Images Through Glass     blog-L     LL-Letters     Photo Gallery     Index     PDF Files     J-Page     BizZ     Home

Note:  This is basically a notes page for myself to dump in computer-related stuff that I think I might want to reference in the future.  If it's of interest to you, great, otherwise, you might want to just ignore this, as it has no particular focus other than that it is computer related....


"In this day and age, computer systems are getting faster and more capable, but they still do not eliminate the need for a sensible, intelligent person to run the show. Computers will never be 'smart enough for any fool to use.' ... When you go looking for a software package, don't just look for which one has the most automation. Don't believe that because it has all that automation, it will make your job or your life easier. It won't. ... There is no substitute for using your own brain to get a job done right." - Howard Chu of Highland

Firewire vs. USB: A Comparison
By Nathanael Copyright © 2005

The following article is based on years of experience. It is provided as a free service to our customers and visitors. However, is not responsible for any damage as a result of following any of this advice.

Copying the contents for commercial purposes is strictly prohibited without's written consent. However, you are welcome to distribute these computer support tips free to your friends and associates as long as it's not for commercial purposes and you acknowledge the source. You are permitted and encouraged to create links to this page from your own web site.

If you have used a computer within the past five years, chances are you've have used USB (Universal Serial Bus) devices many times. From mice and keyboards, to printers and external hard drives, USB devices are nearly ubiquitous ?in fact, over 1 billion USB devices have been sold.

On the other hand, Firewire-based devices are somewhat less prevalent. Nearly all digital camcorders sold after 1995 have included a Firewire connection, as have all modern Macintosh computers. Additionally, many new external storage devices include a Firewire connector. However, lower-end devices such as mice and printers are rarely (if ever) seen with Firewire connectivity.

In the next two sections, we will look at the history of these two similar I/O ports. Then, we'll compare them side-by-side to see which technology is better for specific applications.

USB: A Brief History

Version 1.0 of the USB specification was released in January of 1996 by the USB Implementers Forum (USB-IF) and was followed up by version 1.1 in September of 1998. A theoretical maximum of 127 devices per controller is specified. Both versions 1.0 and 1.1 support a maximum transfer speed of 12Mbps ("Full Speed") and can fall back to 1.5Mbps ("Low Speed") if need be.

Note that these data rates are in Megabits (Mbps) per second, as opposed to Megabytes (MBps) per second ?a commonly confused notation.

USB version 2.0 was released in 2000, upping the theoretical maximum transfer rate by a factor of 14 to 480Mbps ?dubbed "Hi-Speed". USB 2.0 devices are backwards-compatible with USB 1.x devices and controllers, and can fall back to "Full" or "Low" speed in order to coexist with older devices. Nearly all new products on the market are USB 2.0-compatible.

Both USB 1.x and USB 2.0 allow the use of two separate types of connectors ?Type A and Type B ?depending on the requirements of the device itself. Type A connectors are almost always used on the host side (computer or hub), while Type B connectors are smaller and are frequently found on the device side in printers, scanners, and other similar hardware.

Both types of connectors can provide up to 500mA (milliamps) of power to connected devices, though devices that require more than 100mA should be self-powered as each USB port generally has a maximum of 500mA of power to share between all devices. A device that draws all of its required power from the USB bus is referred to as a "bus-powered" device.

Windows 95 OSR2 (OEM Service Release 2) included limited support for USB; the original release of Windows 95 had none. Windows 98 ?and more importantly, Windows 98 SE ?added much better support for USB, but Windows XP's USB support is the best and most robust, by far. Apple's Mac OS has supported USB devices since prior to version 9.0.4, but this release of the operating system added substantially better support.

Firewire: A Brief History

The origins of Firewire date back to the mid-1980s. Engineers at Apple Computer devised a high-speed data transfer technology for Macintosh internal hard drives they called 'Firewire'. Realizing the potential for a technology that allowed high-speed transfer to and from hot-swappable devices, Apple presented this technology to the Institute of Electrical and Electronics Engineers (IEEE).

In December of 1995, the IEEE released an official Firewire specification, dubbed IEEE 1394. This specification, sometimes referred to as 'Firewire 400', describes a hot-swappable peripheral interface with transfer speeds of 100 Mbps, 200 Mbps, and 400 Mbps. During the late 1990s, this standard found its way into Sony electronics (mainly digital camcorders) under the title 'i.LINK'. In January of 1999, Apple released what was probably the first personal computer system to include Firewire ports by default: the Blue PowerMac G3. All Macintosh models from then on have included Firewire connectivity.

