READING:
Plainsong
Kent Haruf

Haruf’s book was nominated for a National Book award, but the book – while well done and an interesting read – doesn’t do anything special for me. Weaving together the disparate lives of a half-dozen or so inhabitants of a small agrarian Colorado community, the story never meshes enough to make it compelling.

All books

I devoted an earlier entry to Why Microsoft Stil Matters; this is the flip side I had planned on writing immediately afterwards, but just didn’t get to. Linus won’t mind the delay.

By Linux, I mean the product of Linux Torvalds and his minions. Much of what I say can also apply to Unix products – actual Unix products (AIX, Solaris) or, like Linux, non-certified Unix-like products, which includes MS-DOS (!). The emphasis on on Linux itself, but the generalizations are of Unix-like products, which I will refer to as Unix for the sake of simplicity.

Also, this list will, in many ways, compare/contrast Linux to Microsoft. This is only natural, as MS is the 800-pound gorilla, the yardstick against which one must compare other similar products or processes. But this does not means that it’s Linux or MS – as Linux has taught us, something can come out of nowhere and give any given gorilla a run for its money (figuratively and literally).

Without any further ado, some reasons that Linux matters:

• Software is becoming a commodity: Microsoft has seen this coming, which is one of the reasons it attempts to lock users in with proprietary tools/code. However, like the file-sharing issues that are keeping record-company execs and movie moguls up at night, the reality is that there is a shift away from centralization to ubiquity. This has profound influences on Microsoft’s model, for example, and favors the work of Linux, which is so strongly decentralized. At the same time, this commodity nature means that software must be portable, so it can be embedded in the next widgit that comes along. Linux is perfect for this; MS’s offerings are not.
• Linux has grown up: We’re finally at the stage where even Linus says that this year – 2004 – Linux will start to become desktop ready. Additionally, thanks to the support of companies such as IBM and Novell, Linux has moved from old 386 boxes to multi-processor supercomputers. The support of the traditional companies, such as IBM and HP, also means that it is easier to get Linux into the workplace. Before, the geeks would just sneak it in under the radar: The company intranet was run as a LAMP deployment, but the CEO would swear the company was an all MS house.
• Linux is transparent: You can see and modify the code. Thus, you – a company/devloper – can work with the OS’s hooks to create new apps/tools. With Apple or MS, there is a complex series of contracts, cross-patent licensing and other non-computer work that all but guarantees a lack of product launch.
• Unix is scriptable; MS is not: This is potentially my biggest gripe with MS products: While a limited amount of scripting is possible (batch files, the scheduler and so on), it does not have the robust scripting cababilities of Unix. In an age where computers have come out of the clean rooms and are on every desk doing every imaginable task, a strong toolkit is a time-saver. While MS makes great tools and allows – through easy-to-use GUI tools and wizards – some control, Unix scripting tools (crontab, tar, piping) allow a user with a little bit of experience to automate menial tasks. The best example I can think of is a simple back-up: Ask the average user how to do this on Window. Huh?? On Linux, a small script with tar and zip and move it to a backup directory/machine. Set the script via the crontab and never worry about it again. Also, MS’s scheduling is strange: I have several tasks on SQL Server running every day. Yet they don’t show up in the master scheduler; you have to know that the task is scheduled (and detailed/editable) in SQL Server. Ouch.
• The SCO Fiasco: While there is an obvious strong negative to the SCO Group suing everyone and their mothers over Linux (follow the almost daily lunacy on Groklaw) – it makes the wary shy away from Linux – there are a few strong positives to come out of this (ongoing) mess:
  

  • It’s pulled the software community together. With the exception of Sun, MS and – to a degree – HP, everyone is pulling together on the side of Linux. And much is done without the often-counterproductive Slashdot-type efforts/remarks.
  • No publicity is bad publicity. Witness the surge in SCO’s stock even today over a year ago’s levels, even as SCO is getting discredited. On the other side of the public courtroom, Linux is getting plastered into every news story. Never heard about/know much about Linux before? Now you do, and – guess what? – you’re getting interested in what it can do…
  • SCO would not sue to win … nothing. Much like MS beginning to attack Linux, the SCO Group’s lawsuit legitimizes Linux to a degree: If it wasn’t of any value, why sue? But there is high perceived value. That helps Linux in that respect.

• Linux is extensible: This is a reinforcement of some earlier points, but bears the emphasis. Because Linux is open, it can be easily extended. This means that more products will be made with/for it, and the base product improved in unexpected ways. While this lower barrier to entry will mean that a lot of dross will be created (witness the abandoned projects at Sourceforge.net), it will also open door that – in a MS-centric world – would not otherwise happen. In a MS-centric world, projects have to be approved by committee, with extensive research and so on, so the products are a lock to be winners: Like MS Bob and Clippy…
• Linux is not designed to make money: Yeah, this one drives the Microsofties crazy. But I see Linux much like the Internet: It’s a tool to do other stuff. Imagine if the Internet had been heavily regulated, taxed and so on. We’d never have Amazon, online banking and reservations, blogs and more. Linux works the same way – it’s an open system that allows you to leverage it in the way you want, which just may be for-profit tool/service (think Google and their thousands of Linux boxes..).

Sure, this is an incomplete list – all such lists are. I didn’t even mention how – now that more and more tasks are getting computerized (and potentially exposed to the Internet), security is a real issue, as is stability. Linux has them; MS doesn’t.

And so on.

Pick your tool; make the most of it. Both Linux and MS matter; however, the balance of power is shifting toward – toward, not to – Linux. I don’t see this changing in the near future; I actually expect this trend to accelerate.