Practical Technology

for practical people.

September 7, 2007
by sjvn01
0 comments

AMD partners with Novell to open source ATI graphic drivers

AMD announced on Sept. 7 a major strategic change in open-source graphic processors support.

The company announced it would provide open-source information and a development package supporting the ATI RadeonHD 2000 series ATI Radeon X1000 series of graphics processing units on Linux desktops.

Beginning the week of Sept. 10, AMD and Novell‘s SUSE Linux engineering team will join forces to release the needed source code and hardware specifications to create open-source 2D graphics drivers for the Radeon chip family. Over the following months, AMD will continue to work with the open-source community to enable 2D, 3D and video playback acceleration to provide the best possible experience on the Linux desktop.

At the same time, AMD will continue to work on ATI’s proprietary, high-performance Catalyst drivers. With the forthcoming Catalyst 7.9 software release in September, AMD will add Linux support for the ATI Radeon HD 2000 series of graphics processors. In addition to expanded GPU support, Catalyst 7.9 is designed to improve performance across the board. In Q4 2007, AMD’s Catalyst software package for Linux will add support for AIGLX (Accelerated Indirect GLX). AIGLX is an enabling technology that allows Linux users to enjoy a rich visual 3D user interface with the 3D Compiz Linux desktop.

An early preview of this driver by the reviews site Phoronix showed that even in beta the AMD fglrx 8.41 driver gives Linux users “truly a new experience on the ATI Linux front. The new driver delivers massive performance improvements.”

“AMD has a proven track record of collaboration with the software ecosystem to support the needs of the open-source community,” said Phil Hester, AMD’s senior vice president and CTO at AMD, in a statement. “With this announcement we’re demonstrating our commitment to respond to the needs of the open-source community and adopt a consistent approach across CPU and GPU technologies. By supporting open-source drivers on our industry-leading ATI Radeon graphics processors, we accelerate innovation across the entire AMD platform.”

Ever since AMD, an open-source supporter, bought ATI in July 2006, there has been speculation that AMD would open source ATI’s proprietary graphics drivers.

As John Cherry, the Linux Foundation’s global initiative manager, blogged recently, “During the state of the Linux round-table discussion on the first day of the Linux Foundation Collaboration Summit, James Bottomley [Linux kernel developer] had asked the panelists what are the top two things each panelist would like from the Linux community. Among the panelists was Google’s Chris DiBona, who is the open-source program manager at Google. His response was interesting when he had said the following: ‘I would love to get either NVIDIA and ATI to actually give us the specs on the drivers we want or let’s just reverse engineer everything and do it ourselves.'”

AMD, with these moves, seems to be actually delivering more than the minimum that DiBona had requested. Novell and AMD have worked closely together for a number of years to bring technology to the open-source market.

“The lack of open-source drivers for graphics hardware has long been a major obstacle for Linux developers and Linux desktop users,” said Nat Friedman, chief technology and strategy officer for open source at Novell. “Our ultimate goal is complete, high-quality, open-source drivers for all video hardware. Today’s announcement brings us one major step closer. Novell is pleased to have extended our collaboration with AMD to deliver the initial drivers that will allow open-source developers to make the Linux experience even better on desktops, laptops and workstations, and we look forward to contributing this initial code to X.Org.”

As part of opening up the ATI drivers, AMD consulted with leading members of the open-source community and the X.Org community to ensure that the driver addressed the needs of both open-source developers and Linux users. AMD hopes that the release of these specifications will allow the development community to partner with AMD to drive new innovations for the Linux desktop.

AMD will provide on-going support for this project with an in-house development team.

“AMD announced its intention to extend support of open source ATI drivers at Red Hat’s Summit conference in May of this year, so we are delighted that the work necessary to achieve this has come to fruition,” said Brian Stevens, Red Hat’s CTO, in a statement. “This action affirms AMD’s commitment to Linux and the open-source community and will greatly improve the ‘out-of-box’ experience for users. AMD’s decision has a strong supporting business case, demonstrating that Linux clients represent a significant market opportunity. It is a bold
decision for AMD, and they deserve full credit.”

Mark Shuttleworth, Ubuntu’s founder, added, “The combination of AMD’s graphics hardware and open-source innovation will accelerate the emergence of new desktop computing technologies. AMD’s timing is excellent, given the surge in Linux desktop adoption and the focus now in the open-source community on next-generation visualization and desktop interaction models.”