Firewire cables come in two variations ?4-pin and 6-pin. 6-pin cables provide up to 30V of power, allowing for fully bus-powered devices. 4-pin cables do not provide power.

In April of 2002, the IEEE released an updated Firewire standard, dubbed IEEE 1394b. IEEE 1394b allows for theoretical maximum transfer rates of up to 3.2Gbps. Apple commercially released a subset of this new standard under the title 'Firewire 800' in 2003.

Firewire 800 devices support a maximum transfer speed of around 800Mbps. Firewire 800 adds a new cable type ?9-pin cables (also called 'beta' cables), which support the full speed of Firewire 800.

Firewire 800 is backwards-compatible with Firewire 400 when 'bilingual' (9-pin to 6- or 4-pin) cables are used. Firewire 400 devices will still run at Firewire 400 speeds, even when connected to a Firewire 800 host.

The Comparison

General Peripherals: USB Wins

USB has almost completely replaced older I/O connectors such as parallel, serial, and MIDI (joystick) ports. Instead of a confusing collection of incompatible devices and connectors, you have a one-size-fits-all connection that works on nearly all PCs manufactured in the last 5-8 years.

While USB has not completely replaced PS/2 ports, USB mice and keyboards are readily available. Nearly all recent scanners and printers have USB connections, as do most other low-bandwidth peripherals.

On the other hand, Firewire is almost completely absent from this category. As mentioned before, Firewire is impractical for low-bandwidth devices; this, coupled with the fact that most computers (besides Macintoshes) do not include Firewire ports by default, has kept Firewire-enabled devices in this category out of this market.

Digital Imaging/Digital Video: Tie

Firewire is much more prevalent in this category. Almost all modern digital camcorders come with Firewire connectivity. Because of technical differences, Firewire is a better bet for transferring uncompressed (raw) video from digital camcorders, even though USB 2.0 has a higher maximum speed (400Mbps vs. 480Mbps). Not many camcorders have Firewire 800 connectivity yet, but this is expected to change over the next few years.

Most digital cameras still use USB for image transfer. This is likely due to the higher level of compatibility with current computers ?nearly all have USB ports, while a considerably lesser portion have Firewire ports.

External Storage: Firewire Wins

This category includes external hard drives, external optical drives/burners, and generic external drive enclosures. Though USB 2.0 and first-generation Firewire are nearly neck-and-neck, Firewire can provide much more power over the bus ?30V as opposed to 5V for USB, which means that external Firewire drives frequently do not need a separate power brick.

Many manufacturers of external storage devices now produce models that include both Firewire and USB 2.0 ports for maximum versatility. These types of devices are probably the best bet for both speed and compatibility.


USB and Firewire both have unique strengths and weaknesses. USB's ubiquity makes it ideal for devices that require high compatibility with current hardware. Firewire's generous bus power and internal architecture lends well to external storage and digital video applications.

If you have any questions relating to this or any other topic, feel free to post them on the Help Desk.

Last updated: 03/24/05

If you find this article useful, please create a link to it:
 - from your website or tell a friend about it. If you have any comments or suggestions about this article, please email

A Quick Start to Debian

Debian is one of the main Linux distributions, and although not designed for the beginner is the basis for many newcomer friendly versions. Aimed at the more experienced, it has the reputation of being rock solid stable and, once installed, extremely easy to maintain.

This page is not designed for the complete newcomer, but it is my hope that I'll save the more experienced Linux user a bit of time. The information is all out there, but sometimes you have to dig a little bit to find it. (It's also to save me the trouble of looking up things I forget if I have to install it on another machine--as one friend said, since it's true that once installed, you don't have to install it again, one does forget.)

Firstly, you have to install Debian. The easiest way is from CD. Although they don't make it easy to find a complete iso, suggesting you use jigdo, a bit of searching will usually find it. (Hint: try However, using jigdo isn't very difficult. One installs it (for example, in FreeBSD it's in ports), gets a url from the Debian web pages and types something like


(If your browser broke that, it should be on one line) After that, just press return when it asks for files to scan and it will ask for a Deb ftp site. You can type in (assuming you live in the US)

and it will do the rest, creating an iso image to burn. (I use stable, woody, as an example, one could use testing or unstable as well). Note that jigdo does require wget. FreeBSD installs it as a dependency.

