This interview reminds me a little of those sci-fi stories that start with the
big bang, and end 40 billion years in the future. Con Zymaris has the
depth of experience to pull this off in the software space, offering some very
astute insights and provacative food for though along the way.
- The early days of computing as the foundations for open source
- Open source growing up alongside proprietary technology
- Circumstances that led to the dominance of Linux among open OSs
- The economic motivation for open source as the result of technological progress
- On the rise of commercial entities as open source guardians
- The effects of OEM momentum behind particular open source technologies
- Low-cost, lightweight computing as an upcoming nexus of power
- Keeping customers for the long haul without being able to lock them in
Scott Swigart: To start, could you please tell us a bit about yourself and your background with open source?
Con Zymaris: I wear a number of different hats, probably the most prominent of which is my position as CEO of one of the world’s first open source companies, a company called Cybersource that we started in 1991. There are other companies called Cybersource, in the United States and elsewhere, but we were the first.
We focus predominantly on value-added services around software and technology. We started the company because we saw the Internet as a fantastic tool for business, and the software that made the Internet work in 1991 was open source software.
Among the other hats that I wear, I’m also the convener of a government-funded industry cluster, which is a group of open source companies here in Victoria, Australia, called Open Source Victoria. We’ve been running that for about seven years now as an advocacy and industry development forum for open source businesses.
My background is predominantly in software development. I started programming on Apple IIs back in the late ’70s. I moved from there to programming on CP/M systems and to UNIX and DOS in the mid 80s, and eventually on Windows and whatnot.
About 10 years ago, my focus shifted predominantly to business development and evangelism of open source software involved. I still dabble in software development, particularly to help my kids along when they’re doing software projects, rather than being directly involved in major coding efforts.
My involvement in open source projects is predominantly in terms of testing, documentation, business development, evangelism, and so forth.
Scott: So, did you skip over the Commodore 64 or did you just leave it out of the list?
Con: At the time in Australia, we had a number of different manufacturers of microcomputers. One of these was actually a very nice machine called “MicroBee.” It had a lot of features and was a very good value for the money. I spent quite a few years programming software on that.
There was a schism of sorts among the high school kids coding either on something like the Commodore 64, VIC 20, or these other eight-bit systems like the MicroBee. So I spent an amount of time in eight-bit assembler and whatnot coding up mostly game software.
Game software got a lot of us into the industry in the early ’80s. I spent the first three or four years coding up what would be classified as very poor games nowadays. But back then they were quite a bit of fun to make. Probably even more fun than to play, and a lot of us were more interested in coding than playing.
Scott: The VIC 20 and Commodore 64 were my entry point into software development. The various bulletin board systems that ran on them were in some ways the precursor to the forums and the social networks we have on the Internet today.
Con: Indeed, yes. There were a lot of people who joined the industry with the VIC 20 and even the TRS-80. It made computing not only affordable but also fun. Actually, that’s one of the things that caught my attention about open source software almost two decades ago. Open source software reinvigorated a sense of fun. The industry had become a bit too buttoned down, mundane, and mainstream.
There was an era just 10 or 15 years before then when it was just a ball of fun. Open source reintroduced that aspect to me, which is perhaps the main reason I joined the community.
Scott: I think the first time I really bumped into open source, I was a programmer for a Fortune 500 company. We were buying UNIX systems from Motorola, and we didn’t want to spend the very large amount of money for their development environment.
We started using this funny thing called GCC which seemed to work just fine, and it seemed to be supported by these strange people who gave us the compiler for free. But we were very happy to use it as part of our development.
Con: Having access to these excellent pieces of software that you otherwise wouldn’t use and couldn’t afford certainly does facilitate the extension of knowledge and the spreading of ideas. The world is somewhat different now, and people expect free. If you go back 20 years or so, to have quality tools at your disposal, which didn’t cost tens of thousands of dollars, was a godsend.
We had access to increasingly useful open source software even back in the late ’80s. That was the beginning of the realization in the industry that good software doesn’t necessarily have to come from a name-brand vendor. You’ve got a job to do, and at the end of the day, this is a functional tool for you to use for your job.
That’s one of the messages that we’ve been trying to get across to new players and new users of open source software. Putting aside all the religious stuff and the politics this is a tool. It either is viable for you or it’s not. But we see our job (as open source evangelists) as at least making such tools visible to the broader community.