A version of this story first appeared in DesktopLinux.

September 6, 2007
by sjvn01
0 comments

Samba 4 Moves Closer to Active Directory Server Compatibility

For years, if you wanted an inexpensive, but Windows-compatible file and print server, you turned to Samba running on Linux. Now, with the first alpha release of Samba 4, this open-source project is moving closer to becoming a complete Windows 2003/Longhorn replacement.

Since Samba 3 arrived in 2003, Windows network administrators have been able to use Samba and Linux as a drop-in replacement for an NT file/print server. You could, and many have, used Samba in place of an NT PDC (primary domain controller).

This enables LAN managers to run small networks’ several dozen Samba and Windows servers and several hundred Windows clients using Samba alone.

With Samba 3, you could also use Samba servers in a Microsoft AD (Active Directory) in both native and mixed mode. What you could not do, however, is run Samba as an AD domain controller or run it in a Windows Server 2003 level Forest or Domain.

For practical purposes since many Windows LANs run in mixed mode to allow the use of older versions of Windows 2000 or 2003 even if Samba isn’t present this means that Samba couldn’t be used in enterprise-level WANs. Typically, it’s only in these most complex of corporate networks that companies deploy a Windows Server 2003 Forest.

Now, Samba is getting closer to being able to replace Server 2003 even in corporate WANs. As Andrew Barlett described in his 2005 paper, Samba 4 – Active Directory (PDF Link), “Samba version 4 is … a massive leap forward in the way Samba is designed and built. This thesis attempts to take that further, but examining the protocol basis and implementation details adding support for hosting the Kerberos network authentication system into Samba4’s partial implementation of an Active Directory Domain controller. Active Directory forms the heart of Microsoft’s modern network architecture, and is the heart of many corporate networks. Producing a compatible product is important, if the Samba project is to remain relevant into the future.”

The problem that Samba faced in particular, was dealing with AD authentication protocols. Microsoft’s proprietary NTLM2 (NT LAN Manager) and Kerberos extensions make creating an open-source, white room compatible AD replacement a difficult task.

Nevertheless, the Samba developers have moved closer to supporting the Active Directory logon protocols in this release. According to the programmers, “Samba4 alpha1 is the culmination of four and a half years of development under our belt since Tridge [prominent Samba developer Andrew Tridgel] first proposed a new Virtual File System layer for Samba 3–a project which eventually led to our Active Directory efforts–and one and a half years since we first released a Technology Preview. We wish to allow users, managers and developers to see how we have progressed, and to invite feedback and support.

However, the Samba Group warns would-be users that, “Samba 4 is currently not yet in a state where it is usable in production environments. Note the WARNINGS in WHATSNEW.txt in the source and the STATUS file which aims to document what should and should not work.”

A version of this story first appeared in Linux-Watch.

August 31, 2007
by sjvn01
0 comments

Adobe adds H.264 to Flash Player 9

n what promises to boost video quality in desktop, mobile, and embedded devices, Adobe announced that it has added H.264 support to its popular Flash Player software. “Flash Player 9, Update 3 beta,” available for download from Adobe’s Labs website, supports Linux, OS X, and Windows.

H.264, also known as MPEG-4 Part 10 (ISO/IEC 14496-10), is already widely used on the Internet, and is also the mandatory format for the HD-DVD and Blu-ray video disc formats. (An IEEE overview of the standard can be downloaded in PDF format here.)

In a statement on its website, Adobe Labs said, “H.264 delivers excellent video quality across the entire bandwidth spectrum, from 3G (mobile phones) to HD (broadcast) and everything in between.” The broadest distribution of H.264 has been via Apple’s QuickTime, included in that company’s iTunes, iPods, and Windows/Macintosh QuickTime Player, Adobe notes.

Now, says Adobe, H.264 has been added to a new beta version of its Flash Player. According to the Adobe Labs wiki, the software will now support high-definition 480p, 720p, and 1080p content, encoded with either On2 or H.264.

The company says H.264 will give users higher-quality, at lower bit rates. However, for backward compatibility, Flash Player will also support its previous video formats, such as the Sorenson Spark video codec (based on H.263) and On2 VP6.

