Mark Hinkle, LinuxWorld Magazine's editor of desktop technology and CIO of NeTraverse, makers of Win4Lin, talks to Kevin Bedell about the latest developments in the Linux desktop and Windows-to-Linux migration.

LWM: Your Dr. Migration column has been very popular with our readers; what gave you the idea to start writing it?
Mark Hinkle: I guess it was mostly a product of my own frustrations as I moved from being primarily a Windows user to being primarily a Linux user. I found it tough to read main pages and search newsgroups for answers to my questions just because the level of technical expertise was so great there and hard for a naïve user to comprehend. I wanted to offer a resource for people like me who were management professionals, doctors, lawyers, teachers, and other vocations that were not IT specific, so they could understand how to run Linux in a less intimidating way. I was also pretty proud of myself when I grasped how to accomplish things in my Linux desktop and wanted to share that with someone else with my background.

Also, since 94% or so of all PC users are using a Windows operating system, I thought it was the logical place to start. I actually am not anti-Microsoft but in any place where there's one dominant player, the lack of competition can make a company complacent in the way they serve their customers. I think Microsoft has done that. That's why I like Linux - the development community is anything but complacent; they are really doing some innovative things that I'm surprised Microsoft didn't figure out right away. For example, Linux has had SpamAssassin, but I don't see an equivalent commercial offering for Microsoft Exchange from Redmond. I just wanted to make people aware that there were some good alternatives to what they were using today and many of them were at least as cost effective.

LWM: What do you think are some of the emerging killer apps for Linux ?
Hinkle: I think the "killer app" designation goes to things like the World Wide Web, e-mail, or maybe the digital camera. I don't know if there is anything in Linux that qualifies as a killer app with the exception of maybe Apache Web server. Actually I think some sectors are emerging that will benefit from Linux. The applications themselves aren't really a "Linux" solution but they could be contenders for killer app status. The first is the spread of pervasive Internet access. I'm interested in seeing when we can truly roam a broad geographic area with our laptops or PDAs and have robust, high-bandwidth applications delivered effectively. One interesting technology in this area is WiMax, based on the IEEE 802.16 standard. The Linux hook for this is that anytime you need to build infrastructure to support new technologies, it usually requires a huge capital investment. Linux can help make the most of that investment because it offers a reliable, cost-effective platform on which to build the supporting WiMax infrastructure or the infrastructure for any wireless network buildout. I think that Linux will enable companies building this infrastructure to leverage their capital better and to build out faster than if they relied on commercial platforms.

The second area that has the potential to benefit greatly from Linux is home entertainment, especially when you look at the trend of digital multimedia convergence. I'm currently enamored with the MythTV (www.mythtv.org) project. It takes the TiVo idea of digitally recording television shows and expands upon that. Since it's an open source project, there is a lot of collaboration for plug-ins and the like for the system, which includes the ability to display digital pictures, and play music and games. I actually built a system and currently hack on it from time to time.

In my mind TiVo should be a "killer app." I have never heard a single TiVO owner complain about the functions it provides. The problem is that it's a closed product so people who want to expand its functionality are limited. I've seen books about hacking TiVo because people want to add more storage or other features that the sealed black box doesn't support. I think that over the next couple of years the market will be ripe for an open source "digital entertainment hub" that offers all the amenities that your computer and traditional home entertainment center offer. Plus, as the ability to manage Linux desktops and servers in the enterprise is perfected, that will translate into the home. I can see the same mechanisms that we use to update Apache on Linux Web servers being used to deliver content such as music and movies to the Linux-powered home entertainment unit.

The other killer app in general is the virtual machine. This space fascinates me since our processors and memory have gotten to the point where our operating systems rarely tax today's latest and greatest PCs and servers. IBM's z/OS on the mainframe is a great example of virtualization on heavy iron. The idea of multiple virtual PCs running on one set of hardware greatly leverages your network card, storage, and memory. I think VMWare did an excellent job with their product and it's the premier way to consolidate x86 servers. Of course, I think Win4Lin is the best way to run Windows OS on the Linux desktop, even if I am a little biased.