I just use Woody (Stable) as it seems to have more on the first CD. I downloaded Sid (unstable) once, and it didn't seem to really be an install CD. (I just downloaded CD-1.) I did a quick google search, saw that didn't sound completely wrong, shrugged and just got Woody. I didn't research it in depth at all, just did a quick google search, so take that statement with a grain of salt.

Since, as you'll see below, upgrading on a reasonably fast machine with a broadband connection probably takes less than an hour, I didn't research this any further.

Once you have the CD burned, installation, despite its reputation of being difficult to install, is fairly straightforward. Again, this page is not for the complete beginner, so I'm not going to walk you through the installation. They have information on their web page. If you are used to text or ncurses based installations, it is fairly typical. Once the CD boots, hit F3 for the various options--I choose bf24 which uses a 2.4 kernel rather than the default 2.2 one. You'll have to pick your swap and / partitions, and if you're only used to GUI installs, this might be initimidating. However, it's no more difficult than say, Slackware, and actually far quicker.

Upon reboot, you are given various configuration options. One is deciding what you will use for your sources. I usually choose the first two or three ftp sites listed in the US. When all this is done, you can log in. (I skip tasksel and the like during this part, it makes the upgrade go more quickly.)

As Debian unstable is far more stable than most distros' release versions, the first thing I do is upgrade to unstable. (Assuming that one installed the stable version.)

Note that you can actually install unstable from the beginning. I learned this afterwards. When first installing, add verbose to your boot arguments--for instance, I install the the 2.4 kernel so when booting from the cd I type at the boot prompt

bf24 verbose

Then, upon first reboot, when choosing sources and the like, unstable is one of the choices.

For those used to source based distros like Gentoo, the upgrade of an entire distribution is amazingly quick. Debian uses binary packages. The first thing to do is change the ftp sites to get unstable sources.

In /etc/apt there is a file called sources.list. It will have lines like

deb stable main non-free contrib

Change stable to read unstable (but leave security at stable) then

apt-get update
apt-get upgrade
apt-get dist-upgrade

This will take awhile, and will require a bit of input while going on. However, when it's done, you have successfully upgraded from stable to unstable.

Debian's apt-get is well known as one of the best package management tools around. People have different ways of doing things and for me, I usually install ssh next

apt-get install ssh

This installs ssh and gives you the option of running the sshd daemon as well. It will also create keys. Next, I install a few other packages, such as wget, sudo and the like.

Something that I've only noticed in Deb and its derivatives is that sometimes, one way or another, a user is added to the sudo group. (Being used to FreeBSD, I usually create a wheel group and use that as the group that can do things.) If a user is in the sudo group, if they run a command with sudo, they aren't asked for a password. So, if that's happened, and you wish to change it, as that user

sudo deluser $USER sudo

will work.

Next, I want to install X.

Debian's tasksel will do this for you. However, it installs, in my opinion, far too many packages that I don't want. So, I just install x-window-system and xlibs-dev. By now you probably have the syntax figured out

apt-get install x-window-system xlibs-dev

As these are binary packages, it is far quicker than a source build.

After installation, it takes you through a dialog similar to xf86config.

If you leave things as is, the next reboot will boot you up into X. I prefer to boot into text mode so I remove xdm from the start up scripts.

update-rc.d -f xdm remove 

In Debian, once it's installed, most things just work. For example, if you install mozilla-firebird and flashplugin-nonfree, Flash will just work in firebird. I actually prefer opera, which doesn't seem to be included (as of January 2004) in unstable (sid). So, I download the static linked deb package, add the /usr/lib/mozilla/plugins path to its plugin directories path and then install lesstif2. After that, Flash works in Opera as well.

Sound and other things in the 2.6 kernel

This one took a bit of searching to get working. Upgrading to the 2.6 kernel and using ALSA took away my sound, which had pretty much worked out of the box when I added the module for my card during installation. Firstly, I installed alsa-base with apt-get. After that, in compiling the kernel, as well as including my card, I added the modules for SND_PCM_OSS SND_MIXER_OSS and SND_SEQ_OSS. Then, add the following lines to /etc/modules.conf

alias sound-service-0-0 snd-mixer-oss
alias sound-service-0-1 snd-seq-oss
alias sound-service-0-3 snd-pcm-oss
alias sound-service-0-8 snd-seq-oss
alias sound-service-0-12 snd-pcm-oss

(This has, at times, if you compile the above options, been automagically done for me.)

