and Trails let you save all or sections of Web pages and share your snippets with others.

ICyte makes sharing easy
Most of the time, sharing Web content means sending someone a link via e-mail, chat, or phone. The ICyte add-on for Firefox and Internet Explorer lets you highlight the important content on the page before you share it, or you can save and send portions of the page rather than the whole enchilada.

You must provide your name and e-mail address to use the service. After you download the add-on and restart Firefox, two buttons are added to the left of the address bar. Click the left button to create a Cyte for a new or existing "project." Here you can assign tags or a note to the Cyte. Click the button on the right to open your Cytes in the sidebar.

Annotate Web pages before you save and share them with ICyte.

(Credit: ICyte)

The Cyte entries in the sidebar show a thumbnail of the page, its name, the name of the project, and its comments and tags. When you click a Cyte to reopen it, a banner appears at the top of the main browser window showing the same information along with the date it was saved and a Live View button that returns to the original page. You can hide this banner to view more of the page itself.

Click the gear icon that appears when you hover over a Cyte in the sidebar to open its drop-down menu with options for editing the Cyte name and other data (but not the page itself), creating a copy, deleting the Cyte, sending it to someone via e-mail, or embedding it in a Web page. You can also share the sites you designate as public with others via RSS, Facebook, Twitter, and other social networks, though I didn't try these features.

... Read more

Mozilla, determined to release Firefox 3.6 before year's end, is also determined to squeeze as many features as possible into the new browser.

The latest example: support for the File interface that adds more sophistication to uploading and some other chores.

Support for the feature is one of the 133 changes that arrived in Firefox 3.6 beta 4, which the Mozilla project released Thursday for Windows, Mac, and Linux.

The File API (application programming interface), a draft standard at the World Wide Web consortium, lets browsers handle files better. Among its abilities are uploading multiple files at once, showing thumbnail previews of images that have been selected for upload, breaking a long video upload up into chunks to protect against network interruption problems, and integrating with drag-and-drop Web applications.

While many software projects use beta testing periods to shake down their code, Mozilla isn't afraid to add new features as it goes. That can mean new ideas arrive sooner, of course, but it also can delay the completion date of the new version. What was to have been a quick Firefox 3.1 release was pushed back months as new features were added and the version ultimately was renamed Firefox 3.5.

For those who want to dig into the File interface, Mozilla offers a Web developer guide to using it.

The beta-testing periods aren't just important for debugging Firefox itself. New versions often don't work with older add-ons that people install to customize the browser, so beta testing gives some time for programmers to update those add-ons. Mike Belzner, Mozilla's director of Firefox, said 70 percent of add-ons are now compatible with Firefox 3.6.

Sometimes it's the little things that count.

The most prominent feature of Firefox 3.6 is Personas, which let you reskin the browser with thousands of different looks. But my single favorite change is a subtler change to the open-source browser's user interface.

Specifically, when you open a link in a new tab, it appears immediately to the right of the active tab. Before, the new tabs would appear to the far right of the strip of tabs.

Yup, that's it. For those of us who spend hours a day in a browser, though, the new tab behavior helps group related tasks together. I constantly shuffle among dozens of tabs, and the new approach automatically brings some organization to my cluttered life.

However, I know it's not everybody's favorite browser behavior. So along with explaining why I like it, I'll also take some potshots and share instructions on how to get the old way back.

Why it's better
The more things I do with a browser--and the number has increased steadily for years now--the more important it becomes to be able to find different tasks amid the chaos. Microsoft and Apple understand this, as evidenced by the new taskbar features in Windows 7 and dock expose in Mac OS X 10.6, aka Snow Leopard. Those features make it easier to pluck out the one window you need from among the many you may have open.

There's a pattern to how I spawn the dozens of tabs I use as a day progresses. On a variety of pages--Gmail, Google Reader, Yahoo Finance, somebody's blog post--I'll encounter a host of links to other pages. I'll middle-click my mouse button to open interesting pages as background tabs, then use Ctrl-Tab to switch to the new pages when I'm ready. I repeat this pattern many times a day.

With the old behavior, each tab appeared to the far right of the tab strip. That's fine when getting started, but when I've moved halfway across the list and want to open another batch, I want the new ones--call them children--to open next to their parent tab. When I go away and come back, or when I lose place juggling tasks, it's easier to find my bearings again.