LWM: You deal with a lot of companies that are deploying Linux. Can you give us an idea of one or two companies that you really feel "get it" with regard to how they use Linux?
Hinkle: I think that IBM obviously got the value of Linux early on with their $1 billion investment in Linux and their understanding that, in the long term, it's much easier to leverage one OS than to develop operating systems for their variety of hardware.

Another company that I think is getting Linux is Novell. They are aggressively embracing Linux and it's not merely lip service. With their purchases of SuSE and Ximian they made great strides in acquiring expertise and quality products. Also, by virtue of adding Richard Seibt and Marcus Rex to their senior management team, they have two Linux guys helping to set their company direction along with Chris Stone, their "Linux-savvy" vice chairman. In addition, they have Nat Friedman and Miguel Icaza with their leadership in Gnome and Mono. These guys are wired into the community and do get Linux. I think that Novell has all the pieces to really bring desktop Linux forward and it's just not up to them to execute.

As for an end user, I think Unilever is really a great example of a global non-IT company (they make Ragú spaghetti sauce and Hellman's mayonnaise, along with olive oil). They have been very active in deploying Linux in their enterprise and were the first non-IT vendor to join the OSDL. They're a good role model for other large corporations: they use Linux and contribute to its future through their involvement in the trade groups that help vendors understand what the needs of the enterprise are.

LWM: You've been around Linux for a number of years. What first attracted you to Linux and where has Linux made the greatest strides in the past couple of years?
Hinkle: Before I took my present job I worked for a large national ISP (then MindSpring now EarthLink) and we were growing at an alarming rate. We had plenty of customers to serve and worked on a razor-thin budget. We started using Linux to provide our internal infrastructure, mainly Web servers. They did a great job from a performance and price standpoint. Over time, we used Linux for technical workstations and our Web-hosting representatives started to use Linux. I was blown away by how well it served our needs. Back then, I tried installing Slackware and Red Hat as desktop operating systems, with pretty awful results. I kept having problems getting Xfree86 to work. Now I very rarely have a problem installing the OS to a working level, barring wireless and state of the art hardware, things just work. To me this is the most noticeable improvement.

LWM: What things are exciting to you about Linux today?
Hinkle: I'm excited at the number of people who are looking at Linux as a desktop platform. For example, my neighbor, a Harvard PhD working in drug discovery not IT, was telling me how he tried Linux last year. I get letters and e-mails from people in education, government, and businesses who are all taking the plunge. It's exciting for a number of reasons. For one, it is expanding the options for the desktop computing market. The fact that we are introducing another competitor causes everyone to try to improve the other products, including Microsoft. Not that Linux is a silver bullet for Windows. It's not. But it's unbelievable how far and fast things have come. It's interesting that there's a lot of Linux around us, embedded in our TiVos, game stations, and the point of sale (POS) systems that are checking us out at the local Home Depot.

LWM: What needs to happen for Linux to become a more viable desktop platform?
Hinkle: Two things: better hardware support, and application availability. For example, on my new Dell laptop I don't have 3D support for the Radeon Mobility 9700 or the Atheros MiniPCI wireless card from my distribution, SuSE. However, I can make everything work with some tinkering. It's just not something that the average user will put up with. Then it comes to application availability, which is why I've chosen to work in the field that I work in.

The most common desktop needs are being addressed pretty well today. I think browsers, office suites, and e-mail clients are improving rapidly with some innovative features included, like Ximian Evolutions virtual folders or Open Office's export to PDF feature. However, these applications won't inhibit Linux desktop migration. The things like our financial applications (Quicken for one), Personal Information Managers (Goldmine and ACT!), and legacy applications like CRM or ERP applications, these last-mile sort of applications are inhibiting a lot of adoptions. That's why it's critical to have a "bridging" strategy. In the mid-term, if there isn't a way for those Windows-only applications to run on Linux, it limits the number of people who can take advantage of the other benefits of Linux, like fewer viruses, remote administration, and high-availability (high uptime). I feel what we do with our Win4Lin (www.win4lin.com) products is allowing more people to make the transition to Linux.