Scott: How do you see the state of open source right now along that whole trajectory of twenty years or so? Back then, most technical people didn’t even know what open source was, and they were used to proprietary combined hardware/software platforms. Maybe they used the GCC compiler without really understanding what was behind it.
Where are we today, in terms of where open source has evolved to and maybe the next steps in terms of where it’s going?
Con: If we go back 30 years or so, pretty much every vendor built, not necessarily their own chips, but certainly their own computer hardware, which was different than other people’s computer hardware, and possibly their own operating system, although more often than not, they were using something like CP/M. There was a degree of “not invented here syndrome” that was quite dramatic.
That has changed a great deal, and Microsoft actually played a large and positive role there. I spent the better part of 10 years building applications for Microsoft platforms back when hardly anybody used Windows, back in the ’80s. I was building applications for Windows 2.0, whereas most people didn’t really start using Windows until Windows 3.0 or 3.1.
Microsoft in particular has helped to make the OEM experience much more mainstream than a company like Digital Research did with their CP/M product. Microsoft began its own operating system business by buying a CP/M clone back in 1981.
That was officially converted in name and eventually in feature-set to what became DOS 2 and then DOS 3 and DOS 4. Another thing that other people probably don’t know that Windows NT initially began life initially as OS/2 Version 3. And as part of the divorce proceedings with IBM, Microsoft got to keep OS/2 Version 3, which in 1989 and 1990, when it was being developed to be a multi CPU-platform version of OS/2.
So you have a great deal of effort by Microsoft to homogenize, if you like, an operating system that all these different PC manufacturers could share. And that was a great success that gave wings to the whole concept of mainstreaming the operating system platform.
So this is something that changed the industry quite radically. Microsoft has obviously benefited quite substantially because of its success with that homogenization. The general public has as well, I think, from the rise of compatible and standardized software. For many years, it was a sleeper that was slowly building momentum, and that standardization that occurred under Microsoft’s control made PCs mainstream.
Scott: And gradually, it became more important, until that model became the most prevalent one in the industry. In some ways, it’s analogous to the rise of the Internet, which existed but only quietly for a long time before becoming incredibly prevalent and important.
Con: That’s a good analogy. The Internet began 40 years ago, but it didn’t really enter public consciousness beyond the computer industry players and the technical cognoscenti until about the mid ’90s. Until maybe ’93 or ’94, most people had never heard of the Internet, and by then, it had been around for 25 years.
That’s an odd thing, and most people don’t really understand that it had been around a long time building up momentum among technical people before it suddenly reached into all corners of everyone’s lives. We don’t make a telephone call nowadays without going through some kind of Internet infrastructure in many countries around the world.
When I was finishing up my degree, I was also working for quite a substantial software distribution company, a very technical company. At the time, I was sharing a house with a system administrator who worked at the University that had just connected Australia to a Internet using a full IP protocol link. This was in the late ’80s.
I experienced Internet access at home because of my house mate’s access to the servers at the University, all of which were running some variation of a proprietary UNIX. That was a fairly early point in terms of adoption of the Internet.
That was spectacularly eye opening for me. I thought this technology and its implications were fantastic, and I believed that we needed to take it to as many people as possible. So, I took this notion to the company that I was working for, and they wouldn’t entertain it at all. They had no concept that anything coming from a group of bearded, sandal-wearing “ex-hippies” at a University could ever be viable in business.
So they said no and I said, “Thanks very much. I’ll quit and start my own business working on this technology.” Which was what I did, going on just about 19 years ago. At the time, it didn’t seem that the Internet would appear viable to the mainstream business world, the government sector, and so on.
Obviously, that changed, and the ethos and the culture of UNIX, which really came from the ’60s and ’70s in particular also rose in prominence, riding on the Internet’s coat-tails. In fact, even while AT&T was very heavily and visibly commercializing UNIX in the 80s, a great strand of UNIX development and culture was happening in the universities by that same group of hacky-sack playing, sandal-wearing bearded UNIX geeks.
These same people, literally, were the ones who built that layer of software that sits atop an operating system and makes it an Internet citizen: TCP stacks, Usenet transfer-servers and news readers, SMTP servers and mail clients etc.
Scott: What you’re describing is really a sort of rising counter-culture within the technology sector, while at the same time, a lot of it was actually government funded, especially by the U.S. Department of Defense and their DARPA project.
