|
Helix Code: Beyond Project to Product
What the Guys Are Up ToOne of them--Miguel--is already instantly identifiable in the Linux community by his first name alone. The other--Nat--is Nat Friedman, fast achieving that status. They're among the Linux rock stars, and together they've formed Helix Code, a company that hopes to make them rich exploiting free software, specifically Gnome. They recently talked with Linux Planet about what they're doing, what they hope to achieve, the future of Gnome, the prospects for the Gnome office suite, the Gnome Foundation, and the possibility of automatic acquisition of tickets to Dave Matthews Band concerts. Like many in the community, they seem to be in perpetual overdrive. And like many, they have a single, passionate purpose. Whether their enthusiasm as Linux programmers translates to success as entrepreneurs is still of course, unknown, but they're giving it their best shot. Making Money from Free Software "This isn't just about free software," he says. "It's about the Internet and communications. Certain types of businesses revolved around communicating data, for instance selling music, selling software, weather services, things like this. Now that we have this communications medium which is becoming fairly ubiquitous, and which allows you to transfer information from one place to another very quickly and at relatively low cost, things are changing--certain businesses are just changing. So people have to come up with new ways of making money. Certain industries will die, and certain companies will not be able to keep up and will also die as a result. "Once the Internet became pervasive, open source started to happen, and collaborative efforts, too." He cites as an example the original compilation of the Oxford English Dictionary. "It took 70 years to produce the first edition of the OED, and without the postal system it would not have been possible. So now you have this new communications medium, which is the Internet and which is fast. And which also has the benefit that when I send you a piece of data, I keep that piece of data myself. Suddenly you have zero cost transmission of information and zero loss to yourself when you transmit the information. This allows things like open source and collaboration to happen. So, suddenly, how can you charge for information, you ask yourself, especially information which you want to be ubiquitous, like software? Software has to be fairly ubiquitous for it to succeeed in certain cases. If it's a word processor or something like that, you want ubiquity for two reasons: one is file formats, though with open file formats you don't need the same piece of software, and the other is user interface--it should be the same. That's mindshare. They know how to use Microsoft Word. They don't know how to use Word Perfect, necessarily. So suddenly you can't charge for software." Actually, there are a number of companies that have done quite nicely and are doing quite nicely by charging for software. Microsoft comes to mind. But while Internet time is faster than world time, business time requires some anticipation of the future, no? "So this isn't about coming up with a creative new business model for the heck of it. It's about the fact that Microsoft's business model is obsolete. It will be superannuated entirely by the existence of the Internet and the possibility of transmitting information. The proprietary business model for selling software just isn't going to hold up anymore. So, how do you make money?"
The Helix Gnome Online Service"Well, once the software's out there, once people are using Evolution, what we're going to do is sell services which are bundled with the software. You can completely ignore these if you want. If you just want the software, hey, it's out there, it's free, it's a gift to the world. "Helix Gnome is going to get passed around. We have over a quarter-million users. And in there is going to be the ability to subscribe to our for-pay subscription service. It'll be about $3 to $5 per month, and it will provide you with a number of services, built into the desktop, that will just improve the entire experience for you. For us, this is about improving the web." An online service? "The basis for all this is a piece of code called the Helix Gnome Updater, which updates your system and allows you to install third-party software and so on. That's one service, and we're going to provide that for free. But in addition to that, one class of services we're going to be providing is called passive information. You're sitting there, you're using your computer--a lot of people use the Web to get information. They want to know what's going on in the world. They want news. They want to know, is it going to rain tomorrow? Can I play my golf game? They want to know when Dave Matthews is playing in town or whatever--so a lot of that information can be done passively, and it actually makes a lot more sense to embed into the desktop. So for example wouldn't it be cool if in my calendar I could have concert listings appear, right in the calendar. I mean, I could opt in--this is not going to be an advertisement; we're not selling advertising space on the desktop by any means--but in the Evolution calendar wouldn't it be great if suddenly Dave Matthews is playing this Friday at the Boston whatever, and you can go ahead and click here to buy a ticket. So that's one thing. Weather reports in your calendar. Browsing the yellow pages with your contact manager. These are all the kinds of things we're going to be doing, which I call passive information. "Other services include storing all your mail, calendar, and addressbook information