Microsoft: Linux is at the end of its life-cycle. Oh really?

Nikolai Pryanishnikov, president of Microsoft Russia, has reportedly claimed that Linux is at the end of its life-cycle.

I think that comes under the category of, "Well, he would say that, wouldn't he?"

There are a lot of companies that'd envy that kind of "end of life-cycle" product!

Linux works on just about every hardware platform bigger than a PIC, supports a mind-bogglingly large range of hardware, is used on everything from the smallest gadget to the world's most powerful supercomputers, and the average person in the street probably has at least five to ten times as many Linux OS licences as they have of Microsoft products. There are a lot of companies that'd envy that kind of "end of life-cycle" product!

ARM devices are squeezing Microsoft out from the bottom up, on embedded devices, and the top down: the new Cortex-A15 based chips will create Linux-friendly servers that should be hugely more power-efficient than the Intel/Microsoft combination manages - and Microsoft doesn't currently have mainline ARM products (Windows, Office) to defend its space - not yet, anyway. I blogged about the consequences earlier this month.

Pryanishnikov's comments form part of a recent FUD campaign Microsoft's been waging against Linux, and Open Source in general. (I blogged about more of it a month ago.) The campaign says more about Microsoft than Linux and Open Source Software (OSS).

The more Microsoft tries to chop off Open Source's Hydra heads, the bigger the problem they create.

The more frantically Microsoft tries to chop off Open Source's Hydra heads, the bigger the problem they create for themselves. Desperation is hard to hide; increasing desperation, even more so. The solution for MS, of course, is to accept Open Source, concede that MS doesn't have answers in a number of OSS's key markets, and probably never will now, and change the business model to work with OSS, rather than paying lip-service in public, and waving knives behind the scenes. I don't think Microsoft has the ability to make that change, though. It would involve admitting that revenues have peaked, which wouldn't please Wall Street one bit - but better restructure now than face a collapse later, which is where MSFT is heading without some creative thinking that has to happen now.

Ironically, conditions in both customer and stock markets could well mean that MS is closer to end-cycle than its competition.

Friday 13th, three days late

Business as usual on Monday. Well, for the first ten seconds.

I couldn't log into my Linux PC. Which was a bit of an issue, because without it I wasn't going to get a lot of work done. A bit of exploring showed that it couldn't see the authentication server. So I try to log into the PC using a "local" user ID instead (one whose details are stored on the machine, not the auth server). No joy - it was still trying to check the auth server, and the login times out before it gets a reply. Actually, as it turned out, the heat death of the Universe would have happened before it got a reply.

Time to dive into what I jokingly call the machine room to check out the auth server. It's running, but won't respond to keyboard or mouse, and the screen is blank. Hmmm. Nothing for it but a hard reboot. Guess what? It's not booting. Take the cover off, and try again. The hard drive sounds like a miniature jack-hammer; looks like the heads are just ramming their end-stop. The auth server (a recycled older PC) is officially toast. And I don't have a drop-in replacement, not to hand.

It's time to go back to my Linux box, and reboot it into single-user mode (which means I can get into the command line without logging in). Set it up as a temporary auth server, in addition to the hundred-and-one other server tasks it's acquired. Reboot it back into a full running state, and log in.

Joy! I can log in now.

The joy is short-lived, however. I need to install PDFEdit, and I roll in a few security updates too. A quarter-way through, the download stalls.

I try to restart the download several times; no luck.

I reboot the Linux box. No difference.

I restart the local network switch. No difference; I can ping other machines on the same switch, but can't see beyond it.

I restart the powerline modem that feeds the switch. No difference; can't even ping the edge router that feeds the whole network.

I restart the powerline modem connected to the edge router, the other end of the connection. No difference. A laptop that connects to the edge router through WiFi seems OK, so we haven't lost the router.

I restart the edge router, just on the off-chance. Hah! I can see out again! Well, for a moment. About 3MB into the updates download on my Linux box, it stalls again.

There's only one thing left to do, and that's replace the powerline modems. They're both Advent units (PC World own brand; yes, I know, I should know better). I happen to have a couple of Devolo models knocking around and pre-configured, so I swap out the Advents for the Devolos.

It works! It was about time something did - I was beginning to lose the will to live.

This time, the updates and PDFEdit download and install without a glitch.

And then I discover that PDFEdit - the installation of which was the whole point of the exercise - can't even do something as simple as move an item from one point on the page to another.

Definitely One Of Those Days.

Linux - a breakthrough technology?

(From a posting at LinkedIn)

I'm interested to see all the folks who've nominated Linux as their breakthrough technology. The thing is, Linux isn't a breakthrough technology, and never has been. When it was first written, it was already outmoded. It was designed only for a limited range of Intel PC hardware. It had a monolithic kernel that had to be recompiled any time you needed to change the options. The code was OKish, but not particularly smart, and there was nothing there that advanced the science of operating system design - in fact, quite the opposite. It was a college project.

What it had in its favour was one thing only: it was free, devised as a free alternative to Andy Tannenbaum's Minix.

And "free" is the breakthrough technology here. Even this wasn't unique. Richard Stallman, with his GNU project, started in the early 1980s, pioneered development and licensing of software that was completely free to use and modify. With Stallman's GNU toolkit welded to the Linux kernel, there was at last a usable, _free_ operating system, with source code for all to see. That's where I came in, at kernel version 0.12. The whole thing was amazingly robust and non-buggy, even at such an early stage.

In the subsequent development - essentially a full redesign - of the Linux kernel, it acquired platforms as diverse as the Acorn Archimedes and big IBM mainframe iron. It gained the ability to load and unload device drivers dynamically, whilst running, meaning that kernel recompiles became unnecessary. Its very freeness lended it to exploratory projects that led to new products, making it the cause celebre of the embedded systems world. And because it was licensed under the GNU Public Licence, modifications became integrated into the mainline kernel. Despite all that, there's still not all that much in Linux that's leading-edge, other than some funky filesystems.

So if you're tempted to nominate Linux as the technology that changed everything, think again. The leading edge technologies that changed your life were twofold. The first was called Richard Stallman, a man who wanted to change the world, and make software free, with source code visible to all. The second, made at his behest, a technology created by a lawyer: the Gnu Public Licence.

A man and a contract. Who knew?