So right from the start, there was this sort of dynamic tension between the upstart nature of open source and the most established organizations, with both enabling each other. Even now, the same is true with some of the world’s largest tech companies contributing to open source and also reaping the benefits.
Con: That’s true, but notwithstanding the mutual enablement you mentioned, the spirit of information sharing going on at the time was completely contra to what was traditional in the proprietary world–certainly in the proprietary vendor UNIX world.
At the same time, there was a long period when a lot of the technologies that made the Internet possible had to be run on top of the operating systems that came from proprietary vendors: things like IRIX from Silicon Graphics (now SGI), SunOS back in the day, or Solaris as it’s called now from Sun, AT&T UNIX of course, IBM’s AIX, and so on.
To make these platforms good Internet citizens, you had to use this stack of software that allowed you to do sharing of files through FTP or Telnetting in from one system to another. Very powerful tools and techniques.
With time, the operating system portion of these Internet machines became less and less important. With the introduction of increasingly powerful commodity hardware, particularly the introduction of the 386 from Intel, it became possible to run 32-bit UNIX-like platforms on that commodity hardware. You didn’t have to spend tens of thousands to buy a proprietary server from companies like Sun, SGI, Motorola, or Hewlett Packard.
At the time, it became commonplace for $3,000 or $5,000 to buy something that you could run all your Internet software on. That all happened in the very early ’90s with the simultaneous growth of Linux and also the BSD platforms, when they eventually were ported to x86, particularly by a number of developers on the West Coast of the US.
Unfortunately the BSD world was hit with a very substantial lawsuit by AT&T against the University of California. It stymied the BSD system for several years, which was enough time for the Linux world to get ahead.
The introduction of commensurately powerful commodity hardware, running a viable and freely redistributable operating system was the shortest path to distance yourself from any particular UNIX vendor. You no longer had to acquire technology from a specific vendor to achieve Internet enablement.
That provided enough of a platform for these techie communities, which we called “hackers,” to thrive. “Hackers,” in the positive sense, as opposed to “crackers” who crack into systems and are malevolent. Hackers build systems, whereas crackers break into systems and wreak havoc.
The hacker developers in the very early ’90s helped propel this core of software that became known as open source, that made the Internet work. At the same time, they used the Internet as the mechanism by which they shared development responsibilities and ideas, in chat forums and so forth. IRC is a fantastic tool for development.
They also made this software available by distribution through the Internet. There is almost a genetic DNA double-helix strand between open source software on one side and the Internet on the other. The two cannot exist separately. You can’t functionally and competently build open source software in a meaningful way without something like the Internet, and the Internet could never have existed without something like open source software.
Scott: I’ve never really heard it framed that way, and it makes perfect sense. Up to this day, there are the IRC rooms on FreeNode, where depending on how big the project is, there might be 20 or 100 or 200 people who are just kind of connected to the room all day long as the software is being developed.
We talked to Justin Erenkrantz at the Apache Foundation, which of course is a core piece of technology for the Internet to even exist. Part of their ethos is if it didn’t happen in the mailing list, it didn’t happen.
The idea is that none of the development is happening in closed-off rooms. Maybe at IBM, they’re writing a piece of code that they want to link to Apache, but ultimately it all kind of gets cleared in the public eye.
It’s interesting to consider how things might have gone if BSD hadn’t been stalled. It might easily have become the dominant free and open source UNIX alternative. It’s got a very different license than Linux does, and you have to wonder how that would have affected the adoption of GPL and other things.
Con: It’s an open question, but I think the people who have looked at it fairly seriously would say that from a number of different perspectives and for a number of different reasons, Linux would probably still have come out ahead of the pack. One reason is the license.
It’s odd to think of it that way, but the license of Linux allows a large company to build software and to inject it into the Linux kernel, under the GPL, without fear of that software being taken up and used by competitors.
A prime example is that SGI contributed a developer-decade’s worth of code in the XFS journaled file-system to Linux under the GPL. At the time, the several of SGI’s competitors didn’t really have a serious, competent journaled file system.
If the target platform for the donation had been BSD UNIX derivative (say FreeBSD, or NextBSD), instead of Linux, it wouldn’t have been possible for SGI to have released the code under a BSD license, because competitors can pop up and say, “Ah, BSD license. Thanks very much for a fantastic journaled file system which we can embed within our own operating system. That’s just saved us a load of development. Now it makes us more competitive, and we’ve just leveraged your R&D. Thank you very much.”
