I just reread the announcement of a partnership between IBM and RedHat in 1999 to promote GNU/Linux to the enterprise. Here’s a quote:

Continue Reading

Remember the olden days when folks used that other OS for the applications? We were told GNU/Linux would never make it because it lacked applications. GNU/Linux is making it and there are a lot fewer applications lacking.

I have been using Debian GNU/Linux lately with 28000 packages of software. That’s an amazing number of packages you can install in seconds anywhere in the world on many hardware platforms but it’s a drop in the bucket compared to what’s available under a Free Software licence.

SourceForge and GitHub host about 1.75 million software projects with about 3 million contributors. Collectively, those people deliver far more code than M$ with its few thousand developers. M$ delivers bloat. What word could be used to describe this much Free Software? Cornucopia? Abundance? Wealth?

It’s no wonder I and many millions of people are happy with the number and kinds of applications for GNU/Linux. The world can create its own software. The world does not need monopolists to produce software. Indeed, monopolists in software are endangered species.

Exploring GitHub gives (repositories, any language):

• 37 found for “search engine”
• 91 found for “pascal”
• 16 found for “mortgage”
• 356 found for “CRM”

Using Google to explore (site:github.com) finds:

• 276 hits for “double-entry”
• 21 hits for “ballistics”
• 540 for “ERP”

No wonder Android is thriving. It’s just the tip of the iceberg. The rate of growth of availability of Free Software is phenomenal. In just a few years SourceForge has multiplied and GitHub came into existence.

UPDATE Found a gem on GitHub. It’s not an application but an important document, a transcript of an interview with Donald Knuth, one of the GodFathers of computer science. He started out in the days of hard-wired decimal machines and punched cards a decade before I touched a computer. There’s a lot of sharing of code in the old days and that is thriving today. This article really underscores the idea that one builds software from algorithms and refinements of others’ ideas. Software is not a creative art but a conversation with computers and people .

I finally got around to testing KVM. I usually use VirtualBox. The test was very easy. Everything worked. The only real difficulty I had was getting an .iso to boot. It turns out there is a “connect” button I did not notice far to the right in the configuration widget. My eyes? User interface? I don’t know but it turns out to be every bit as easy to use the GUI for KVM as VirtualBox. I even ran KVM inside VirtualBox. Everything worked.

Then I loaded the driver (avoiding a reboot) and ran virt-manager and it was very similar to VirtualBox from there.

From now on I will prefer KVM over VirtualBox thanks to the shenanigans of Oracle v Google etc. If KVM is solid enough for RedHat, it is solid enough for me.
“KVM is included in the mainline linux kernel since 2.6.20 and is stable and fast for most workloads.” There are technical reasons to use KVM as well. Now being in the Linux kernel means it should be continually available. I have no confidence that Oracle will be smooth with VirtualBox. Parts of VB (“extensions”) are already non-free (“free for personal use”…).

see TheVarGuy for a review/comparison earlier. My first experience with KVM was much better than his but that may be because KVM was evolving rapidly back then. Phoronix did some benchmarking with peculiar results (some virtualization was faster than native… clock error? caching?) but it is clear that KVM and VB give similar performance in many cases.