Home Page



May- or Jun-2000
Post Database Year 11 Special

A Personal Journey into Linux
By Thiravudh Khoman

As a computer professional, I've always had an interest in Unix. I've used many types of computers in the past - a few mainframes, a few minicomputers, and practically all of the major personal computer operating systems from the Apple II onwards. But never one that ran Unix.

The Epiphany

Not long after I got my PC, I found a program called "MKS Toolkit" from a Canadian company called Mortice Kern Systems (https://www.mks.com) which brought me a step closer. MKS Toolkit gave me the Unix command line utilities and a Unix shell for my PC - a very decent one - but that was it. It was a DOS world back then (Windows hadn't reared its head yet), and after playing around with MKS for a spell, I asked myself a cold, hard question: Why did I need or want this when I could do much the same thing with DOS and the odd utility? Obviously, what was missing was the "killer app".

Several years later, after chatting with a friend about Minix (Linux's predecessor), I started hearing about another implementation of Unix for the PC. At that time, companies like QNX and SCO already had Unixes for low-end Intel 286/386's, but since this was basically an intellectual exercise for me, I wasn't ready to pay the asking price. Besides, this new Unix for the PC - Linux it was called - was free. Eventually, I tracked down an early copy of Slackware Linux as well as something called Yggdrasil Linux. Unfortunately, the Linux documentation back then was sparse and terribly technical (for me at least), and I didn't get very far. Nonetheless, I was intrigued by what I saw and made a mental note to keep an eye on Linux in the future.

My breakthrough with Linux finally came about a year ago. I had just completed 15 years of IT work at a finance company and was looking for a low-cost network operating system for home use and some long delayed Internet R&D.; The obvious choices were Novell Netware which I had used for the past 5 years and Windows NT, but again, I didn't care to pay 10's of thousands of Baht for such software. As a proponent of public domain software and in light of Thailand's financial crisis, I wanted something legal and cheap, partly to save money, but also because I eventually wanted to impart what I learned to some IT compatriots. As a matter of principle, therefore, the Panthip route was out.

By that time, I was well into reading Linux magazines to prepare myself ("trying to read" is a better description). It was one particular article on Netware emulators that turned the proverbial light bulb on in my head. Wow, not only could I run my old friend Netware, but I would get Linux as a bonus. There were in fact several such emulators, but the one I fancied most was something called "Netware for Linux" by Caldera Systems (https://www.caldera.com). This was REAL Netware, licensed from Novell and ported to Linux, and therefore, was as compatible as any emulator could ever hope to be. What's more, a 2-user license was free for use. You could even download it, but unfortunately the software was over 100mb in size and required Caldera's own OpenLinux v1.x distribution to run it. I had found the killer app that would get me started on Linux.

That Netware for Linux ("NWL") would finally start me on Linux is ironic because NWL now appears to be a dead-end product. Caldera no longer seems interested in developing the program and hasn't even upgraded it so that it could run on their newer 2.x versions of OpenLinux. Furthermore, while waiting for my NWL CD to arrive, I obtained Red Hat Linux v5.2 locally and discovered "Samba", a program that allows Linux and Windows PCs to share files and printers under the auspicies of Windows networking. In short, I no longer needed NWL - I could network using Samba instead.

Linux Coming to You

The purpose of the above preamble was to suggest that it's possible that Linux could sneak up on you somewhere, somehow, sometime like it did for me, and that the question of whether you should try it may be a moot one. In my opinion, Linux IS for real and odds are, you may have even "tried" it already. How's that? Many websites on the Internet are already running off Linux hosts (look for the "Powered by Linux" banners) and I believe more will be coming online in the future, especially as affordable high-speed links make running servers on your own premises a reality.

Closer to home, more and more household items should be running "Linux Inside" in the near future as well. Two consumer items already available in the U.S. (albeit not in the category of refrigerators or washing machines) are the Tivo home recording system (https://www.tivo.com) and the Empeg car MP3 player (https://www.empeg.com). It's a no-brainer that other Linux-endowed home systems and consumer devices are on the horizon, especially as efforts gain steam to have Linux embedded into devices and even chips. A case in point is the Kerbango internet radio (https://www.kerbango.com) which runs MontaVista's embedded Hard Hat Linux.

And of course, the Linux community's curiosity is at a bursting point wondering what Linux's Linus Torvalds is doing at chip company TransMeta (https://www.transmeta.com). Indeed, we may see those aforementioned fridges and washing machines with Tux the Penguin stickers on them sooner than we think. Hey, if my National water heater can have an elephant on it, why not a Penguin stumping for Japanese fuzzy logic air-conditioners?