So in many ways, the Linux license actually does assist in things like corporate code contributions. Because the potential competitors know that you can’t incorporate GPL code within software that you want to remain proprietary, it acts as a deterrent from purloining that code which various corporate or vendor bodies donate under the GPL.
That means that Vendor A can trust that Vendor B can’t jump in and grab the fantastic bits of contributed GPL code and inject it into their proprietary competing code and sell it in competition against Vendor A. That’s actually a great little warranty, if you like, which allows these disparate vendors to contribute that code into the common pool available for Linux.
There are hundreds of such vendors contributing code in different ways and in different extents, into the kernel, beyond the major players like IBM, Sun, Oracle, Red Hat, and so forth. There are a bunch of other companies that contribute other code as well.
As to why Linux took the lead; there are cultural differences between the Linux and BSD communities, too. Traditionally, the BSD world is much more buttoned-down and serious in terms of the acceptance of code contributions, although I think the Linux world has tightened up how some of its code is contributed nowadays.
Scott: And even though Linux has become dominant, it’s good that we have different platforms, including things like OpenSolaris. Variety is good, and I think there are risks in monoculture, even in something like Linux. It’s best not to put all our eggs in one basket.
Con: That’s true, although for decades, the industry has been looking towards having some kind of common platform. It did this with networks. If you go back 20 years, we had SNA from IBM, IPX from Novell, NetBIOS from Microsoft, and a dozen other competing network protocols. And along comes this open, unencumbered protocol which has also got, interestingly and importantly, open source reference implementations, that can be readily embedded into your own product range technology namely the Internet TCP/IP, and the UDP protocol suite.
Along these come, and essentially sweep everything else out of the way within 10 years. There is the corresponding potential for something like Linux, in connection with cousin alternatives like NextBSD, FreeBSD, and Open Solaris, particularly if they share enough reasonably-common, public, standard APIs. The desktops environments these are running, are pretty much standard across the board: things like Gnome and KDE are standard across these different operating systems, making them appear much the same to the casual user.
From the vendors viewpoint, there was therefore the potential for Linux to achieve something in the operating system space, similar to what the Internet achieved in the network protocol space; the Internet became the common standard that “everybody can trust, nobody needs to distrust”; the place that you could have a meeting of minds with your followers and your competitors alike. Linux could become the Switzerland of operating system platforms.
The Internet did that for networks. IBM could not really adopt DECnet, and DEC couldn’t adopt SNA. Well, as none of these industry players developed Linux, it becomes the neutral thing that they can all adopt to build products and solutions with, as long as they stick to the spirit of the community and the dictates of the license.
And you see a lot of that already in those areas where there isn’t an encumbrance (or expectacton) to adopting alternate operating system technology. Anything from mobile phones, to GPS systems, to DVD players, to Smart TV systems, to in-car computing systems. In all of these areas, to put it bluntly, people don’t necessarily expect to sit in front of a Windows interface; that expectation therefore doesn’t deter vendors from selecting Linux.
Users don’t expect to see a system menu with a Start button and the blue sky and green fields of an XP desktop in their GPS system, or on a phone or DVD player. All those product segments are very quickly becoming the Linux space, because there’s a relatively minimal amount of negative reaction from users when presented with that alternate user experience.
If they see something that performs how they want it to perform, that’s what they then use. Because of a number of different licensing and development costs and all sorts of stability reasons, many manufacturers are adopting Linux technology for devices. It’s obviously much harder for Linux to make headway into desktops, where people have been using a particular user-interface pattern for some years.
Scott: It also seems that eventually, economics will dictate change on the desktop. Once you get $100 computers, $150 operating systems become problematic.
Con: The netbook is that point of inflection for many people. At $200, you can go and buy one from major OEMs like Dell or HP. A $200 netbook is a more than “good enough” computer for many people. And at that point, you start asking questions about the viability of a $400 office suite or a $150 operating system.
And that’s probably the most likely path that increasing numbers of users will move, over time, to desktop Linux. It’s one of those things that might take years. After all, it took Unix decades to reach broad acceptance in the business community.
The same thing is true with the Internet. Remember, we’re not even 20 years into the Linux epoch.