With this update, Adobe Flash Player also branches out from its native .FLV file format. The company says it will now support MP4, M4A, MOV, and other files derived from the standard MPEG-4 container format, as long as they contain either H.264 video and/or HE-AAC encoded audio.

Other new features in Flash Player 9, Update 3 (code-named “Moviestar”) are said to include:

* Multi-core support for vector rendering
* Full screen mode with hardware scaling.
* Flash Player cache for common platform components, such as the Flex framework
* Support for Microsoft Active Accessibility (MSAA) in the Windows plug-in

Playing HD video will require a relatively new PC, Adobe concedes. Stated hardware requirements are a Pentium-class processor clocked at 2 GHz or better, 32 MB of RAM.

For an “optimal experience” and full-screen video, Adobe recommends a 3 GHz PC or better, or a 2 GHz Macintosh. The above unretouched image was captured from Adobe’s demo video using a 2 GHz Core Duo system with integrated graphics.

The Flash Player 9, Update 3 beta is available from the Adobe Labs website, here. Linux versions are available for download in gzipped tar and rpm formats. Mac OS X and Windows versions are also available.

A version of this story first appeared in DesktopLinux.

August 30, 2007
by sjvn01
0 comments

Open XML standard war grows heated

The day is fast approaching when the comment and voting period for ISO/IEC DIS 29500, the draft ISO specification based upon Microsoft’s Office Open XML formats, will either be approved or not. As Sept. 2 comes closer, Microsoft appears to be stuffing the ballot boxes of some countries’ ISO organizations while open-source and standard organizations are firing back with furious words.

In Sweden, 23 new companies, all of which were Microsoft partners, joined the Swedish Standards Institute at literally the group’s closing meeting on the subject and were allowed to vote in favor of making Microsoft’s Open XML a standard. One anti-Open XML attendee said that Sweden’s vote had been hijacked.

It wasn’t just the forces that were arrayed against Open XML that cried foul. A pro Open XML supporter, Wictor Wilén, reported, “We all entered the meeting at the last possible minute, and we all was signed in to the meeting.”

Then, “starting with IBM, a number of representatives ran out of the conference room mumbling and cursing when they realized in which direction the vote would go—and the vote result was clear to most of us.” The vote was a forgone conclusion. “The vote went quick, and it was 26 votes for yes and six for no.”

Similar Microsoft vote-stuffing activities have also been reported in Denmark and Norway.

Microsoft also released a Microsoft-commissioned study on Aug. 25, by IDC, on “Adoption of Document Standards.” It contains such statements as “Organizations do not put emphasis on discussions about the ‘openness’ of standards. Instead, more practical aspects are rated highly: Cost is very important as is the ability to have an easy transition of existing documents to a new standard.” Since many organizations use Microsoft Office, the clear implication is that Open XML is the better standardization choice.

Andrew “Andy” Updegrove, a partner with Gesmer Updegrove LLP, a Boston law firm, and the editor of ConsortiumInfo.org, rhetorically asked, “The OOXML Vote: How Bad Can It Get? (Keep Counting).” From where he stands, it can get very bad indeed.

Updegrove opens by quoting Microsoft’s director of corporate standards Jason Matusow on Microsoft’s attempt to turn its format into an international standard, “There is no question that all over the world the competing interests in the Open XML standardization process are going to use all tactics available to them within the rules.”

Updegrove then reports that there is a “sudden surge of interest among ISO members in upgrading their privileges to ‘P’ status, which will entitle to them (just in time) to a more influential vote on OOXML.” He continued, “When I first noted that I had heard concerns over upgrading at the global vote level, only two nations had upgraded. When I wrote about it the second time, that number had risen to six. It’s now only a few days later, and the number has risen to nine (bear in mind that the original number was only thirty). And there are still a few days left during which stealth countries, their votes already taken, can make the cut. Where will it all end?”

His fear is that it will end when Microsoft has manipulated the process to the point where Open XML will become a standard. “As someone who has spent a great part of my life working to support open standards over the past 20 years, I have to say that this is the most egregious, and far-reaching, example of playing the system to the advantage of a single company that I have ever seen. Breathtaking, in fact. That’s assuming, of course, that I am right in supposing that all of these newbie countries vote ‘yes.'”

He assumes, as you might guess, they will vote for Microsoft. Other countries, such as the U.S. and Germany, have already voted yes for Open XML.