Linux on Your Computer

Of course, I'm skirting the issue here. No doubt, the focus of the question, "Should I be using Linux?" refers to what we see on the screens of our PC's. And in fact, the question begs the follow-on "... instead of Windows".

For people who'd prefer not to learn another style of computing, who aren't ready to forego the current wealth of Windows software, or who aren't excessively dissatisfied with Windows, switching to Linux may feel like an ordeal. Such people are probably Windows 9x users and I doubt if Linux (at least at this time) has any overriding advantages over Windows in meeting their needs. Linux is a different animal from Windows and while recent developments make conversion from Windows easier, no Linux developer is going to use Windows as an exact role model, and therefore, some adjustments in modus operandi will be inevitable.

To Windows users, Linux is probably not unlike living in a foreign country and conversing in a foreign tongue. Certain people are fascinated by this challenge, while others just can't cope without their native language, their native friends, or even their native foods. Of course, there's nothing inherently wrong this. Some people can and some people can't. People who are reasonably satisfied with Windows, shouldn't consider changing just because someone ballyhoo's Linux as the flavour of the month. One switches becauses it has something better to offer (the "pull effect") or one is dissatisfied with Windows (the "push effect").

On the other hand, for people who find learning different ways of doing things both a joy and a challenge, who find the Windows experience exasperating, who wish to experience the full range of computing applications without financial impediments, or who are fascinated riding the crest of a new wave (those who witnessed the birth of the earliest CP/M computers, the Apple II, the Macintosh and the PC will know what I mean), Linux has much to offer, especially for those who are willing to grow with it. Certain people, it seems, embrace Linux because it's the antithesis of Bill Gates and everything that's Microsoft. Personally, I'm not motivated by such negativities, and I believe it's the wrong reason to use Linux.

In my opinion, Linux's strength lies in its "free" nature and its server-side capabilities, especially its close tie-in with the basic internet applications. Naturally, other platforms (e.g. Windows NT, Sun, etc.) are also well geared to run internet apps, but they'd be hard pressed to match Linux's cost-efficiencies. More often than not, software for Linux can be obtained for free or at little cost; and not just hobbyist utilities but serious business-building apps. On the hardware front, it's generally true that Linux's hardware needs are more modest than the likes of Windows NT. A low-end Windows 9x PC these days (Celeron CPU, 32mb RAM) can comfortably run server apps, while older Pentium I's and even 486's with 16mb RAM could still eke out a living serving text-only apps, something next to impossible with Windows NT.

Of course, these aren't the features that are going to light the fires of casual Windows users. Given that Linux is first and foremost a "network operating system" (NOS) rather than a personal one, this shouldn't be surprising. Rather, the primary interest group should be technically-oriented people like system administrators and developers who are provided a functional alternative to the mainstream NOS'es, one which costs nothing to try or to implement. If my finance company were still around today, I have little doubt that I be running Linux by now, although it wouldn't necessarily replace my Netware file and print servers (at least not yet). Running Linux as an intranet web server, as a mail server, as a caching proxy server, as a backup server - these are all jobs well suited to Linux, and at nearly zero cost!

Running Windows and Linux

So far, I've discussed Windows and Linux as "either-or" propositions. In truth, recent developments allow you to have your cake and eat it too. The easiest way to try Linux while keeping Windows in ascendancy is to install a Linux "distribution" that lives happily on a DOS/Windows FAT partition. Two such distributions come to mind: Phat Linux (https://www.phatlinux.com) and WinLinux (https://www.winlinux.com). Both appear as nothing more than another program in Windows' space and has no effect whatsoever on your existing Windows setup. However, neither actually run under Windows - they both require a reboot out of Windows after which a DOS loader is used to load Linux.

While both Phat Linux and WinLinux give you a taste of Linux - real Linux mind you, not an emulation under Windows - they're both somewhat truncated, a fact that you'll discover if you dig deeper. To install full-blown Linux so that it co-exists with Windows is a bit trickier, but still eminently do-able. This involves stealing some free space from Windows by using a re-partitioning tool (e.g. PowerQuest's Partition Magic), and then installing Linux into that free space using its own native ext2 filing system. With two operating systems on your hard disk, you'll need a "boot manager" to allow you to select which O/S to boot into. Linux comes with something called "lilo" for this, but you may prefer the likes of PowerQuest's BootMagic instead (it's prettier). Caldera's OpenLinux v2.x (https://www.caldera.com) is probably a good starting point since it bundles both Partition Magic and BootMagic, not to mention a nice, comforting GUI installation procedure.

