A Conversation with Bernie Thompson, VistaSource
Put simply, VistaSource is Applix's Linux Division spun off. As Bernie Thompson, president of VistaSource makes clear below, the increasing diversity of Applix's business initiatives--namely, Windows-based customer relationship management processing on the one hand and "UNIX-centric, Linux-centric" office suites on the other--made spinning off the Linux division as VistaSource almost a no-brainer. Now with an arsenal that includes Applixware 5.0 and their new Applix Anywhere technology, Bernie's VistaSource is set to take on the StarOffices, the Corel WordPerfect Offices ... even Microsoft Office itself. In and among telling us how he thinks VistaSource will make its mark, Bernie offers some interesting thoughts on licensing, ASPs and mobile technology, and the joys of being "native." Read on.
David Penn: The first thing I've got to say is that this must have been one helluva year for you. From Cosource and then the big triumph moving to Applix and now the spinoff to VistaSource ... could things be going any faster for you?
Bernie Thompson: Yeah, it has been crazy. Things move very quickly in this business environment. I didn't necessarily know at the point where Cosource was acquired by Applix that this would be the end result. But it's what needed to happen, and the situation is a really good thing, all things considered. We're working hard to make it happen in the right way.
David: Is the VistaSource spinoff going to mean another geographical move for you?
Bernie: No, we're going to remain headquartered here in Westboro, Massachusetts, which is about 35 minutes west of Boston. Kind of a nice area, a little bit rural, but close to all the Boston universities.
David: When did the discussion on the spinoff of VistaSource begin, as far as you know, and how involved were you in that?
Bernie: It did actually start early. There had been some discussions I hadn't necessarily been aware of, and those discussions really accelerated in January. Applix has a number of businesses under one roof, and those businesses target very different markets. Everything we do that's related to our open-source stuff and related to our office products - which aren't open source - is all UNIX-centric and Linux-centric. The other parts of Applix are businesses which deal with customer relationship management and on-line transaction processing and analysis of financial data. Those things are primarily Windows-based, and they're just very different markets. So we had one company, which was Applix, with two very different focuses. And it was very confusing for people, especially for those in the Linux community, I think. You'd come to the Applix web site and see all this stuff, and some of it was stuff you cared about and some of it you didn't. It was really confusing for people, so it was a good thing to have a very clear and defined split of what Applix does, and let each part focus and do better through focusing better.
David: One of the things I read yesterday or the day before was that Applix's Linux division proper was a profitable division within Applix. Is that correct?
Bernie: Yeah, that's correct. It's somewhat interesting; the overall financial perspective ... what we're doing at VistaSource, the products and the services, is actually the core of what Applix was originally built on and was very profitable. And Applix did an initial public offering and became a public company in 1995, based on these office products and related software. What started happening, though, was that even though Applix had won the office wars in the UNIX space, UNIX was declining heavily and Applix felt the need to diversify in order to grow. So they diversified into the other businesses I talked about.
Well, in recent years of course, there's been a big change in the UNIX world - and that is Linux. Linux really has begun to eclipse UNIX in units, and is now becoming a major force overall. We all know the numbers: 25% of the overall server market according to IDC, close to 34% of the web server market, and making real inroads on the client side, too, especially in the thin-client and embedded space. So, this kind of business that had been declining for Applix now has this great potential to grow again if we are able to do our jobs well. What happened with VistaSource is that we've got a very substantial business of selling office suites into the UNIX market, a business that is large, that has been declining, and we bring that along with us. We had $18 million in revenue in 1999, and out of that, $1.5 million was Linux-related. So you can see we're still doing a lot of business outside of Linux. To put that in perspective, for example, Caldera, who did an IPO several months ago, had about $2 million in Linux revenue. So we're really a substantially sized company targeted in the Linux space.
David: One of the things I noted is that Applixware also has some presence on other platforms as well.
Bernie: We haven't necessarily done a good enough job at communicating that. One of the real strengths of the product is that it's very much a cross-platform product, and it's really a tribute to our engineering team. We've traditionally been available on a wide variety of UNIX platforms. We've also had a Windows port in the past, and for a major customer, we recently did a port to the IBM system 390. It really just shows you how portable the code base is. Now, with Applix 5.0, we're GTK-based, and GTK and the GTK team (which is, of course, all open source) have done a great job of making GTK itself very portable. So now our underlying code is portable, and with GTK as our user interface, it is even more portable. It's a good situation being able to target wide-platform support. And then, on top of all that, you add a really cool functionality: the ability to access the office suite from any web browser through our Java interface, and you get really good client coverage.