Scott: Linux definitely seems to do very well on devices where you don’t place expectations on the operating system. TiVo is the classic example; many people don’t even think of it as having an operating system, let alone identifying which one it uses.
Con: It’s a tool; product makers need something that does the job well, at the best possible price, as long as you have an open and competitive market. Of course, there have always been questions in the IT space about that “open and competitive” aspect. Microsoft has gotten a lot of attention in recent years, and before them it was other major players, like IBM. People are always asking whether strings are being pulled in the background, strings that actually prevent competition.
There have been courtroom cases about this, but at the end of the day, it won’t really matter, because it’s impossible for any one company to maintain that “king of the roost” position forever. IBM knew this, and Microsoft probably also knows it as well.
And many of us who have been in the industry long enough remember the period when IBM had dwarfed all of the major players. The IBM universe was the universe, and if you didn’t do what IBM said, then you weren’t really playing in the IT game at all. That position has now been usurped by Microsoft.
My hope, which is probably shared by a number of people in the industry, is that if we do move to the post-Microsoft world, it won’t be controlled by a single, specific vendor. And if you look at it from a strategic point of view, realistically, the only platform that can deliver vendor-neutral, open, unencumbered computing for the decades to come, is something like Linux.
Scott: We’re sort of entering a new era. There used to be a lot of talk about whether open source companies can be competitive, and now we have begun to see a lot of acquisitions of open source companies by traditional companies. And we have begun also to see that when you buy an open source company, it isn’t exactly like acquiring a proprietary company.
Con: And that can surprise the people who acquire these companies, sometimes for what seems to be exorbitant amounts of money. They seem not to realize that, while they get the name and the trademark, the code base with most of these acquisitions remains open source, and is therefore harder to monetize.
And even if you buy the technology and decide to make it proprietary, that doesn’t stop someone else from taking the current version and continuing with the open source development. The right to fork the codebase is one of the intrinsic freedoms of open source software.
Scott: It seems that they are also acquiring the intellectual capital that is in the developers’ heads. If you buy MySQL, you get all of those developers, but the problem is that you have to actually keep them. If you’re Sun Microsystems, and you acquire MySQL, and then everybody leaves, what do you really have at the end of the day?
Con: That’s right, and it’s a non-trivial problem. Companies have to be very cautious about their interaction with open source communities. To date, I’ve seen only a handful of companies who have done this successfully, and generally, they have been companies that have been embedded within open source communities themselves.
After a non-open source vendor acquires an open source product, they need to convince the staff that the new parent values the way the staff has been working, and they want them to continue. The acquiring company has to work hard not to change the acquired company’s development culture and corporate culture.
There is nothing stopping, as you said, the developers from forming a separate company after the appropriate deadlines have passed in terms of contractual negotiations and non-compete clauses and so forth.
Once that time has passed, they can jump and form their own company with the same code and developer community. The user community will find out very quickly that they are the same smart people who created the technology in the first place, and chances are that those users will follow the original developers rather than the acquiring company.
This is something that those acquiring vendors need to be mindful of. They need to strategize accordingly and do their due-diligence. You can’t really do business as usual with open source software. It has to be done differently, but traditional companies can still be successful here.
Scott: IBM seems like they’ve done a very good job about figuring out how to derive a lot of revenue from things around open source, including services, implementations, and their own proprietary software that runs on top of Linux. They have also figured out where it makes sense for them to contribute to things like the Linux kernel.
They’re still a big corporation, and they’re still out there to please investors, but they seem to have kind of figured it out. On the other hand, Oracle’s database runs on Linux and some of that kind of stuff, but I wonder about the future of some of the technologies that Sun had been holding, like Java and MySQL.
I wonder what these things are going to look like two, three, four years from now in a company like Oracle, where you just have to imagine that there are tens of thousands of people who just fundamentally don’t get open source.
Con: I believe that many organizations, possibly most, embody the philosophy of the person at the top who runs it. IBM had major financial problems in the early ’90s, when it was losing billions of dollars every quarter. They brought in Louis Gerstner, who I can tell, even from an outsider’s perspective, did a fantastic job. One of the reasons he did a fantastic job is that he used to be a customer of IBM’s five years beforehand, when he was a fairly senior MIS player at American Express.
