LWM's senior contributing analyst, Bill Claybrook, spoke with Jim Zemlin, executive director, Free Standards Group, about the Linux Standard Base, the support it has received in the community, and the importance of having an operating platform that is not only open source but open standard as well.

LWM: Tell us what the Free Standards Group (FSG) is and what it does.
Jim Zemlin: FSG creates a binary standard for the Linux operating system so Linux applications and platforms work together seamlessly. It's a nonprofit organization located in San Francisco. One of the goals of our organization is to make Linux the overwhelming choice of application developers and users by delivering a standard that will offer users the most cost-effective solutions. We believe that having an open standard allows Linux to have a fair shot in the marketplace against Microsoft.

What is Linux Standard Base (LSB) and what is important about it?
LSB is a standard that allows application developers to build LSB-compliant applications that will run on any LSB-compliant Linux distribution. Our mission is to provide an open standard for people who own applications on Linux. We talk to a lot of people who run applications, and they want to run them on the most cost-effective platform available. We believe that Linux with an open standard is going to provide that now and in the future.

What is the business value of LSB to Linux distributors and ISVs?
For ISVs, it makes the job of building applications and targeting the Linux platform much easier. It lowers their costs and reduces their risks in the Linux marketplace. They know that the open standard will remain stable regardless of how the marketplace for Linux distributors evolves around the globe. For Linux distribution vendors, the standard allows them to add value to Linux by enabling them to guarantee interoperability and a broad range of application choice to their customers. You can historically look at the number of applications on Unix versus Windows and see that fragmentation doesn't lend itself to more applications. Another thing that it does for distribution vendors is provide them with a forum for listening to application developer feedback: What kind of standards do they want to see in Linux? What things do they expect to see in a Linux platform? For end users, it provides them with lower maintenance costs and more available applications. This reduces their risks by providing for various degrees of portability for the things that they care about most - data and applications.

Red Hat has the largest Linux market share of any distributor, and it appears that it would like RHEL to be the de facto standard for Linux. Is Red Hat supporting LSB 2.0 and, if so, what are they doing?
They are supporting LSB 2.0. They support LSB in general and have recently expressed interest in running for our board of directors.

When LSB 2.0 was released in late August 2004, the announcement generated about 200 articles. Have you been able to take advantage of this publicity?
In multiple ways. The number of application vendors interested in working with the LSB Workgroup has gone up dramatically, and we'll be announcing new members. We have signed up a dozen new members. I expect our membership to double or triple in the next six months. The Linux Core Consortium (LCC) is one example of an organization that saw the momentum LSB was gaining, and it's now seeking to bring together distribution vendors to comply. The LCC is a group of Linux companies (Conectiva, Mandrakesoft, Progeny, and Turbolinux) collaborating to form a reference implementation based on the LSB 2.0 standard plus a common kernel.

Of the new members interested in the LSB, how are they interested? Are they interested in funding, in participating in workgroups?
All of the above. They are interested in providing technical input into the standard and in financially supporting FSG. They are really interested in making LSB work. The Chinese government is interested in working with the LSB as far as the Chinese National Standard is concerned. We have the attention of many organizations around the globe and have been able to leverage that. The number of participants in the LSB workgroup has tripled.

The group that has the most to gain from LSB is the ISVs, but as far as I know there are no ISVs who are LSB-compliant. Is this still true?
Yes. There are two types of applications out there. There is a large group of applications that depends on a lot of components and middleware. The LSB needs additions to its current state to allow them to be compliant. The feedback that I get from these ISVs is that the guidance that the LSB gives them in developing applications, although they may not have every library that they require in the LSB today, reduces the cost of developing applications for Linux. These ISVs are looking for ways to reduce risks in their Linux investment. They are looking at long-term viable standards. They are very enthusiastic in my conversations with them. I think that you'll see many of them becoming compliant over time.

There is another group of applications, more simple than the other group, that just need to go through the certification process. Sometimes the smaller ISVs are unaware that LSB exists or don't have the tools or resources to become compliant today. Since we made the LSB 2.0 announcement, we've been getting increasing interest from small ISVs that normally wouldn't target Linux platforms. They are asking what the most cost-effective way is to get onto the Linux platform. There are thousands of these small applications.

There seems to be a lot of interest in Java becoming part of LSB. Are there plans to incorporate Java?
We plan on including Java or referencing Java in the next specification releases. The technical details of how we do that and licensing issues are currently being discussed, but it is something that we are looking at. ISVs are interested in having Java as part of the specification.