David: One report I read that was talking about how the Applix Linux division was managing to win a lot of new customers, but it wasn't seeing a lot of the old customers buying upgrades and keeping up that tack ... is that true, and is it very much of a problem?
Bernie: It's something we have to struggle with: the decline of traditional UNIX on the client, which is where we were really dominant several years ago. I guess we don't have any evidence for this at this point, but I think UNIX on the client is going to make a resurgence because we've really had a period here of several years where everybody's movement was toward Windows, and a lot of professional organizations - for example, the financial community, Wall Street - they previously, even in the mid-nineties, were largely UNIX-based. Then, in the late '90s, they were all heavily transitioning toward Windows and that hurt UNIX and Applix. Now with Linux, the UNIX market as a whole is having a resurgence. I think you see that, partially, in Sun's financial results - a lot of that is server, but there's also some client in there. So to the extent that UNIX as a whole recovers, and becomes an operating system platform which people are choosing, we have real good markets to sell into.
David: Let me ask you a little about SourceAccess ... which I believe is what the license is being called. I think of it in the way I think of different sorts of licenses: the Mozilla license, the Sun-community license ... all these different variations on the open-source theme. I wonder if you could tell us a little bit about what SourceAccess means, and why that's the way to go with VistaSource.
Bernie: The first thing to know about the SourceAccess license is that it's not an open-source license, in the sense that it wouldn't satisfy the OSI definition. That's because what we're trying to do is take the freedom aspects of open source, the liberty aspects, but not the "free beer" or "free lunch" aspects. Open source is basically the ability to inspect the software, the ability to modify the software and the ability to copy it without paying license fees. So we're the first two, but not the third. And that does separate us from most of those other licenses. Now, most people would say it separates us in a bad way, because they're having to pay for their software now. But that's the model we're going forward with - we're a software company, and we live and die based on whether or not we can produce products that people consider worth paying for.
David: How much debate was there internally in terms of figuring out just what to include and what not to include in the SourceAccess license?
Bernie: I think there's still a debate, because what we've done is we've done it, so far, on a case-by-case basis. If you look up on our site right now, you can't just grab the source code - although that's a goal we have for several months from now. So, there's still some internal work we have to do in order to figure out exactly how we're going to do it. Probably the biggest area of contention is in converters to and from Microsoft Office. We have several dozen, if not a little over a hundred, man-years invested in converters to and from Microsoft Office. So you know we're kinda faced with this choice: "well, we've invested a lot in this, and if we make that source code visible, people will be able to copy that from us, that knowledge we've gained about how to read and write Microsoft's file formats ..." Then we ask ourselves, is that the business we want to be in? Having to differentiate ourselves based on how well we can reverse engineer Microsoft's file format? And the answer is no, that's not the business we want to be in. So what we're focused on is trying to do our best to participate with the industry to get this process going of having common file formats between the office suites. We're kicking that off ourselves in the easiest way we know how, which is to fully publish our own file formats. Right up on the VistaSource web site, you'll see in our product section the full information about our file formats. That's a step we definitely feel is a great step, or a great first step, in this process. And then the next step, which is the obvious answer for the industry, is to have an XML file format, which can be a standard to which every office suite can read and write, and use as their standard file format. And we're beginning discussions with other people in the industry to try and get together a consortium to work on that kind of XML-based standard file format. If we're able to do that, if we're successful at doing that, it means the game no longer is "can we lock customers in with our file formats?" Rather, the game becomes, who's gonna produce the best software? And that's the field of competition we want to compete on ...
David: I just want to follow up on that. We hear a lot about standards from time to time, but how important do you think this file format compatibility is? Not just so other companies can compete with Microsoft, but from the point of end users who are trying to deal with each other and trying to swap files back and forth, and are running into these sorts of problems ...
Bernie: Oh, it's just absolutely critical. It's a huge barrier. I mean, for us, just to speak for a second about our issue specifically, if we have one person who's an Applixware user in a sea of Microsoft Office users, and that Applixware user constructs a document and mails it out to everybody, no one can read it. It's just an overwhelmingly huge problem. File formats are just like communication protocols, such as TCP/IP or HTML/HTTP; there really need to be standards in order for computer systems to work together. Right now, within the industry, there is no standard. If anything, you clearly have a de facto standard with Microsoft Office, and it's unfortunate. It would be better if it were basically an IETF-style open standard, so that everybody could participate equally. And that's where we'd like to see the industry go. I think all of us have a lot of work to get there, but we're in a position of having a complete set of word processor, spreadsheet, presentation, etc. software to help that along significantly.