He got the rough side of IBM as customer, and he wanted to make sure that no other customers got that rough end of IBM. And he set about putting mechanisms in place to change IBM’s culture into the current one that IBM has. It was still a pretty big and byzantine machine, but it overcame much of the negative attitude and stifling process that it had 20 years ago.
It seems that Sam Palmisano is a master’s apprentice to Louis Gerstner, with much the same philosophy and the same outlook about how to deal with customers, competitors, and R&D, as well as what to invest in and how to make money by building solutions for customers.
That is a fairly different species from Oracle, where Larry Ellison exudes a very different type of personality from Louis Gerstner or Sam Palmisano. Oracle has traditionally been much more product and sales revenue focused, which makes their outlook and their positioning about how to use open source quite different from IBM.
Scott: On one hand, you sense that when these companies buy an open source company, they don’t understand entirely what they’re getting. On the other hand, you sense there’s an awareness, or at least this sense on these very large companies that they know they need to figure it out.
You see Microsoft dipping its toe into the water by contributing some code to the Linux kernel and some other stuff. Linus has even kind of defended them. It was interesting that Microsoft contributed integration components that let Linux run as a guest OS on top of their virtualization platform.
Immediately, lots of people shouted that it was self-serving. Linus responded by saying, essentially, “What do you think Dell’s doing when they contribute device drivers for their particular hardware?” That’s the whole point of open source: everybody gets to be completely self serving and in the process of being self serving, they build this great thing that everybody gets to use.
Con: That’s absolutely correct, and there’s no negative issue associated with that. The only reason for the pushback by the open source community is that in the previous decade or so, a broad mistrust of Microsoft has taken root.
In a perfect world, Microsoft becomes just another major industry player, like IBM, Oracle, HP, or what have you. In reality, though, many in the open source world still see Microsoft as having a bit too much control over what happens within the industry.
People will point to Microsoft and say they’re making mistakes and being silly and everything else. Microsoft are smart guys. They will sit down, and with mathematical precision they will itemize almost like a strategic battle plan for why and when they have to flip from having a positive or a negative stance on open source.
If there’s a mathematical point where they gain from distributing GPL drivers to allow for optimized hosting ofLinux on their virtualization platform under Windows Server, they will do it at that point. Not before and not after. And as you’re saying, that’s no different from how Dell or some other company would approach the problem.
Unless a vendor can visualize how they can gain strategically in specified areas by broadly supporting open source, they’re not likely to take a forward step. They will do it on an as-needed, case-by-case basis where it makes sense for their business.
If I were a shareholder at Microsoft, I would expect no less. The interesting thing is that, of course, Microsoft has run the numbers and calculated that they would be at a strategic disadvantage if they weren’t doing these injections of open source code and also normalizing relations with the open source community.
Scott: I don’t claim to know their inner thinking, but I would echo what you’re saying and observe that at some point, Sun Microsystems did something very similar. They had a proprietary Solaris operating system running on proprietary hardware, and at some point they decided to go with open source.
They did it very differently; they decided under Jonathan Schwartz to make a fundamental shift and get completely to open source in the time frame that they could, but they saw the writing on the wall. Proprietary UNIX systems were competing against freer, much lower cost Linux, which was just going to erode their position year after year.
IBM saw similarly that they had to get on board with Linux. They did the calculation and saw that they could still make a lot of money off of these other things, but that if they didn’t get on board and have their products running on Linux, their proprietary UNIX was just going to decrease in terms of share.
Con: All of these different firms have a particular strategic positioning. It also shifts with time and is impacted by the culture of the organization. Sun is a high-end engineering firm, and I imagine the high priests of the Solaris development scene looked down their noses at Linux and said, “Why on earth would we have to compete with this thing?”
But that’s an outlook which was there maybe 10 years ago or so, and now they can’t really make that kind of statement. The growth of quality and functionality within Linux far outstripped what you would get in terms of features and capabilities within Solaris, because of their different development models and licenses
And at the end of the day, you just can’t compete with Linux.
Scott: Right. You can’t compete with Dell and HP and all of these companies that are contributing to it. They are doing it in a very self-serving way, but you end up with a Linux that has an enormous number of drivers and supports a huge amount of hardware.
When Intel comes out with a new chip set, they contribute code to Linux to make sure that Linux takes full advantage of their new features and the chip set, while other operating systems, like Solaris, AIX, or HP-UX sort of schedule it. They get around to it when they get around to it.