How will you get some of the relatively well-known ISVs to make their applications LSB-compliant?
Go and meet with their organizations, show them the benefits, sell them on why LSB is important. Tell them what we are going to do in the future and about the things we have already done. The book, Building Applications with the Linux Standard Base (www.amazon.com/exec/obidos/ASIN/0131456954), helps developers develop ISV applications. We're hosting tutorials at the OSDL conference in February 2005 and developing tools that will make it easier for developers to build LSB-compliant applications.

As it turns out, there are some applications that are not easy to make LSB-compliant. Are you focusing on ISV applications that are relatively easy to make compliant?
Yes. The LSB is a different kind of medicine depending on the organization. With certain applications, although we don't have all the libraries in the LSB that they may require, the LSB provides them with a framework that allows them to reduce development and/or testing costs. This is highly valuable to them. For the second tier of applications that can certify to the LSB and run on any LSB-compliant system, we can certainly make it as easy as possible to realize the standard's benefits.

How is funding coming along?
We can always use more money. There's a lot of work to be done in building a standard like LSB. Right now, the question to the industry and community is: Is this important? So far, everyone in the industry and community says that it is. I'm in the process right now of making sure we have an infrastructure that is fiscally sound and diversified so we can maintain our neutrality. We are receptive to anyone who wants to help us financially. There are costs around developing test suites and managing the certification process. The goal of FSG is to have a small staff of people who work the resources of our member organizations in order to make this standard effective for Linux. So far the response I've gotten from people to underwrite FSG has been very positive.

You talk to OSDL. What is your relationship with them?
OSDL provides us with marketing research and market feedback. One of their workgroups has a technical representative who provides input into the LSB.

Does OSDL fund FSG in any way?
Presently, not at all. However, OSDL could be one of the groups of diversified funders that we have and seek.

What do you think the future of the LSB is, beyond LSB 2.0?
If you look at the world of Linux standards efforts, there have been several previous attempts, but the LSB remains the enduring vehicle to make this happen. It may not have happened as quickly as everyone in the industry would like. I think that in 2005 we will be reaching a turning point where having a standard is something that end users and ISVs are going to demand, and for distribution vendors it's in their best interests. The reason that open source is so successful is that there is an implicit promise of open standards. If Linux does not have open standards, it will end up just like another Unix. We have a historical opportunity here to provide a platform that allows end users to get what they care most about: their data and applications running on the most cost-effective, open platform available. And they won't be locked into any single vendor like Microsoft.

Where does Microsoft play in terms of an incentive to get more people interested in supporting LSB?
Two things. First, Microsoft is attempting to derail the Linux industry. That's why they bring up fragmentation. The LSB is a clear and actionable response to Microsoft's FUD and marketing message. Second, and I think the more important issue, is this historical turning point - we finally have an operating platform that you can not only see all the code for, but you also are truly not locked into any vendor. It's not just open source that delivers this. It's open source plus open standards that allow for interoperability no matter what Linux distribution you might be running, no matter what architecture you might be on. It will provide end users with a way to ensure that their investment in applications and data will run on the most secure and cost-performant platform out there. That will only happen through open source and open standards. It will make any technology more efficient. Microsoft isn't for that because it wants to define what efficient is with Windows being that definition.

On your recent two-week trip to China and Japan, were the folks whom you spoke with familiar with LSB? Do they have staff tracking it?
They are extremely knowledgeable. I met with government officials, major industry leaders, and open source community developers. They are aware of the LSB and FSG, and they want to work with us. They want the LSB to be the basis for their Linux and open source efforts in Asia. The purpose of my trip was to build a framework for facilitating increased participation in the standard from folks in Asia and to make it easier for them to participate in providing input and development resources. To this end, we will have announcements about new members from Asia, opening of new offices, etc., in the near future.

The FSG did a survey recently, interviewing several end users, Linux distributors, and ISVs. What are you going to do with the information that you collected?
Put it into action by developing engineering and process plans to build out our standard so that it's relevant and actionable to fulfill the goal of Linux standardization.

Any big announcements at LinuxWorld in Boston in February 2005?
We will have some announcements at LinuxWorld, but over the next couple of weeks we will announce the opening of a lab in China cooperatively with the CESI, the government standard organization in China. And we will be announcing several new members from Asia.

When will LSB 3.0 be announced?
It will be announced this year. There are details on the workgroup site - www.linuxbase.org.