offline, being able to access it through the Web. And there are a couple of others that I'd rather not reveal just yet. Those are the kinds of services, and we're going to charge somewhere between $3 and $5 per month. Of course, Gnome will be perfectly usable without them, but we hope that they'll be good enough that people will really like them and feel compelled to use them--feel excited about using them because it makes their desktop all the better." Does this mean that Helix Gnome is going to be crippleware? "We're not going to nag people. There may be a greyed-out menu item somewhere, but that would be the extent of it. We're certainly not going to nag anybody. The desktop is complete without the services. But it's a really great Internet desktop, one step up, once you turn them on. My image of the thing is that when you turn the thing on, it lights up, when you plug it into the services it lights up suddenly, and you've got information coming in, you've got lots of other things available to you that you just didn't have before. The software itself is free, totally free." Miguel: Excited About the Software "I don't think of Gnome as a desktop. I'm more interested in Gnome as a platform. I met Nat Friedman on a trip when I went to interview at Microsoft. I was interviewing with the Internet Explorer team at Microsoft that was doing the port to Unix. The interesing thing is that I went to the office where one of my friends was working. He was just working on the Java part. And then some other guy was just working on the HTML part. But nobody was running the actual browser, together. I was like, 'Guys, how do you expect that to work someday?' They showed me how it worked, and the whole component technology. And that's how the Gnome project actually started. It began as a project to bring the component technology to Unix. We had an idea of making a component system that would work across toolkits. And that idea didn't really quite work because we didn't have all the energy to actually work on that project. "Back then, I was working on Linux on the SGI, so these were more of my side projects, while I was actually concentrating on the X server and the kernel support for the SGI and Linux."
Where Gnome is HeadedIt's difficult to sort out where Gnome ends and Helix Code begins, or vice versa. And while this may or may not be the actual case at Helix Code, Miguel seems excited particularly about Gnome itself. And he's delighted to describe its future. "Gnome 1.4 is slated to release sometime in October or November. Gnome 2.0 is basically going to be many things you see in 1.4, like Evolution, the groupware client which has mail and calendar and so on. And Nautilus. And with the introduction of these two things, we introduce for the first time Bonobo, which is our component technology as a usable platform, and other things like the configuration engine, the virtual file system, the Gnome printing architecture finally reaches 1.0 status. So a number of these technologies are going to be available for Gnome 1.4. I think the big shift between 1.4 and 2.0 is going to be that the API is going to be updated. For instance, we're still binary compatible with Gnome 1.0 that was released, what, a year ago? a year and a half ago? and we have a commitment to keep binary compatability. Gnome 2.0 is the first time we've scheduled a big API cleanup. This will happen at the GTK level. "One of the other features is that GTK 2.0 and hence Gnome 2.0 use a new engine for displaying text on screen and getting characters together and printing those. That is the Pango infrastructure. Pango is a system that lets you output Unicode text to screen and printers and so on, but it's a pretty complex process if you're taking typography into account, and that's where the strength of Pango is. It's interesting, because in some languages you cannot just use a different character for representing a letter. So the big change is obviously Pango, and porting all that stuff over to the GTK 2.0 platform. "The Gnome libraries are also going to become more integrated. Right now, we have many various packages. Even though we are doing this in a way that is binary compatible, many packages were being developed outside of the 1.0 release libraries. With 2.0, we have a chance of integrating everything into a single spot, like the Bonobo components architecture is now going to become part of the Gnome core library, so everytime you use the Gnome core libraries, you automatically have the full Bonobo support in your application. So it's mostly an integration. I wouldn't say it's such a groundbreaking thing. It's more of an ongoing work, into porting stuff we have right now. Most of the stuff I've mentioned is already developed and already finished; it's a matter of porting the applications over to this new platform and doing the unification that we have planned." The Gnome Office Suite "Well, right now we have a couple of applications to make up the Gnome office suite: Gnumeric, a few others, Gimp, and the StarOffice people are releasing their stuff open source next month, and we're going to be working with them to get that stuff ported. What we're doing is, we're taking a hammer to StarOffice. StarOffice is this big thing right now. We're splitting it up into a bunch of tiny pieces with a hammer. These are all called components. They're all going to be Bonobo components, and then we're going to reassemble them into an office suite which is totally Gnome native." But what about outfits like the AbiWord project, which has been the Gnome word processor? "AbiWord, yeah. It's fairly larval. It's up to that project. I'm not driving that. That's up to the AbiWord guys."