Keep in mind, by the way, that this setup is for "trying" Linux only. Serious implementations of Linux are always installed onto dedicated "boxes" (Unix vernacular for "computers"), and in fact really "Pro" implementations segregate each application onto their own dedicated boxes! You may also wish to know that many Linux die-hards would rather drink hemlock than share a hard disk with Windows!

Is Linux Good Enough?

An on-going argument between the pro- and anti-Linux camps centers on how well Linux performs compared to the mainstream NOS'es (most notably Windows NT). I prefer not to get into a number slinging match, but AM of the opinion that Linux is most certainly "good enough" (or "por piang" in post-IMF speak), and would like to offer the following observations:

  • Technical superiority alone isn't necessarily the deciding factor. Looking back into history, one finds that most if not all of the PC "super-servers" of yesteryear have all bitten the dust, while non-specialized PC providers like Compaq, HP and Dell have conquered the low-to-high end server market. I believe they did this by gradually incorporating features of the super-servers and making them available at affordable prices.
  • By being forced into showdowns with Windows NT, Linux is addressing its so-called "enterprise" computing weaknesses and is gradually supporting features of the super-servers and making them available at affordable or zero cost. Of course, they're doing this from the software end, not hardware.
  • One area of enterprise computing that Linux is ahead of the curve is in the area of "clustering". If you'd like to see how a nearly off-the-shelf Red Hat Linux can power a "somewhat" powerful computer system, check out: https://www.lwn.net/2000/features/FSLCluster/. This article talks about the NOAA's (National Oceanic and Atmospheric Administration) Forecast Systems Laboratory which runs a Linux cluster comprising 276 stock DEC Alpha's (667Mhz Alpha CPU's with 512mb RAM each). And this is only Phase One. Phase Two and Three will grow the system even larger.
  • Another example of a "serious" Linux roll-out is the well-known Google search engine (https://www.google.com). Google uses over 4,000 "white box" (i.e. generic, "no name") servers running RedHat Linux to perform searches. You can read more about this at: https://www.techweb.com/wire/story/TWB20000530S0011.

Linux and Thailand

In my opinion, Linux represents a strategic opportunity for Thailand, especially now in the aftermath of the financial crisis. This importance is rooted in Linux's status as "free software". Granted, the politically correct interpretation of "free" is in terms of source code availability and not in terms of cost, but in truth, both meanings are important here.

Many years ago, U.S. interests argued that local software development would blossom if only Thailand would pass a software copyright law which they had vetted (as if Thailand had a choice?). While this may be true to a certain extent, the challenges of creating a major work like an alternative operating system or office productivity tools from scratch are formidable. Inevitably, Thailand (and practically every other country as well) often has little choice but to buy into the likes of Microsoft Windows and Microsoft Office.

The significance of Linux is that it provides Thai developers a headstart, a starting point far above ground zero. Thai programmers don't have to re-invent the wheel with Linux; all they have to do is to join global efforts in localizing the Linux apps already part of the distribution as well as contributing apps of their own making. This is the crux of the Thai Linux Working Group (https://www.linux.thai.net), as well as more commercial ventures such as Kaiwal Software's (https://www.kaiwal.com) KW Linux.

Meanwhile, the free availability of Linux has made possible the implementation of various Linux-based projects with fewer financial hindrances. An excellent case in point is the SchoolNet project (https://www.school.net.th) which might have languished for want of funding if commercial software was the only alternative. I was going to suggest that institutions of higher learning embrace Linux, but this is clearly ignorance and naiveté on my part - most universities are already well down that path, a fact I'm happy to report.

It should be noted that despite my pro-Linux standpoint, I'm not too gung-ho about converting people to using it. My reasons are simple - I don't see it happening in the current generation of users, at least not to the scale I think it deserves. Windows is just too entrenched. But I do have high hopes for the next generation, one which I hope has been well-schooled in Linux. Assuming Windows is still around then, the next generation should be proficient in both and will be able to pick and choose the best features of each.

Perhaps this is the silver lining of the financial crisis: being less able to afford the "brand name" stuff, we're forced to pay serious attention to alternatives like Linux, and in doing so, discover an alternative that breaks us of costly software habits and the tyranny of upgrade cycles. We'll see.

Copyright © 2000, Thiravudh Khoman