Against Microsoft and its supporters, open-source and standard organizations are fighting with words. The Linux Foundation, which heretofore has not entered into the fray, is now calling upon the national bodies that have not yet cast their votes to vote “No, with comments.”

The LF explained its venture outside of Linux matters by stating that “central to the mission of the Linux Foundation is the creation of standards that become widely adopted. Therefore, the Linux Foundation is not only familiar with, but has a vested interest in, the preservation of the validity and integrity of the global standards adoption process. When that process works well, everyone wins. The modern world has become utterly dependent upon technology, and therefore upon the ability of standards organizations to provide interoperability and other open standards as well. With the conversion of paper documents to digital formats, the world has also become utterly dependent upon the ability of those documents to be accessed in the future. Creation of documents in proprietary formats at best jeopardizes that ability and at worst guarantees that easy access in the future will be impossible.”

Updegrove, it should be noted, recently became a legal adviser to the LF.

The LF statement concluded, “Finally, the Linux Foundation notes that there already exists an ISO/IEC standard intended for a similar purpose—the Open Document Format—that has been implemented in at least a dozen products, both open source as well as proprietary. These products have been developed and released by multiple vendors (including several Linux Foundation members). While the current voting in ISO/IEC JTC1 is based upon the technical merits and issues relating to OOXML, the Linux Foundation believes that the marketplace would be better served by all vendors—including Microsoft—uniting around the implementation and further development of a single, common specification.”

The Linux Desktop Architects have also taken a public stand against making Open XML a standard. John Cherry, the global initiative manager, explains why technically speaking Open XML is not standard worthy. “OOXML is a direct port of a single vendor’s binary document formats. … It lists a large number of ‘Compatibility Settings’ for legacy applications (i.e. footnoteLayoutLikeWW8, autoSpaceLikeWord95, useWord97LineBreakRules, etc.) which would be difficult for other developers to implement and hardly what you would find in an aspirational, consolidated best practices document. There are literally 100s of technical flaws that should be addressed before standardizing OOXML, including continued use of binary code tied to platform-specific features, propagating bugs in MS-Office into the standard, proprietary units, references to proprietary/confidential tags, unclear IP and patent rights.”

So from where Cherry sits, while the Open XML may be standard material someday for now, “OOXML is simply not ready to become an ISO standard.”

Speaking for Google, Jeremy Allison, the well-known Samba architect, and Dan Kegel, a Google staff software engineer, said, “Google is concerned about the potential adoption of Microsoft’s Office Open XML format as an ISO standard. Google supports open standards and the Open Document Format an existing ISO standard that has been a driver for innovation. We do not think it is beneficial to introduce an alternative standard when the Open Document Format already meets the common definitions of an open standard, has received ISO approval and is in wide use around the world.”

The pair added, “Multiple incompatible standards are a bad thing for customer choice, as purchasers of Betamax video recorders discovered to their cost,” while “Multiple implementations of a single standard are good for both the industry and for customers.”

Some countries, such as Brazil, China, India and Canada, have already announced that they have voted against Open XML. The final result on whether Open XML will make the grade as an ISO standard will be announced on Sept. 2.

It’s possible that there will not be enough Yes votes and enough No with Comments votes that the ISO won’t be able to give the proposed standard a thumbs up or a thumbs down. If that happens, there will be an ISO Ballot Resolution Meeting on Feb. 19 through 25 in Geneva, Switzerland. There, either a decision will be reached on the Open XML as it stands or revisions will be made to it. In that case, the revised standard would then be voted on.

A version of this story first appeared in DesktopLinux.

August 24, 2007
by sjvn01
0 comments

Microsoft kills off anti-Linux ‘Get the Facts’ site

In Linux circles, Microsoft’s anti-Linux site, Get the Facts, was better known as Get the FUD, and was seen as more of a joke than a convincing argument in favor of Microsoft products over Linux. Microsoft may have come to agree that the site was not serving any useful purpose, as the company closed it down on Aug. 23.

From the beginning of the “Get the Facts” ad campaign in 2004, Microsoft’s “Facts” were often questioned. Reports favoring Microsoft’s TCO (total cost of ownership) from research groups like the Yankee Group were published. When the Yankee Group published a report showing that Linux and Windows were neck and neck in TCO, Microsoft didn’t tell the world about the pro-Linux report.

