Why Linux

Here are some initial thoughts:

First and foremost: it's great for hobbyists: we get to peek at a kernel ... and experiment!
it raises the bar for other unix platforms (commercial and free)
it provides some pressure on microsoft
it's a great server platform
There is no chance that Linux will ever be unsupported.
There is no chance that Linux can ever be used to lock people in to a particular vendor.
it looks pretty good for embedded systems and appliance sorts of things. (Embedded systems and server appliances (like the Qube) need to be very inexpensive. Paying a hundred bucks a pop for the operating system is not a competitive move. That's why Tivo is built on Linux. More than servers, and firewalls though, stuff like a "web tablet" might be another place that we will see stuff built on Linux. Unfortunately, I'm not very involved in the embedded systems world. For all I know, Linux might be running on my digital watch.)
For the same reasons that Linux makes sense in the embedded systems space, Linux makes sense on low-end desktop machines. The $500 PC systems you can buy could be $420 systems with Linux instead of windows. If low-end desktop machines fall below the magic "I don't have to ask my wife before I buy" mark (which I believe is about $300) there won't be much margin above the hardware for shelling out to the OS vendor.
There are at least very good long-term prospects for OpenSource OS software, possibly better than proprietary OSes. Linux is already great for hobbists and will continue to be. Of the "niche" markets that have adopted GNU/Linux, some are quite large, others are small, some are growing. What might give OpenSource OS software better prospects for the future than comercial software would be if large companies find Linux (or something similar) to be a better value for their desktop systems, as well as thier servers.

But WhyNotFreeBsd??

Some aspects (big and small) of Linux are probably interesting in some way to a small chunk of people, but not an issue that is of interest to the vast majority of people making OS choices for themselves and/or others.

Some aspects of Linux (server platform, impact on other unices) make it good for niche markets, but again not likely to impact most people making choices.

The quality of new MicroSoft OSes will cut off what few advantages linux had.

There is not, as yet, any evidence of that. You are also implicitly making the mistake of ignoring non PeeCee hardware.

Outside of Zambia, if you suggest to anyone that that measly $80 difference is worth getting Linux instead of Windows, they'll chase you up a tree and set fire to it. :-)

Not sure what trees have to do with a sensible discussion of this issue. Apparently you are stuck thinking about home-user market. Try buying site-licences for 3,000 people some time. The advantages of linux (and I will be the first to admit they are not universal) are not in the initial price tag. Although the TCO is very good, and can be a factor.

Until you can adopt into your worldview an honest understanding of why paying $80 extra for Windows is not only understandable, but, in the vast majority of cases, a common-sense no-brainer, I wouldn't make too many predictions about the future of the OS market. :-)

It appears you have little grasp of what the vast majority of cases is. I fully agree that for joe 6-pack, windows is a no-brainer. Also for certain office prod. uses, as long as the server requirements are not too complex. For many other uses, Unix has some serious benifits, and Linux is a very good Unix for some spaces (of course it has it's warts too).

There is no single reason why everybody should use Linux. If there were, Linux would just be another religion, and who needs that? Several of my friends run OpenBsd and some even run FreeBsd. So there is plenty of choice, and you have to find out what works best for you. I got stuck with Linux in 1992 when there was no free BSD for the affordable PC platform, and I haven't had reason to change since. --LarsAronsson

Outside of Zambia, if you suggest to anyone that that measly $80 difference is worth getting Linux instead of Windows, they'll chase you up a tree and set fire to it. :-)

The problem with the above statement is what, so conveniently, LinuxIsTheAnswer? people always neglect to mention:

Linux will save you $80 or whatever Windows costs these days.
Linux will not save you x hundred pounds on the price of a computer.
If, therefore, we say that computers are being donated to countries which cannot afford to purchase them; will the operating system not also be donated?

Does anyone have any experience of the above which may change what I said?

-- MatthewTheobalds

Yes, any intel-like computers donated to Zambia would almost certainly come with a free copy of MsWindows? ... but that hardly seems the point. I believe the point that was being made was that in the post-industrialized countries, we won't care about an $80 difference in our purchase.

This is a straw man that often comes up. The $80 is nearly irrelevant. A big part of the minority world (e.g. NA/western europe etc.) shrink-wrap software market relies on planned and enforced obsolesence (and lock-in). These hypothetical people in Zambia aren't going to be ordering a new version of Office for those machines, nor are they going to be playing the upgrade game with any of the other pieces of software they might need. So this leaves them a few options

Pirate current copies of software (which probably won't run on the donated hardware)
Try to manage with old versions of the software, which may be donated with the machine as well (but if there is a bug or feature lacking, they are completely SOL)
Use and operating system and tools that are libre, and allow them to make bugfixes and system modifications.

Despite microsoft's opinions to the contrary, even in NA my experience has been that the TCO of linux or *BSD is lower than win32. In these countries, where the cost of sysadmin labour is very low, it is a no-brainer. Granted, some places don't have the technological base to bootstrap this sort of system nation-wide. But they do have smart people willing to learn and work their butts off - so it isn't that large a hurdle. Besides, in a scenario like that you aren't much better off with win32 boxes, you still have to train a core number of people.

Please consider the lifetime costs of Windows versus Linux. With WIndows you get almost yearly upgrades and your are forced to change your hardware platform every 3 - 4 years. We had some UNIX servers (on 386 machines) running a proprietary application for 10 years. About 2 years ago, we began replacing the servers with LINUX (on currently available PC hardware) due to hardware failures on the original units. We expect the LINUX boxes to run for another 10 years. Know anyone who ran Windows 3.1 for 10 years, did a relativley painless upgrade to Windows2000, and expects to do no upgrades for 10 years?

Indeed. I am often found saying: "If this machine were 12 years old, I couldn't write this any faster." Then I point out that this machine is indeed, 12 years old, running NetBsd / mac68k (the latest version, of course). Thank goodness (and the developers) for NetBsd. -- MatthewTheobalds

Where Linux really would have made a difference: A certain public body I know has over 60 workstations available for casual use. These are all running Windows NT. There is one machine per person. Each machine runs only a web browser and a stripped-down, custom written email client.

When asked why do you have a box-per person for such a light load, the answer is "because they're set-up in racks" and we can hot-swap the entire system units if they crash.

Since they don't need any flashy office applications I'd say that instead of 60+ machines running NT like this, they might have been better off with 30 or fewer machines running Linux. Maybe they could have dumped the email client completely and used a cluster of machines doing webmail.

They could skip the NT license fees entirely, and spend a fraction of it training a few staff to support linux. That's a cost that's comparatively low compared to all these server licences. Given the level of customisation that you can achieve with Linux , they could probably tune the whole set-up to provide an even thinner, more robust client. --

Even if the Linux is great in the above mentioned Zambia, the people of Europe and North America could probably care less. However, the lifetime cost of an OS might make a difference to more than a few companies, that remains to be seen.

CategoryUnix CategoryLinux UnixOperatingSystems CategoryOperatingSystem