I also (but only once) had an issue when upgrading to the 2.6 kernel--suddenly, although DNS was working, I was getting connection refused from the apt sites and couldn't get to web sites either. A bit of googling indicated that it was the tcp_ecn flag. I fixed it with

sysctl -w net.ipv4.tcp_ecn=0

XFree 4.3.0

NOTE: (Since this was written XFree 4.3.0 is now in unstable. Another thing to PLEASE note about the advice below is that experimental IS experimental.)

The downside of Debian's well-earned reputation for stability is that its packages are often slightly older versions. If one needs, or just wants, Xfree 4.3.0 this is how I went about getting it. First, I added this line to my /etc/apt/sources.list

deb ../project/experimental main contrib non-free

Then, run apt-get update. After that

apt-get -t experimental install x-window-system

(By the way, one can always search for packages at

One of my mentors has created the following guide for determining whether to run stable, testing or unstable. He pointed out, after looking at this page, that despite my blithe assurances, much can go wrong in unstable, which is why it is called unstable. I'm putting his guide below.

Is it a server?
Y - run Stable
N - next question

Is it an important workstation?
Y - run Testing, or even Stable if it's really mission-critical
N - next question

Does it matter if this workstation actually works or not?
Y - Next question
N - Run Unstable, and install anything from Experimental that strikes
your fancy.

Is it a workstation on which you can afford some possible downtime?
Y- Run Unstable

N - Go back to the top and try again, or write another question :-)

There's a few other things that are a bit different in Debian. Many of them are covered in the distrowatch article mentioned above. This covers, among other things, the Debian way to build a kernel. Debian's way to build from souce is in one of their faqs, but I also have it covered in the Debian section of my fluxbox page. I also cover the steps necessary to get Japanese working in Debian in my Japanese in *nix page.

The main purpose of this page is to save the person new to Debian a bit of searching.

An Open Letter from Hector Ruiz, AMD Chairman, President and Chief Executive Officer

The microprocessor is the brain of every computer, a transformational technology in today’s world. And as in all markets, innovation in the microprocessor sector depends on competition – the ability of consumers and businesses worldwide to choose solutions based on one microprocessor over another.

Our competitor has harmed and limited competition in the microprocessor industry. On behalf of ourselves, our customers and partners, and consumers worldwide, we have been forced to take action.

We have filed a 48-page, detailed Complaint in federal district court. Because, as our Complaint explains exhaustively, Intel's actions include:

  • Forcing major customers to accept exclusive deals,
  • Withholding rebates and marketing subsidies as a means of punishing customers who buy more than prescribed quantities of processors from AMD,
  • Threatening retaliation against customers doing business with AMD,
  • Establishing quotas keeping retailers from selling the computers they want, and
  • Forcing PC makers to boycott AMD product launches.

For most competitive situations, this is just business. But from a monopolist, this is illegal.

These serious allegations deserve serious attention. Earned success is one thing. Illegal maintenance of a monopoly is quite another.

Intel's behavior is much more than meets the eye. You may not have been aware, but Intel's illegal actions hurt consumers - everyday. Computer buyers pay higher prices inflated by Intel's monopoly profits. Less innovation is produced because less competition exists. Purchasers lose their fundamental right to choose the best technology available.

We believe the legal process will work. In the meantime, the men and women of AMD will continue to drive innovation, focusing on our customers and on the people who use computers at home and work every day.

At AMD, we know innovation. We thrive on competition. And we depend on a market based upon freedom of choice.

Read our Complaint. Demand innovation. Choose fair and open competition.

Hector Ruiz
Chairman, President and Chief Executive Officer
Advanced Micro Devices

To share your thoughts about innovation and fair and open competition with us, please e-mail

Novell Looking for Acquisition Targets?
Written by Bryan Richard  
Thursday, 04 January 2007

Matt Asay recent blogged about how Novell might be in the market to make an acquisition this year in the virtualization space. He lists XenSource and Altiris as possible targets.

If Novell wants to maximize the potential of their Microsoft alliance and bring about a scenario like Canonical founder, Mark Shuttleworth, outlined in a recent Red Herring interview...