LWM: What exactly do the Win4Lin products entail?
Hinkle: Win4Lin products allow the Windows operating system to be executed as a process on Linux but not in a very integrated way. For example, the Windows process looks like any other Linux program, so Linux manages memory and processing. The file system is not a special system; it's using whatever file system Linux is using. You can even edit a document in Windows using Microsoft Word and then hop right over to Linux and open it in Open Office. Finally, it's lightning fast because Windows is executed right on the processor. It's not an emulated x86 processor, so there's very little overhead. That's the 1,000-foot view of what we do. Here are some of the nuances: the Win4Lin Workstation is a single user version that runs locally on the Linux desktop. The Win4Lin Terminal Server runs off the server and is served in a thin-client model, so there's a single point of administration. Plus it makes your Windows session portable so you can pull up that session from your desk, a conference room, or at home.

LWM: Can you give me some examples of how this solution is being used today?
Hinkle: A good example is Gruppo Ventaglio, headquartered in Italy. They are a tour operator with branch offices all over the Mediterranean. They serve their Windows desktops using Win4Lin Terminal Server and Tarantella (www.tarantella.com) over the network to all these small offices (over 500 concurrent sessions I believe). Then the IT staff runs the sessions from a data center since it's impractical to have on-site IT staff in all these locations.

Another example is LiveDoor (www.livedoor.com) in Japan. They are a systems integrator that has converted PCs to thin-client machines using the LinspireLive! Bootable CD, then they use the Win4Lin Terminal Server to serve Windows applications. They're duplicating this configuration in schools in Japan so that they can take advantage of Linux as a learning environment, and they can still teach students about Microsoft Office, since it's currently the most popular office suite.

Finally, probably my favorite example is in the Netherlands schools where Siceroo, another integrator, is putting together ICT (Information Communications Technology) solutions for schools using the Win4Lin Terminal Server, Linux, and Sun's thin-client devices - the Sun Ray. In the schools, they're dropping in servers supplemented with thin-client devices. The students still get to use all their rich multimedia Windows applications running on Win4Lin, but the thin-client/ Linux combo is very easy for Siceroo to maintain on behalf of the school. What's really neat is that the students authenticate to the server through Sun's Sun Ray using smart cards. They can work on their projects at their desk by sticking in a card, then when it's time to show their teacher they can pull out the card, walk up to the teacher's desk, and pull up their desktop in the same state it was in at their seat. That's an incredibly useful tool in a school where many users share the same PC.

LWM: How does Win4Lin stack up against other Windows-on-Linux solutions?
Hinkle: I think there are four ways to run Windows on Linux today: virtualization, emulation, integration, and native Windows redisplay. Virtualization is a VMWare type model, which is a powerful way to run Windows on Linux. They create a total virtual machine (VM) and then install an operating system inside that VM. This works well but it's expensive because of the system resource requirements to actually create the VM. However, if you have very beefy hardware, it works well and the per-seat licensing of this solution is somewhere in the $180 range last I checked. But this is an excellent solution for developers who need to emulate a lot of environments for testing. Also, it's really great as a server consolidation tool in which you want to run a bunch of servers on one piece of hardware. For enterprises moving to Linux but with legacy Windows servers that's where a VM solution shines.

Another way to run Windows on Linux is to use Wine (www.winehq.org). This would be the dream solution since it runs Windows applications without a Windows license. The approach is to provide an API that the applications use in place of Windows. The problem is that since these applications weren't designed to run in this way, application support is severely limited. I've been using Wine on and off for a couple of years as a point of reference and I find it works okay for some applications, but it's a little difficult to install for the uninitiated Linux user. My assessment is that it's like these collegiate engineering competitions where they build bridges out of matchsticks. With a lot of hard work they can make things work and it's a very admirable effort, but it's not really practical for widespread use.

