Other LWN stuff:
 Daily Updates
 Calendar
 Linux Stocks Page
 Book reviews
 Penguin Gallery

 Archives/search
 Use LWN headlines
 Advertise here
 Contact us

Recent features:
- RMS Interview
- 2001 Timeline
- O'Reilly Open Source Conference
- OLS 2001
- Gaël Duval
- Kernel Summit
- Singapore Linux Conference
- djbdns

Here is the permanent site for this page.

See also: last week's LWN.

Leading items and editorials

The OpenLinux press releases and web pages say very little about licensing for these products, which could lead many to believe that OpenLinux is yet another free distribution. If, however, you dig into the OpenLinux Workstation Reviewer's Guide (warning: 5MB PDF), you'll find the following text:

OpenLinux Workstation licensing has changed from previous versions. OpenLinux Workstation, as a product, is licensed per system and cannot be deployed without limit. Caldera will provide a certificate of license authenticity (COLA) with each unit sold, and Caldera expects each customer to have a COLA for each system that deploys OpenLinux Workstation.

Linux users have grown accustomed to being able to copy distributions at will. It's not uncommon for a company to buy a small number of distribution CDs and to install them on dozens or hundreds of systems. By imposing per-system licensing, Caldera is challenging that practice, and, in the process, raising some interesting issues. We'll look at Caldera's licensing from two angles; the next article goes into why Caldera is taking this path, and the following will examine whether members of the free software community are right to criticize the company.

Why is Caldera doing this? Given that a move of this sort is guaranteed to provoke criticism from the free software community, what could have inspired Caldera to go with per-system licensing? We talked with John Harker, Caldera's vice president of server product management, and asked him that question.

In general terms, Caldera is trying to do two things:

• Expand the Linux market to a new class of users, and
• Actually make some money for a change.

The folks at Caldera have made the observation that Linux has been most successful in situations where it can be made to do something right out of the box. Tasks like running web or file servers, for example. In situations where integration of tools is required, adoption of Linux has been slower.

Caldera's purpose is to make a version of Linux that works "out of the box" in a wider range of situations, and to help others do the same. The company has a particular focus on independent system vendors who integrate the operating system with some other, task-specific software and resell the result. Those vendors want an operating system that "just works." Something that is stable, lacking bleeding-edge software, has a development roadmap, and will be supported by its vendor for a long time. In other words, these people want something that "looks more like a traditional operating system." Caldera will be trying to provide that system for these vendors.

So the company's target is not, in general, the current set of Linux users. Caldera is thus not afraid of losing customers as a result of the new licensing. As Mr. Harker put it, Caldera may lose a number of "users," but the company does not expect its paying customers to go elsewhere. And, in fact, it expects to attract quite a few more of them through the new features in this distribution.

Thus, according to Caldera, the company remains committed to Linux and free software, and it wants to continue to see them succeed. It's just taking a different path to make that happen. Meanwhile Caldera continues to support free software projects (including XFree86 and KDE). They are also, according to Mr. Harker, looking at whether there are any interesting parts of SCO Unix that could maybe be released under an open source license, and would appreciate input from the community on that topic. Interestingly, Linux has progressed far enough that, it would seem, traditional Unix systems have little to offer in the way of additions.

Caldera has chosen a path that is unproven, to say the least. But then, that could be said of almost every Linux company. But it is a path that remains based in Linux, and is not the complete departure that some people have made it out to be. Whether it succeeds or not depends on whether the company can find customers who agree with its vision of what is needed.

Should Caldera be taking this path? Needless to say, per-system licensing has drawn some criticism. Some pretty severe things have been said about this business approach and Caldera's position in the free software community. Much of the energy that has gone into that criticism might have been better expended elsewhere, however.

There is no evidence that Caldera is violating the license of any free software package that it is including in its distribution. Source for free software packaged by Caldera is available - it's included in the box. Caldera is not attempting to restrict the redistribution of the free software it uses; instead, it is restricting a compilation which includes a fair number of proprietary packages. These packages, including Volution, JBuilder, and Forte, do not allow unlimited copying. So a distribution which includes them can not allow copying either.

Caldera's reliance on proprietary software is neither new nor surprising. A mix of proprietary and free packages has always been part of the company's plans. This plan has always implied restrictions on copying; it is, perhaps, a bit late to criticise Caldera at this point.