    Microsoft is going to claim that deploying Linux anywhere, unless you pay Microsoft a patent fee, is a violation of their patent and they haven't proved that yet. But they certainly seem to be positioning themselves in such a way that they could do so.

... then you have to think they'll buy XenSource.

Why XenSource? Because it's at the heart of Red Hat's pending RHEL 5 virtualization features.

If you're into doomsday scenarios -- and you kind of have to be these days -- you have to wonder to what extent Novell would be willing to use as a competitive weapon the agreement with Microsoft that excludes Novell customers from patent litigation.

If Microsoft has a patent covering Xen-like virtualization tucked away somewhere in their intellectual property vault then Novell could use that to plant doubt in customers minds about upgrading to RHEL 5.

Novell paid handsomely for that patent indemnification -- both in cash and community PR -- you have to assume they're going to put it to use and acquiring XenSource would put them in a position to leverage it.

Of course all of this idle speculation on a slow news day could amount to nothing. But regardless of whether the intellectual property threats are real or implied, the Open Source market seems to have graduated from feature wars to information wars. The open solutions of 2007 could start to be judged not just by if they solve technical problems but if they also pass muster with a company's Chief Legal Officer.

And that's a shame. The last thing that Open Source vendors need is customers asking, "Is it safe?" It's what SCO aimed for and failed to accomplish.

But if done correctly, Novell could show SCO a thing or two about how the game is played.

Linux: Ubuntu Founder On Microsoft "Challenge"

Canonical CEO Mark Shuttleworth talks why it may finally be time for Linux to out-innovate Apple and Microsoft on the desktop.

December 29, 2006

By Falguni Bhuta

Taking a trip into space hasn't been Mark Shuttleworth's biggest challenge. Instead the one-time space tourist counts building an open-source company and working to hook users on the Linux as his most testing venture.

Mr. Shuttleworth founded the Ubuntu project in 2004 to distribute a free desktop operating system based on Debian Linux that would compete with Microsoft Windows.

In 2002, Mr. Shuttleworth took a trip into space becoming the first South African in orbit and the second space tourist ever. The entrepreneur previously founded Thawte Consulting and sold it to Verisign for $575 million, has headed a venture capital firm, and has finally settled on pushing open-source software as his next career move. The Ubuntu project is controlled by UK-based Canonical, where Mr. Shuttleworth is the CEO.

Mr. Shuttleworth spoke with Red Herring about recent developments in the world of open source and his plans for Ubuntu.

Q: How are the events in the open source industry in the last few months affecting Ubuntu, if at all?

A: There were two big strategic announcements that were made this year by non-Linux players. Oracle [said] that they will be providing support for [a version] of Red Hat Linux without the trademark and Microsoft and Novell would collaborate in a number of areas, and that announcement had some interesting intellectual property considerations so those both have been significant for us.

On the Oracle front, it hasn't really changed our position, because we remain the only group that's focused on providing Linux free of charge but on a commercially sustainable basis. So Oracle competing with Red Hat may be the cause of the commercial distress without really impacting on our strategy.

The Microsoft announcement is potentially more interesting. It is setting us up for a situation where it will be almost impossible potentially for Linux to remain a free platform, if Microsoft is able to assert a considerable level of intellectual property ownership of a Linux account. That will have a very significant impact on developer participation and innovation in Linux, we feel that it is a very significant challenge.

Q: So are you saying that if Microsoft finds a way it's going to be hard for Linux [to] stay as a free piece of software?

A: Microsoft is going to claim that deploying Linux anywhere, unless you pay Microsoft a patent fee, is a violation of their patent and they haven't proved that yet. But they certainly seem to be positioning themselves in such a way that they could do so.

They are really trying to get something to legitimize their claim, so the deal with Novell had a lot of money attached to it. And as a part of that deal, Novell is lifted up in its stand to legitimize much of its claim. So it's a very interesting strategic move and there are a lot of people on the other side saying that there are absolutely no intellectual property issues with Linux that this is kind of a game.

Q: What are your thoughts, do you think that they have enough Intellectual Property to threaten Linux users to sue them if they don't use Microsoft?

A: It's very possible that Microsoft does have a patent amongst hundreds of thousands of patents out there, which covers something that Linux does, but until they come out and say which patent it is, it's impossible to know. No one is ready to overrule the patents out there [and] no one could possibly make a decision about that other than Microsoft.