There are two companies that offer commercial support for Wine and both have a somewhat narrow range of applications. TransGaming (www.transgaming.com) focuses on running Windows DirectX games and they offer a subscription of $5 per month. The other is CodeWeavers (www.codeweavers.com), which has three levels of support for applications, indicating the varying degrees of success that applications should run under Windows. Their number of "gold applications" is around 12 or so right now; if you need just a narrow number of applications or supplemental browser plug-ins, it's about $40 for their standard edition. I think their solution works best for hobbyists who need limited Windows application support (for example, Word and Excel 97/2000, Internet Explorer, and a few others). I do think it's a very neat idea but not something I would use in a business.

Their integration, which is where Win4Lin comes in, gives you that true application experience of virtually all Windows applications. It has low-resource requirements and it really integrates tightly with the Linux operating system.

The final way to run Windows applications on Linux is to redisplay Windows applications via a Windows Terminal Server (with or without Citrix) to Linux. This solution does allow you to run Windows applications as intended but really is counterproductive to a Linux strategy. My problem with this is that you still don't get that end-to-end Linux footprint; you still need to admin Windows and Linux servers. I wouldn't even begin to speculate on pricing because the Microsoft licensing is pretty confusing, at least to me.

The bottom line is that all these solutions are bridging strategies. Ten years from now they will be less of an issue, but over the upcoming years, as we see more and more migrating Linux users, they'll need a migration path for their applications.

LWM: Nobody's perfect, so where does Win4Lin fall short?
Hinkle: For the vast majority of users I would say nowhere, but then again I'm biased. Currently we support only Windows 9x/Me as the Windows version, but we are expecting to go to beta with a new version that supports later versions of Windows in the third quarter of 2004. Also, if you want to play Windows 3D DirectX games, Win4Lin is not for you. We do support DirectX 2D but when you directly access the hardware the way DirectX 3D requires, it's tricky to run it on Linux. Also, we don't support audio input, which is on our requested features list. The key for us is to focus on running Windows applications on Linux, not to manage peripherals or do high-end gaming. We feel that we are an excellent solution for moving applications and extending existing Windows investments. I don't know if the majority of people out there get a lot of value from newer releases of software and operating systems. For its intended use I think Win4Lin fits the bill. For high-end processing applications like 3D gaming, we're not what that profile of user is looking for. Also, the one thing that's not really a weakness but an inconvenience is our need to patch the Linux kernel with our hooks that are open source but not used outside of our program. We have worked with Linux vendors to add these hooks, including Xandros, Lycoris, and Lindows, but we would like to have more distros included with them.

LWM: Outside of Win4Lin, what projects or software excite you right now.
Hinkle: I'm in the process of relaunching our Web site (www.win4lin.com) and I have been doing a lot of work on Mambo Server (www.mamboserver.com); I've been blown away at this very robust content management system. You can actually install a fully browser-administered professional Web site in minutes and allow it to be maintained by people who don't know a lick about HTML, PHP, or MySQL. On top of that there's a huge community behind it, churning out all sorts of neat components for functions such as blogs, file repositories, and document management, as well as thousands of templates that can change the whole look and feel of your Web site in seconds. I'm also very interested in the blogging software out there. I just started blogging (http://blog.luckydog.org) - I guess that's a word now - in the last couple of months and have gotten a lot out of it, mainly from reading other people's blogs. I am currently using Bloglines (www.bloglines.com) to read a lot of RSS feeds and some select blogs. It's nice too because it's Web based and if I'm working on another PC I still get my news updates from a consolidated Web interface.

LWM: What are some of the things we should look forward to from you in the future?
Hinkle: I'm very excited about the next release of Win4Lin. I have a couple of product ideas that could take the Linux desktop to your Web browser with and without Windows applications; that could happen yet this year. I also have a couple of book deals in the works, writing about Windows-to-Linux migration, and, of course, one of my favorite activities is writing my monthly column for LinuxWorld Magazine.