David: Let me ask you something else with regard to Source Access and these sorts of licenses that aren't the purest open source, but are sort of the offspring. What I'm wondering is, when I think of what VistaSource is doing, and then I think of Mozilla and Java and Zope ... I'm wondering if there's really starting to open up this space that's not fully open source, so to speak, and is certainly not completely proprietary either. But you're finding a lot of these technologies that many people are really excited about, but needing to find something in between those two extremes, for lack of a better term. Do you see that also?
Bernie: Absolutely. Basically, the situation we see is that open source as we see it today is going to continue to grow. It is a substantial and important force within the software market today. And all the same reasons why people are developing pure open-source today, all those reasons will continue to be valid. But what we believe is going to happen is that the rest of the software market, the software market that today is totally closed, will change. It is not going to become completely open source, but it's going to shift in spectrum toward open source, towards licenses like the SourceAccess model we're putting forward today. SourceAccess, by the way - and that may not be the final name of the license or anything - is meant to be descriptive, not a name like the GPL which you'd use to represent one license. The problem is we don't want to call it open-anything, because the "open" term carries heavy connotations and we don't want people to get confused and think we're trying to say it's an open-source license.
David: I wanted to talk a little about the application service provider arena. One of the things I wanted to ask is, at the present time, who is it who is really excited about the notion of application service and where is that excitement coming from?
Bernie: I think the excitement is coming from people who are mobile. It's coming from people who need to be able to access a particular application. Let's take a salesman: he wants to be able to enter his sales data and view the sales data while he's still inside the corporation. But he also may need to access that information when he's at a customer site, and he just wants to be able to plop down at a web browser, one that may not even be his own machine. Or he's at an airport, and he wants to sit down at an Internet terminal that doesn't have any special software installed on it. If he's able to just sit down at a web browser and access full applications with full application functionality - not just an HTML page, but real applications - that's really a breakthrough. And that's what we do with our Anywhere technology. We have a demo available right from our site. If you go up to our site and go to our products section, we have a "try us out" link from which you can try out the Anywhere technology ... and it's very cool.
David: Let's talk a little bit more about the Anywhere technology and the ASPs. I've read that VistaSource has a "lead" in the ASP area, and I was wondering if that was a lead in terms of a strategic advantage, or something else technological?
Bernie: It's primarily technological. What we've done is we've created an office suite where we've separated the underlying functionality from the user interface. A word processor is the ability to create text and the ability to change the attributes of that text, to spell-check that text, all those things. That's all kinds of underlying functionality. And then there's the actual display the user sees, and how the user interacts with it. We've separated those two things, and in the case of our Anywhere technology, the underlying portion (the server-side portion) is running as compiled C code on a server somewhere. And there's just this thin client, which is less than 850K of Java download, running on the client. So it's a pragmatic and intelligent split of the workload between the server and the client. And that is unique. Everyone else who's trying to solve this problem of creating an office suite that you can run from the browser has gone the route of trying to create a 100% Java office suite. The problem with that approach is that Java has an enormous amount of performance problems, both in terms of the big download as well as in terms of how fast it runs. The results of all these projects have been pretty poor. You've had poor-performing applications that just don't work well enough for people's needs. By doing what we've done, and splitting the load between the client and the server, you get very good performance. On a good Internet connection with a modern machine, the difference between the native and the Java version is practically unnoticeable.
David: Given the fact that this separation of power is fairly unique at this point, what is the ASP market like right now, then? Is it laboring under some pretty flawed applications?
Bernie: We have all the Citrix-style access to applications, the screenshot protocol. Almost any application can be accessed Citrix-style; the open-source version of Citrix is VNC (Virtual Network Computing). What this consists of is an application running on the server and drawing all of its graphics to the server. There's a separate program running that. Any time the screen updates, the server sends a snapshot of the screen to the client. Then, any time the client clicks a mouse or clicks a keyboard, it sends that key press back to the server, the server redraws the application image, and then the image is sent back over to the client. The problem with this is that it requires a very fast Internet connection, and it puts a high load on the server. The response times are just not acceptable in many cases.