It's like being in a library. When you're in the European history section, you don't want to find books on rewiring your house and on vegetarian cooking.

As a longtime Firefox user, I didn't realize tab positioning could be better. When I started using Google's Chrome, which introduced the new tab behavior to me, the scales were lifted from my eyes. I immediately could get to the next tab with a quick press of Ctrl-Tab on the keyboard rather than have to use the mouse to click over to the far end of the list. I use both browsers daily, but until the Firefox 3.6 beta arrived, the new-tab position had become a sore point for me when in Firefox.

The change is actually a big deal in a couple ways. First, even seemingly minor changes in software can be disruptive. Old habits die hard, and computer users wrestling with constant change can get angry when more is foisted upon them.

Second, though, browsers are assuming an ever greater role in what people do in their personal and professional lives, and keeping one's bearings is commensurately important. That's especially true for those people for whom a gaggle of browser tabs represents a collection of chores going on in parallel.

Internet Explorer 8 categorizes related tabs by color.

(Credit: Screenshot by Stephen Shankland/CNET)

How the competition handles it
Tabs are now universal among browsers, but new-tab behavior isn't. Firefox and Chrome handle it the way I like best, but how do others tackle the issue?

First, let's look at Internet Explorer 8. Microsoft showed it understands some of the challenges of tab management in its latest version of its browser by coloring child tabs the same hue as their parents, but I have a gripe with how it works. Specifically, although child tabs get the same color as their parents for easy grouping and arrive to the right, grandchild tabs are the same color as child tabs. Similarly, grandchild tabs appear to the far right of the whole group of child tabs.

In my mind, I consider grandchild tabs a separate group from the child tabs. But with IE, grandchildren get the same color and position treatment as children. The only way to get a new color is to start a fresh empty tab There's no easy way to give grandchildren a new color without causing some confusion, though--should the child be the same color as the original parent or change color to be grouped with the grandchildren?

Next is Opera, which gives users a choice. By default, it opens new tabs to the far right, which I don't like, but in the Advanced|Tabs section of the preferences dialog box, you can check "Open new tab next to active." Huzzah!

There's a subtle change here I don't care for, though. Tabs always appear immediately to the right of the active tab. I'd rather have all one tab's children appear in sequence to the right. For example, if a parent tab is in position 1, then the first child would be in position 2, the second in position 3, and the third in position 4. Opening three child tabs in Opera leaves the parent in position 1, the third child in position 2, the second child in position 3, and the first child in position 4.

Last, there's Safari. It does it the old way I loathe with no option to change. Too bad.

Firefox can show thumbnail previews of new tabs, but I find them hard to recognize in front of busy Web pages.

(Credit: Screenshot by Stephen Shankland/CNET)

Why it's not enough
Most browser makers are excited about the fact that their software is subsuming more and more computing tasks that previously ran on computer operating systems. But as browsers inherit this central importance, they also inherit some of the hassles.

The new tab positioning behavior in Firefox is a step in the right direction, but there's more that needs to be done. Moving from one tab to a related adjacent one, whether through a keyboard command or mouse clicking, is a minor change. But things get harder when you need to switch from one group of tabs to the next.

There's work under way here. Opera is perhaps the leader with the ability to show thumbnails as you use Ctrl-Tab to cycle your list of open tabs.

Firefox has been noodling with the approach too. It tried then dropped tab thumbnail previews earlier, but the technology is still present. Using the about:config system for tweaking the browser (more on this later), you can change the "browser.ctrlTab.previews" setting to "true."

But for reasons that aren't clear to me, I don't find this effective either in Firefox or Opera. Perhaps I haven't used it enough, or the thumbnails are too small to be immediately recognizable, or they're just hard to see against the noisy background. There's a good reason that Apple dims the background most of the way to black when using Expose.

Aero Peek in Windows 7 lets the task bar show a glimpse of Firefox and IE tabs.

(Credit: Screenshot by Seth Rosenblatt/CNET)

Windows itself is helping, too. The new taskbar in Windows 7 can show individual tabs, once browsers support the feature. It's in Internet Explorer 8, and it's in the new Firefox 3.6 beta.

Add-ons such as Firefox Showcase can further tweak Firefox. (Indeed, for a wealth of options, check Mashable's handy Firefox tab management guide.)