There are, in fact, a couple of good aspects to Caldera's techniques. Here's why free software advocates should be glad that Caldera is taking this approach:

• Caldera is showing that a business model that mixes free and proprietary software can be executed in a way that does not violate the GPL or any other free software license. The company is demonstrating a way of exercising the freedoms that come with free software. And, despite the fears being spread by a certain large software company, Caldera is not losing rights to its intellectual property in the process.

• Caldera is also demonstrating the great strength Linux has in the form of its many distributions. No doubt many Caldera users will conclude that they do not like the new licensing. Every one of those users has the option of switching to a new vendor without abandoning Linux.

It would be difficult to overemphasize that second point. History is full of captive IBM, DEC, CDC, etc. shops which found themselves in a difficult position when their vendors failed them. No Linux shop need ever fear that fate. Transitioning to a new distribution is not without its challenges, but those challenges are miniscule even compared to switching between proprietary Unix variants. Any Caldera user who is unhappy with the new licensing will find several other, well-engineered, well-supported distributions waiting.

So Caldera is causing little, if any, harm with this approach. In the best (for Caldera) scenario, it will have found a way to add value to Linux that suits its customers, and the company will flourish. Otherwise, its users will make their feelings clear by voting with their feet for one of the many alternatives. Either way, we're seeing a demonstration of the freedoms that come with free software.

The end of the Alpha. Compaq this week announced a new deal with Intel. The end result will likely be the end of the Alpha processor architecture, and a reduction of choices for Linux users.

The details: Compaq will be moving away from the Alpha processor, which it picked up as part of its acquisition of Digital Equipment Corporation a few years ago. In the future, Compaq expects to use the Itanium processor for all of its 64-bit servers. Much of the Alpha technology, and staff that goes along with it, will be transfered to Intel. This isn't happening right away, of course; Compaq is trying to give advance warning to its Alpha customers. Thus, for example, the goal of having all Compaq 64-bit servers using Itanium isn't to be met until 2004. In the mean time, new Alpha-based servers and a new generation of Alpha processors will be forthcoming.

Of course, Compaq isn't the only company working with the Alpha processor; most of them are made by Samsung, and an important partner is API Networks. We talked with a couple of managers at API (Guy Ludden, senior marketing manager and Tom Morris, director of product strategy) about API's plans for the Alpha. API sees no real impact in the short term (the next couple of years), but they acknowledge that there is little chance of the Alpha living on after Compaq loses interest.

The more cynical among us have been heard to mutter that the Alpha will live on in the form of new Itanium processors that actually work. But we can't speak to that. In truth, the Alpha appears to be dead.

The Alpha processor is important to Linux historically, of course. Way back in the beginning, Linus didn't think that Linux would ever run on anything other than the 386:

I'd say that porting is impossible. It's mostly in C, but most people wouldn't call what I write C. It uses every conceivable feature of the 386 I could find, as it was also a project to teach me about the 386.

Now, of course, Linux is one of the most portable (and ported) operating systems available. That change came about when Jon 'maddog' Hall, then working at DEC, got an Alpha-based system delivered to Linus in 1994.

Since then, Linux has been ported to numerous systems, but the Alpha has retained a strong following. It remains the processor of choice for people building high-end clusters, and for anybody with serious processing requirements in general. Many users will be sad to see it go.

What is truly unfortunate, though, is the reduction in choices for all Linux users. The end of the Alpha takes us a little closer to a world completely dominated by a single processor architecture. And that, of course, can not possibly be good for anybody.

Next week's LWN.net weekly edition will come out on Wednesday, July 4, so that we can enjoy the U.S. Independence Day holiday. We'll return to our normal schedule the following week.

Inside this week's Linux Weekly News:

• Security: Samba remote root exploit, new vulnerabilities in pmpost, ePerl, w3m, cfingerd, ntping and more.
• Kernel: 2.5 is coming; kbuild transition; a couple of approaches to memory management improvements.
• Distributions: OpenLinux Workstation 3.1, LNX-BBC, many Red Hat updates.
• On the Desktop: Office options, chess interfaces and new games
• Development: Linux Clusters, embedded contest, net monitoring tools, Gnome Basic (GB), debugging Perl, Python 2.0.1, Kawa.
• Commerce: VA Linux Systems bails out of the hardware business; IPOs in the news.
• History: Software wars map; The Magic Caldron, Hard Hat and KRUD; MySQL goes GPL.
• Letters: 'On the desktop' critics reconsider; shared library dependencies.

This Week's LWN was brought to you by:

• Jonathan Corbet, Executive Editor
• Elizabeth O. Coolbaugh, Managing Editor
• Michael J. Hammel, Senior Editor

Next: Security