Even at the start, Microsoft took other reports and published them out of context. For example, the first report Microsoft published was a 2002 vintage IDC report, which was sponsored by Microsoft, comparing TCO of Windows 2000 to Linux. IDC found that W2K beat out Linux in four out of five common enterprise tasks. This may have been true in 2002, but in 2004? I don’t think so. By then there were lots of Linux network-smart administrators.

I could go on and on–for example, about Microsoft trying to hide that it was sponsoring anti-Linux research–but there’s little point.

Now, though, Microsoft is of at least three minds about Linux. One, represented by Steve Ballmer and his patent lawsuit fantasies, still wants to stomp Linux and open source into the ground. Then there’s the side that wants to give Linux and open source lip service while doing as little as possible. Here, I count Microsoft’s open-source projects and its recent efforts to get its own open-source licenses.

Then, there’s the part of Microsoft that gets that Microsoft is going to have to learn to live with Linux. Mind you, it really doesn’t want to, but Linux isn’t going away. In this group, I count the people who came up with the technology interchange with Novell and the people who are still, I’m certain, trying to work out a similar agreement with Red Hat for Red Hat Global Desktop.

The new replacement for the Get the Facts site, the Windows Server “Compare” site, isn’t as bad as the old one. Still, at the end of the day, it’s just a propaganda site masquerading as objective information.

No surprise, really. I mean, Microsoft and FUD go together like stink and… Ahem. Anyway, if you want the real facts about Windows Server 2003 versus Red Hat Enterprise Linux, why not try Linux out yourself? After all, business server Linuxes are never more than a download away.

My recommendations for a business Linux server are RHEL 5, SLES (SUSE Linux Enterprise Server) 10 Service Pack 1 , and the RHEL clones, Oracle’s Unbreakable Linux and CentOS for experienced Linux users.

A version of this story was first published on Linux-Watch.

August 17, 2007
by sjvn01
0 comments

How many Linux desktop users are there?

Desktop operating systems numbers, even when gathered by top research companies, such as IDG and Gartner, are often a bit fuzzy. When it comes to uncommon desktop operating systems, like Linux, the numbers often amount to little more than an educated guess. Now, a new open-source program, statix, promises to give accurate data on how many Linux desktops are actually in use.

Cole Crawford, a Dell Linux IT strategist, created statix to anonymously track the number of Linux desktops worldwide via an opt in Python script. Crawford is a long-time Linux user and has been involved in the Linux community since the launch of Slackware 1.0 in 1993.

Statix uses a Python client and a hosted Python CGI (Common Gateway Interface) back end to track the country in which each desktop is running. Eventually it will also be able to track the kernel and distribution of statix-using Linux desktops. The project is currently seeking a MySQL database administrator and a user interface UI developer.

While Crawford is a Dell employee, this is a purely open-source project. In a brief conversation at LinuxWorld in early August in San Francisco, Crawford said, “This is my attempt to give something back to the Linux community, which has helped me so much over the years.”

This project has Linux Foundation’s support. As John Cherry, the Foundation’s global initiative manager, explained, “We have been playing with market data for years now, and with freely distributable software, it is very difficult to glean this market data.”

Linux desktop surveys such as DesktopLinux’s own annual survey tend to be biased towards Linuxes with large, active communities, which encourage their supporters to vote. Other popularly quoted numbers, such as those from DistroWatch Page Hit Ranking, tend to reflect the interests of community-supported Linuxes and quickly evolving Linux distributions. So, for example, Mint, which has released several different editions in the last eight months, gets far more attention than the comparatively slow changing Red Hat distributions.

To make statix more accurate than these sites, or the results from the interviews of CIOs and CTOs that the research companies use, Cherry wants all the Linux distributions to adopt statix and help with its development. “The trick,” according to Cherry, “will be to get this deployed in all of the open and commercial Linux distributions. If it doesn’t get fully deployed, we will have yet another skewed source of desktop Linux market data.”

So it is that Cherry asks that, “If desktop Linux market data is important to you or your company, please become involved in the statix project and in the deployment of this tool on every distro. I realize that market data can be a double-edged sword, so distro vendors should get those issues on the table early.”

A version of this story first appeared in Desktop Linux.