Linux heavies play version name game
Linus Torvalds and other central Linux programmers are debating whether the next version of the OS should be numbered 2.6, or if its new features merit the grander designation 3.0.
- Shankland covered the tech industry for more than 25 years and was a science writer for five years before that. He has deep expertise in microprocessors, digital photography, computer hardware and software, internet standards, web technology, and more.
The debate began last week with a post to the Linux kernel mailing list by programmer Jeff Garzik. Amid a discussion of some storage-system features in the kernel, or core software, of Linux, Garzik asked, "Is it definitely to be named 2.6?"
Torvalds, the founder of the Linux kernel effort and still its leader, replied that he saw "no real reason to call it 3.0," saying there aren't enough significant changes to warrant such a major new designation. Later, though, Torvalds opened the door for the loftier label.
Version numbers, while in one sense mere labels, carry a message about the magnitude of improvements and changes from one version of software to the next. For that reason, corporate marketing departments often have a hand in choosing them. Sun Microsystems, for example, took its Solaris operating system from 2.5 to 2.6, but when it got to 2.7, it changed its naming convention, labeling the product Solaris 7 to signal it was an important new release.
In the Linux debate, Red Hat programmer Ingo Mollnar disagreed with Torvalds' assessment. Mollnar said changes to how Linux uses hard disks as extra memory--the "virtual memory" subsystem--and to how its input-output system moves data are some of the biggest Linux tweaks in the last five years.
Torvalds responded that he likes the virtual memory system changes but is worried they haven't been tested well enough by mainstream users. But, he concluded, "If Ingo is right, I'll do the 3.0.x thing."
Torvalds himself has used version-numbering sleight of hand to try to coax Linux development. For example, he applied a naming convention normally reserved for final versions of the OS to a test version in order to encourage developers to up the ante on their experiments.
But in the present debate, Torvalds seems to be trying to keep people focused on the technology rather than the naming scheme.
"It's just a number," Torvalds wrote. "I don't feel that strongly either way. I think version number inflation...is a bit silly."
"Anyway," Torvalds wrote later, "people who are having (virtual memory) trouble with the current (test version), please complain, and tell what your workload is. Don't sit silent and make us think we're good to go."