More interesting to me, though, is work under way to expand Firefox's "awesome bar" abilities. Today, typing in it opens Web pages and retrieves ones you've already visited or bookmarked. In the future, it could be able to move you to another open tab, too. I'm a keyboard guy, so particularly appreciate this idea.

You can get a taste of the idea now. If you've enabled the "browser.ctrlTab.previews" option, hitting Ctrl-Shift-Tab will not only show you thumbnail previews, but will put a cursor in a search box.

Typing the letters of the Web page name will winnow down the thumbnails. For example, typing "netap" will cull my open tabs so only Net Applications and NetApp show. If you have a bunch of similar tabs all open, this might not help much, of course.

However, the feature only works with the tabs of one browser window, so if you can't use it to search among other browser instances.

How to get the old way back
Perhaps I've convinced you that the new approach is better. But perhaps not--in which case I encourage you to share your thoughts in the comments so people will hear more than my opinion.

For those who don't like the new tab positions, you can revert to the old method.

To get the old style back for new tab position, use Firefox's about:config system.

(Credit: Screenshot by Stephen Shankland/CNET)

First type "about:config" in the Firefox address bar. You'll get a warning that you're tinkering with Firefox's innards and you should be careful, but this isn't brain surgery, so don't be frightened. Click the "I'll be careful, I promise" button, and you'll see a big list of all the browser settings that can be tweaked.

Next, in the text box labeled "Filter:", type "tabs.insertRelatedAfterCurrent"; you should see just one entry below. In the column marked "Value," double-click on the word "true" to change it to "false." You're done.

But I'd encourage you to at least give the new way a try. If you don't like it, you can always change back.

The World Wide Web Consortium has published a draft of an interface that browsers can use to manipulate files better, one of a series of steps aimed at gradually improving the sophistication and polish of Web site interfaces.

The draft File API (application programming interface) defines a number of ways that browsers and Web sites can handle files better. One big part of it: being able to select multiple files for upload, such as on photo-sharing sites or Web-based e-mail, a task that often relies on Adobe Systems' Flash today.

But there are other aspects, too. For example, the Files interface governs the use of "blobs," or packages of raw binary data such as video files. Google has touted blobs for its Gears browser plug-in as a way to divide large videos into small chunks so that uploads can be more easily resumed if a network problem interrupts the process.

Another benefit: files are handled asynchronously, which means the browser won't freeze up while a file is being uploaded or otherwise handled, and the browser reports progress on file transfers.

The technology is one example of work to transform the Web into a better foundation for interactive applications, a move that usurps some power from computer operating systems such as Windows and that's embodied most boldly in Google's Chrome OS project.

Here's one example of use of the Files interface provided by Mike Smith, who works for the W3C on matters relating to HTML--Hypertext Markup Language, the language used to describe Web pages:

A user uses a Web-based application for reading and sending e-mail. She wants to attach multiple files to particular messages. The Web application provides an user interface that allows her to select multiple files to attach at the same time. After she selects the files, they are uploaded to the Web application asynchronously, allowing the user to perform other actions while they are uploading (for example, finishing the rest of the message she was composing before you added the file attachments). As the attachments are uploaded, the Web applications shows progress bars to indicate how much of the contents of the files have uploaded thus far.

The interface can work in conjunction with various standards including the drag-and-drop support in the HTML 5 now under development and the Web Workers technology that lets browsers better perform multiple operations simultaneously.

The interface also can help Web applications process the contents of files. For example, Smith describes a lyrics finder:

A user has on her local file system a playlist file from her favorite desktop music player. The playlist contains a list of song titles and information, and she wants to be able to easily fetch the lyrics for particular songs without needing to manually search for the lyrics on the Web. So a site can provide a Web-based application that allows her to upload her playlist. The Web application then parses the file and then presents a user interface to her, show in the contents of the file as a hyperlinked, sortable list. She can then retrieve the lyrics for any given song just by clicking on a particular song title.

Arun Ranganathan, Mozilla's standards evangelist and chairman of the WebGL working group, wrote the specification, according to Chris Blizzard, Mozilla's director of developer relations.

Standards for the Web are advancing rapidly with W3C representatives including Microsoft working in conjunction with a parallel effort, WHATWG. New standards require actual implementation in browsers before they are accepted as finished, a fact that can lead to some chaos but that helps ensure the new ideas are tested in the real world.

Firefox 3.6, in beta testing now, will support most of the Files API, according to Blizzard.