See also: last week's Linux History page.

This week in Linux history

Former Debian leader Bruce Perens raised a heated discussion in the Debian community when he announced plans to build a new Linux distribution, based on Red Hat Linux. Bruce never built that new distribution, at least not directly. The Linux Standard Base, initially headed by Bruce, now has the reference implementation he wanted. He then later accepted a position as CEO of the Linux Capital Group, which funded Ian Murdock's Progeny Linux Systems, which created Progeny Debian, yet another new distribution.

In any case, Bruce's set of goals for his dream distribution remain interesting.

A nifty new search engine, then known as "google.stanford.edu," hit the net.

According to this ZDNet article two big database providers had no plans to port to Linux.

Database makers Oracle Corp. and Informix Software Inc., whose wares near the top of the Linux community's wish-list, both say they have no plans to support Linux. "We do 16 ports right now, and I could easily add another one. But we don't want to do it unless it makes sense," says Mike Saranga, Informix's senior VP of product management and development.

I guess it finally made sense because database products are now available for Linux from Oracle, Informix, SAP AG and others.

Two years ago (May 6, 1999 LWN): The Mindcraft III benchmarking effort was announced. The results of Mindcraft II had never been released. This time Mindcraft invited Linux experts to tune the Linux system used in the test. In this round of testing some performance issues were found in the Linux kernel; the exposure helped get these issues fixed. Also, countering the claims that Linux tuning information was not readily available, two new sites, TuneLinux.com and Linux Performance Tuning announced their existence. TuneLinux has not been updated in over a year, but the Linux Performance Tuning page is still active.

It is also worthwhile to take a look at Dan Kegel's site which reports, "As of February 2001, performance on the SAP database benchmark on a 4 CPU machine is dramatically better with 2.4 compared to 2.2."

Here's how VA Linux systems looked two years ago, according to Forbes:

Duplicating a Dell-type direct sales model, VA Research is solidly profitable, with a net margin of more than 10%. [Larry] Augustin estimates that revenues will double every quarter for the next two years. VA is shooting for $1.5 billion-plus in sales during the next five years. So far, for the first quarter 1999, Augustin has kept his promise: Sales are up 300% sequentially.

Development of a USB subsystem for the Linux kernel got a bit of a jump start when Linus decided to ignore the existing Linux USB efforts, and tossed a completely new implementation of his own making into the 2.2.7 kernel.

One year ago (May 5, 2000 LWN): The media wanted to know where the 2.4 kernel was. There were several articles remarking on slipped schedules (as if an actual schedule ever existed). This ZDNet article quotes Linus Torvalds:

"We didn't much have a timetable for 2.4 originally, except that everybody knew that the two and a half years between 2.0 and 2.2 was too painful," Torvalds said. "The original hope was to have a release schedule between nine and 12 months, which everybody thought was wonderful, but at the same time a lot of people wondered about how it would work with a minimum three-month testing cycle. Right now it's been about 15 months since 2.2, and it's almost certainly going to be at least three more months," Torvalds continued. "Oh well. More than I would have liked, but not surprisingly so." Torvalds said a big part of the reason that 2.4 is running behind schedule is the same reason that Windows releases so often run late: Developers always want to add just one more feature.

In reality the 2.4 kernel wasn't released until January of this year, about 23 months after 2.2. Well the cycle was a bit shorter, only two and a quarter years instead of two and a half. Besides we all know that when it is released is not as important as how well it works and, of course, we like all those cool new features.

Dr. Dobb's Python-URL, a weekly look at postings on comp.lang.python, included Martijn Faassen's story of the "really early days of Python".

Decades passed while Gordon and the timbot did their work. The timbot sent off some surreptitious messages to various people. John McCarthy got a few hints on recursion. Bill Gates was funded by an anonymous investor. (so, they needed the money! what?!) Larry Wall received a pamphlet on postmodern linguistic analysis. And one christmas, Guido van Rossum received a thick envelope containing the complete plans for a working time machine.

Just in case you were wondering...

Linuxcare laid off a substantial portion of its workforce - estimated at about 35%, and canceled its pending IPO. LWN wrote that "Linuxcare has the dubious honor of being the first open source downsizing. One can only wish that it will be the last." Unfortunately, Linuxcare was not the last open source company to downsize, nor was the last such event for Linuxcare itself.

In contrast, VA Linux Systems was on a spending spree. It was in the process of acquiring Andover.Net (but the $60 million cash payment had just been dropped from the deal - VA is probably happy to have that money now), and also announced the acquisition of Precision Insight.

Next: Letters