To give you some specific examples, if you hit the "page down" key very quickly in our Anywhere word processor, it is almost instantaneous because all that data is actually sitting on your client and there's no communication happening with the server. If you do that with a VNC or Citrix-style interface, it's going to have to keep getting new screenshots from the server every time you do the "page down" because there's really nothing running on your client. It's all on the server side, and there are just pictures coming back to you. That style, that technology, is the technology that is very widely available. It's available from Microsoft in the form of Terminal Server. It's available from Citrix, which is kind of the pioneer. It's available from SCO with their Tarantella technology, and there's a good open-source implementation called VNC. Any application can be viewed through that mechanism, so there are a lot of ASPs who've been running applications that way; maybe it's MS Office, or maybe it's StarOffice. Then they see our application, and they see the better performance you get with our application through this split workload, and we really come off as a breakthrough because now the application is usable. And it's usable on a 28.8 or 56K modem, where the other styles simply won't work in a case like that. Or won't work well enough. There's a lot of activity. I think most of the technology solutions are not perfect. In our case, I think we really have something that's a step forward. There really isn't anybody else who is able to do quite what we do the way we do it.
David: You mentioned that when you presented this to some of the ASPs that they would be able to tell this difference fairly readily compared to the others ... How is the work going of selling some of the ASPs out there on what VistaSource is doing?
Bernie: That's where a lot of our sales activity is focused. It is where we see a lot of opportunity to move forward. Because Linux runs on 34% of the web servers on the Internet, and we're very Linux-focused, we have kind of a double advantage there, in terms of what we care about and the platforms we sell on, vs. for example Sun with StarOffice where they're really interested in pushing Solaris solutions. But we've got a lot of sales activity going on there. As far as actual sales, we have a substantial amount of our first-quarter sales from this kind of business, but we're really early on still. We really have been focusing on only these ASP markets in just the 2000s here. Prior to this, we really haven't focused on this to the extent we should have. We're in the early days, but the signs we are getting are very good. We think this is going to be a big business for us, but all the cards aren't in yet.
David: The last thing I wanted to ask was about the relationship between VistaSource and some of the other Linux distributions out there, particularly in terms of Applixware showing up with some of the Red Hats or Calderas or whatever. What kind of relationship do you anticipate there?
Bernie: This is where we're really needed, because in the end, you need great applications to make your system useful, right? And today, the major application providers are ourselves, Sun and Corel, and then KOffice and some of the work around the GNOME office suites, but really it's the three: Sun, Corel and ourselves that are very far along. In the case of Sun and Corel, they aren't independent software providers. They have, in both cases, an operating system they care about, which in Sun's case is Solaris and in Corel's case their own distribution. It's such a shame. It's a recreation of the same Microsoft problem again, where you've got companies that own both the operating system and the applications. We're never going to get into the Linux distribution business. We're never going to do that, and the reason is because we want to be the pure, independent software vendor that's able to port widely and support a wide range of platforms - on both the operating-system side and the hardware side. In Sun's case, they really care about the SPARC architecture and in selling SPARC and Solaris servers, and we want to be able to have this be an application that's portable to a wide range of platforms. That's a key focus for us: to be an independent software company that's able to focus on just doing great apps, and doesn't have any of these conflicts of interest.
David: Is there anything else you wanted to add that you thought might be of interest?
Bernie: Probably the biggest thing is just related to our Office 5.0 product. For the first time, we're going to the retail market. You're going to be able to see our office product all over the place in stores around the country. It's pretty exciting, because you know there weren't really Linux sections in a lot of these stores only 12 months ago, and now there are. And, a wide variety of applications are starting to become available, and that's a real sign of health for the Linux market. We've done a lot of work to make Applixware 5.0 the most native of all the office suites. Our office suite comes from UNIX, so it was already a short port for us over to Linux. Then, we've even taken it down the next step further. In Applixware 5.0, the user interface is totally rewritten to use GTK, so we're able to do some very cool things, like be able to take a GTK theme and have that replicate over the application and have our applications fit in with the desktop. We also have "drag and drop" integration with the desktop. We're doing a lot of things and focusing our efforts on really integrating well with Linux, so we're the most native of the office suites, leveraging our existing strength from UNIX.
Applixware 5.0 is a really exciting product. We may not keep the Applixware 5.0 name; we'll be transitioning that, probably leading into the 5.1 version. But everyone knows our product is Applixware, so we can't switch it just on a dime. Branding will be transitioning as we go to the next major version, but it's really a great native product for Linux. It's something which we have and continue to put a lot of work into.
- Linux Kernel Testing and Debugging
- RSS Feeds
- NSA: Linux Journal is an "extremist forum" and its readers get flagged for extra surveillance
- Tails above the Rest, Part III
- Wanted: Your Embedded Linux Projects
- Tech Tip: Really Simple HTTP Server with Python
- Docker: Lightweight Linux Containers for Consistent Development and Deployment
- What's Your i-Name?
- Tails above the Rest: the Installation
- Tails above the Rest, Part II