Helix/Gnome? Gnome/Helix?It is confusing, now--is Gnome a Helix product, or is Helix Gnome plus services, or what? At the code level, Miguel says, Helix is Gnome but faster. "Everything we do is basically Gnome. Every contribution we do is basically put into the Gnome CVS. The only difference between Helix Code and Gnome is at some point in the Gnome history there is a freeze and we say, 'Here is a package that is 1.2.0.' But as development there is a bug found and fixed and we say it is 1.2.1. Or 1.2.2. So the only thing we do at Helix is basically tracking those versions and publishing the latest version. Helix Code is nothing but a continuous version of Gnome. We never release anything out of CVS. Everything must be released first in tarball format by the Gnome maintainers. There's so much stuff that you get the illusion of being a continuous distribution. Everytime a maintainer decides to do an update, we do the same, but in binary format for all these platforms." And there is, of course, the Helix installer, which makes upgrades easier. "It's a nice piece of work, and the new version is awesome. We've learned from our mistakes in the previous version. It is actually a full rewrite of our previous installer, and the this time we actually had GUI designers helping us, so that also makes a difference." The Gnome Foundation "Different people joined for different reasons. I think one of the biggest commitments is the one from Sun. One of the biggest things they're doing is regression tests on the whole system. They're starting to provide patches and fixes to the system, because they need to have a system that is reliable and that can be run in very controlled environments. Solaris is used in many places, and they need to guarantee that the software is going to work. From the quality assurance point of view, in terms of the solidity of Gnome, it's a very good thing to have someone like Sun contributing that effort. The same thing happens with different companies. All of those companies are working together towards the same goal, some with different priorities than others. In the case of Helix, we try to make installation easy; we try to make Gnome beautiful; in the case of Sun, they're trying to take what we have and make it robust and reliable. Different people contribute in their different areas of expertise. And the same is true with Compaq. Compaq is interested in the handheld market. H-P is also interested in replacing CDE, and so on. "The foundation was something we needed. When companies became interested in the past, it was difficult to get the companies to grow with the community. They talked to one or two people, and we're individuals and sometimes we dropped the ball, and things were slower than they could have been if there had been a single entry point for these people to get in touch with the community. So the foundation addresses the need of organizing the collaboration of the different companies into Gnome." Is Helix Gnome a Fork of Gnome? "No, not at all," says Nat. "There's no forking at all. Certainly not. What we do is the same thing Red Hat does. Red Hat takes Linux, and they take all these underlying tools, and they make a distribution of those. Sometimes that involves polishing them and adding patches, but no one would say that Red Hat forks Linux, right?" So, what, then, is Helix Gnome? "We see ourselves as a distribution. We apply those patches that are necessary to make things work. No one ever ships a Linus Torvalds kernel. No single distribution ever ships a Linus Torvalds kernel. There's always some change they make before it goes out the door. But I consider the things to be so small, and necessary and important. You can't ask maintainers to release new versions and to be the distributors. There's a step beyond project to product, and that's where we see ourselves. Every contribution we make, every bug we fix, gets contributed back into the main Gnome CVS. It's all gnome.org for us." The FSF's Place in All this "Miguel is on the board of the FSF," says Nat. "We also know those guys pretty well. We hung out with them a lot. We believe very strongly in what they do. Gnome is part of GNU. I'm talking about Helix Code--when you're talking to me, you're talking to a representative of Helix Code." Distributions "Red Hat's not shipping Helix Gnome yet; they may," says Nat. "Turbo Linux is shipping Helix Gnome with their next revision; IBM's bundling it on their laptops, but we just popped onto the scene. You can get it from our website, right now. That's the basic story. That's going to change in the next couple of months. We're a productized, branded version of Gnome, basically. What we do is take Gnome and we spit-polish it and we make it usable." The Big Money In an article in the Boston Globe it was said that Helix Code will do $120 million business next year. A pretty tidy sum for an outfit based on free software. Nat is a little more circumspect. "I didn't say that. It was not me who said that."
|