It is very difficult for any one entity that wants to control all the code to compete with that sort of thing. We’ve talked a lot about these players that have been around for a long time, and then there’s this sort of new kid on the block, in the open source world, and that is Google. Google shows up and it becomes sort of the command line to the Internet.
It’s a text box and a button, and you don’t really have a lot of sense for how they are doing what they are doing but you get the sense that have figured out something with commodity hardware. They are not buying a bunch of Windows licenses. They are using some kind of UNIX-ish operating system.
They’re basically building a massive computer that spans the globe. They start not necessarily giving away their search algorithms, but they start contributing a lot of stuff to open source. They have GWT, they roll out Android, they’ve got a Chrome browser. They are talking about a Chrome operating system and these potentially very disruptive kind of things.
What does your crystal ball tell you about that?
Con: Google is in the position to have a very dramatic role in the next decade, for two reasons. One is that they have a fantastic name brand, which gives them the reach to maximize penetration into the mainstream. The average Joe or Jane is not that keen on acquiring technology, because technology is complex and potentially even scary.
Because of this, they want technology to be incredibly simple and from a big name vendor, so they feel confident that they can get it to work. Google can deliver that name brand clout, and of course they have enough money to be a serious competitor to anyone in the industry.
But, the question to ask is, “If you were Google, what is your end-game strategy? What are you trying to achieve?” From Google’s perspective, my guess is that they want to have computer terminals in as many places as possible to access ‘their’ Internet. So you will get your smart browser plugged into simple operating system terminals. You might have five of them in your household: you might find one on your desk, one on your phone, even one on your more powerful netbook, or notebook, or what have you.
But at the end of the day, they are looking for the standard utility model of computing. That’s where they have immense strength, because more than anybody, they know they have to maintain and make money by running hundreds of thousands of servers in the backroom and delivering a consumer-grade interface to as many people as they can.
Google will inject money into every possible technology or process they can, to help to remove the roadblocks towards that reality, and to erect bridges to make sure that that reality happens. Browsers have gotten pretty good, and increasingly, they’re all coming to the party with regard to supporting standards.
Google wants a complete operating system in the browser. They want to be able to run Quake 3 in a browser. Your browser does currently not run Quake 3, so Google make a browser that can, because they do not (in reality) want an operating system; they want the whole world to fit within the browser.
Of course this is the exact antithesis of what Microsoft wants.
Google also recognizes that they can increasingly tell people that they can use the Google web ‘operating system’: office suite, and mail and calendaring suites, all of which are online, free, and available for Mac, Windows, and Linux, all on your web browser.
However, they also still note and understand the industry’s maxim that “he who controls the desktop operating system can control the universe”, and this is what Netscape found to its detriment a dozen years ago when Microsoft decided to compete in the browser space. Microsoft controlled the desktop operating system, and so it could do things in such a way as to outmaneuver competitors’ applications that are running on top.
But from Google’s perspective, the end-point would be essentially zero-cost, friction-free software, on as many devices as possible, all linking back into the Googleplex in some way, shape, or form. One way to do this, of course, is to develop a current-generation OS that they can make available to as many low-cost, web tablet manufacturers as they can possibly take on board.
And these manufacturers will be making computing devices that will be increasingly low-cost and ubiquitous, so it will become essentially irrelevant what OS you’re running on this web tablet, as long as it is consumer-friendly, ultra-low cost, and can get you into the Internet, which of course means into Google’s arms.
Scott: I totally agree with your analysis, when I look at Android targeted at handheld devices, and maybe netbooks, as well as their plans around Chrome OS. I also have to think about HTML 5, which as you mentioned lets you run Quake in a web browser with storage, offline capabilities, and a lot of other stuff.
It’s a lot of separate efforts, but they are obviously all converging on this goal. I don’t know if this is the right characterization, but Google could become the next Microsoft to some degree. It becomes a ubiquitous platform. It’s not necessarily on your device, because whether your device is running Chrome, Android, or some other OS, it’s going to have a browser, and as more of these experiences become web based, Google could control the entry point to everything.
Con: As long as they are perceived as roughly living by that credo, “Don’t be evil,” that will work. I had this discussion a while back when someone asked me, “what will stop people from using Microsoft?” I responded, “How many years’ worth of Microsoft documents have you got in your PC hard disk?” He said “Like 12 years worth.”