The other thing of course is that as soon as it's clear, if there is some sort of infringement of Microsoft intellectual property, usually it's very easy to work around that, essentially by re-writing the PC software in question. So a lot of people are very confident that even if there is something that is an infringement of a patent it could very quickly be resolved. So Microsoft didn't actually want to be on the hook for saying this is a specific patent that is infringed, because then the Linux developers would work around it. 

On the other hand, what they do want to do is they want customers to feel slightly nervous of Linux. I think Microsoft is certainly sort of becoming a smarter operator into how they interact with Linux and with free software. They spent a lot of time saying it doesn't exist, it is a toy, it is a cancer, and it is dangerous, and calling it anti-capitalist, and now they seem to be engaging in a much more realistic competitive pragmatic fashion to that problem.

On Desktop Innovation

Q: So do you think that Ubuntu, and maybe another player such as Linspire, will have to eventually team up with Microsoft to avoid any conflicts?

A: We would never pay a patent license fee to Microsoft, because we don't believe that there are any patent issues, and our economic model is essentially to make the software freely available. And if you are making the software freely available, you can't obviously pay a patent license to somebody else for a copy of the particular version. So while Novell may feel that they can do a deal like that, Ubuntu would absolutely never do it.

I think obviously we will be leading the charge in terms of showing the weak points in Microsoft's assertion. It also means that we will be able to represent at least a threat to Microsoft because not only we will be competing for an end product but we will also be competing with their way of doing business [because] we have a different economic model, fundamentally different from theirs. There is no per-seat license fee.  If they call they can negotiate with us in the same terms that they would negotiate with someone like Novell.

Q: You said it has been a big year for Ubuntu. Why do you say that, what are some of the milestones that you have reached?

A: A couple of milestones, this is the year that we put out our first release in enterprise. We still have a reputation over the last of two years of focusing very heavily on the desktop so that made us a very popular version of Linux for people who are Linux power users. We have grown to the point where those power users were starting to say "I would really like to start using Ubuntu on my servers at the company where I work. But in order to do that, you need to provide support for a much longer period of time" and so this is the year that we put out our first enterprise quality release.

Also, that was the year where we formed a couple of very significant partnerships and alliances. Our relationship with Sun was very significant; for Sun Microsystems, it was a pragmatic way to embrace Linux. They had some very interesting new hardware technology in the form of a massively multi-core chip. The Sun guys are coming up with processors that have six to nine cores. So they wanted to be able to take advantage of that hardware in the Linux base, and we were able to work with them on that.

Q: What about growth in adoption rates, any kind of numbers that you can give me?

A: We know now that there are probably at least 8 million [Ubuntu] users.

Q: Do you think 2007 is the year of Linux on the desktop?

A: It's been the next year for the last five or six. I certainly think that we are seeing an acceleration in interest in the open source community solving desktop type problems. If you [go] back five years or six years, the people writing the core software in Linux were all people who were responsible for service. [Now] we are seeing a real shift in that the opens source community [itself] suddenly wants to solve the desktop problems. They want to show that innovation in free software on the desktop really can demonstrate exciting ideas that people expect from their Windows or their Mac. 

Microsoft and others, a lot of them say that free software and open source is all about copying what was being done before in proprietary software, and for a lot of time that was true. The world we are seeing is that, as soon as the free software reaches a point where it's as good as the proprietary software, suddenly all the innovation shifts to the free software. Original innovation, and we saw that for [browsers] where once Firefox had reached the level of parity with Internet Explorer, suddenly it became this hotbed of innovation, and now it's Microsoft that's scrambling to catch up with the free software browser.

Now, what may happen in 2007 is that we suddenly feel desktops being very hot ground for innovation, and new ideas, desktop style ideas, coming through in Linux suddenly made the proprietary software guys feel like they have to catch up. Which is different of course [from] saying that 2007 will be a year when all suddenly really want to switch to Linux, but it could well be the year when suddenly Linux starts to pull ahead in terms of innovation with the pace of developments in change.

On Microsoft Vista

Q: 2007 also saw the launch of Microsoft Vista: how is that going to affect Linux on the desktop especially. Do you think that Vista will kill Linux on the desktop, or do you think it will actually make more people adopt Linux?