And there you have it. How easy would it be to convert all those documents into something different? You can’t do it, because there are all of these macros, complex embedded formatting, etc. It can’t be done, so many are stuck with using Microsoft Office forever.
The corresponding question is, “How easily can you shift from using Gmail to Yahoo mail?” It would be a bit of a pain, but with a couple of hours’ work, you could bring across all of your contacts and copy down all of your mail.
And again, how easy would it be to use an alternate search engine? Well, you can do it pretty much instantly. You can start using Bing, or you can start using Yahoo Search. That’s the key difference between what Google does and what Microsoft has done. It’s a lot harder for Google to achieve serious lock-in.
They remain where they remain for as long as people perceive them to be doing a reasonable job, possibly a good job, and as long as people perceive them to be complying with their own personal ethos of how they think a business should work. Otherwise, it’s very easy for any consumer to say, “Well actually, no. I’m not using anything from Google. I get my mail from Yahoo. I get my online storage from Microsoft. I use FaceBook for my chat, and I don’t use Google search.”
The key reason, of course, that this is ease-of-exodus is possible is that to a large extent, Google understands that this is a double-edged sword for them. In order to be successful and to accelerate the adoption of these kinds of technology, Google has to use open standards. And by definition, a vendor cannot achieve lock-in through open standards.
So for its talk protocol, for example, it uses XMPP, which is the Internet-standard chatting protocol, the open source Jabber chatting platform. They don’t develop their own proprietary discussion protocol. It’s the same for mail, calendaring, and the list goes on and on.
Now, at no point are they in a position to be able to fully dictate terms on their software. They can say, at any point, “Well fine, we’re not bringing out new versions of this software.” But if their software is of enough interest to enough people in the community and to other vendors or technology companies, these will continue to maintain it, or provide viable migration paths to open standards-based alternatives.
Once again, it’s all made possible by open source software that uses open standards. This combination can’t put Google in any more of a strategic leverage position than any other vendor; as long as they are doing a good job developing Android or Chrome OS or the Chrome Browser, you will probably continue to come back to them.
At that point, it becomes just a standard consumer choice option. If Google does turn evil, consumers can say, “Look, Google’s turned evil, but I really like what they are doing, so I’ll go with them.” Or “It’s just as easy for me to go and use the newly named Bob’s Browser, forked from Chrome, from the group over in Cincinnati which is doing fantastic work. I don’t like Google any more.”
That’s the key fundamental difference. Google wanted to build a a viable client-side platform quickly and to make it as broadly-acceptable as possible, which introduces that double edged sword for them to contend with: Google has to use open standards and open source to maximally gain penetration, on the one hand. On the other hand, it can’t really achieve consumer lock-in.
Scott: You see a sort of meritocracy already playing out. When I go to a conference like OSCON, where it’s just all open source people there, I see an enormous number of Macintoshes, and I see an enormous number of iPhones. I know that the people there know that Apple is not an open source company. In fact, it’s very secretive. It likes to kind of surprise the marketplace with new products and this kind of stuff.
Yet people just want the products so bad that at the end of the day, they want the products for what they do. They want the iPhone for what it does. But there isn’t an enormous amount of lock in with an iPhone. It’s not like they couldn’t go buy a different phone tomorrow, and probably move their contacts over and move their music over and be in business.
Con: Apple is in a very enviable position. Although at the end of the day, Apple builds proprietary software, proprietary firmware, and proprietary hardware, people want what they offer.
I have been, for the past 13 years, a user of the Opera web browser, which is a proprietary piece of software. Now it’s free, although I used to have to pay for it. I love that web browser, but if Opera was not adhering to what I would perceive to be good business practices, it wouldn’t take me too long to move what I’m doing away from Opera onto Firefox, or onto Chrome OS or what have you.
So that’s how I would judge Apple in this instance. I don’t really use Apple operating systems, but people who do use them use Firefox and they use OpenOffice.org and whatnot. So from my perspective, these users get the OS environment that’s suited to them, but what interests me is that as long as Apple adheres to open and Internet standards, so that I can share material and can correspond with Apple users, I’m OK with Apple doing what they do, as they don’t have a chokehold on our industry in any real sense.
Scott: Well, that’s probably a great place to wrap up. Thanks for taking all the time to chat.
Con: Thank you for the opportunity. It’s been wonderful.