Q: Well, that's going to get into a very complicated set of factors. On the one hand, Vista is a very polished product, [it has good] features there and [Microsoft] needs to be credited for that work. On the other hand, it is expensive, more expensive than previous versions of Windows. The way they have priced it and the number of different versions of it, you will figure out what most people are going to end up paying, it's gone up. 

They are going to try to enforce their licenses much more aggressively in parts of the world where that pricing is an issue, where people have - until today - continued to use pirated versions of Windows. Suddenly [they will] have to consider other options, and we do see that in emerging markets in particular, Linux is taking off because people want a stable platform that is cheaper.  So in that sense, yes, Linux might well benefit from the release of Vista from a pricing and licensing control point of view.

In other senses, Vista is going to drive a whole round of applications development, a whole round of other things, which will become incompatible with Linux. In Vista, Microsoft has taken steps to break compatibility with pieces of the Windows' infrastructure, which Linux has already reverse engineered. So if you look for example at the pop-sharing capability in the previous versions of Windows, Linux is very compatible with that. So you can easily deploy Linux in the same environment as Windows, and Microsoft has taken some very specific steps in Vista to make that hard for Linux to do, so one thing, it is going to be something interesting and competitive. I was somewhat surprised at the low-key nature of the Vista release, I don't know if you are seeing a huge amount of publicity of the same, virtually nothing yet.

Q: What are your feelings towards Microsoft?

A: It is difficult to have a simple opinion about an organization of about 75,000 people, they are always going to be individuals there with very bad ideas and other individuals with very good ideas.

I think that they should be credited with competing with software developers, and if we go back to the days before Microsoft, software was enormously expensive because everybody customized their product. [They] turned it into a real commodity for computing and for software development, and at the same time they are a convicted monopoly.

I think there is an interesting thing that Microsoft is going to learn and that is that maybe in the eighties and nineties, the most efficient way to produce software was to hire the smartest guys in the campuses far away. But maybe in the year 2000 and beyond, the most efficient way to produce software is allowing people to gravitate to the parts of a software environment that they are most interested in and then to choose to collaborate in real time from wherever they want to be in the world.

Which is why I am so interested in 2007, potentially, as the year in which Linux innovation on the desktop starts to up-shine the innovation of Apple and Microsoft.  You've got to see that getting brilliant people scratching their niches from anywhere in the world collaborating on the Internet is not just a great way to make cheap software, it is a great way to way to make a phenomenal, surprisingly good, and unexpected software, breakthrough software.  If you look at the great companies that have been produced over the last ten years in the technology sector, it is hard not to see that the vast majority of them have found deep relationship with open source, such as Google, Yahoo, Ebay.

[These companies] and others are fundamentally driven by open source and they were built by people who were empowered to do what they could do with open source.

Q: What is the way for Ubuntu to expand the use of Linux on the desktop? Do you think working with proprietary companies to port their codecs to Linux for multimedia use such as video and audio is necessary?

A: Sure, anything like that helps the extent that software people are familiar on Windows or on the Mac is also available on Linux that we have and that can cut both ways. One thing is to go to the proprietary software companies and get them to port to Linux, but the other way to do it is to take open source software and make that available on Windows and on the Mac. 

Having said that, this is by far the most complicated thing I have ever been part of. I often sit here and wonder, will the world switch even if we can produce a platform that is more exciting and more robust and more distinctively virus-free, that is unlikely to get spyware and completely made every single measure better. Will the world switch? I don't know.

Q: Is it more complicated than, for example, going into space?

A: No, I had very specific responsibilities in the crew once I was certified and trained for those responsibilities. In this case and in many instances what we tried to do has never been done before. We are trying to change the way people think about the economics of software fundamentally, not just substituting one product, a $99.99 product with another product that is $49.99, to change the way people think about the economics of software and change the habit that the people have with computers. So today I am privileged to be a part of it.

Q: What is your next big challenge other than open source and Linux?

A: I don't have a list. I am not going to be multitasking, I would like to go deep into one problem and [get a] conclusion. I love this project, so I hang out with many interesting people and I think we are changing the world, so I don't daydream a lot about other things that I am not doing at that time. When the Linux project is done, whichever way it works out, what I'll do is take a look at the state of the world and see what problems are interesting at that time. Everything in life is contextual partly in place and partly in time, what's interesting today is not going to be interesting when I have the time to see it.