LINUX BLOG SAFARI

Linux: The Gold Standard of Code

There are few things more gratifying to those of us here in the Linux blogosphere than seeing the many and varied virtues of our favorite operating system get officially recognized.

It happens with increasing regularity these days, of course — after all, there are so very many virtues to consider — but recently an example emerged that has been warming FOSS fans’ hearts ever since.

“Linux code is the ‘benchmark of quality,’ study concludes” is the headline that started the ball rolling down at the blogosphere’s Broken Windows Lounge, where toast after toast has rung out in Linux’s name as a result. ‘The Right Way to Do IT’Linux Girl

Is the Linux desktop a “mess”? Are distros getting too fat? Perish the thought. Linux is the gold standard of code; this week, at least, bloggers have no time for such trivial matters.

It’s time to celebrate Linux’s latest success — not that it comes as any great surprise.

“Of course Linux’s code quality is excellent,” began blogger Robert Pogson, for example. “Developing FLOSS cooperatively among a wide array of businesses, OEMs, retailers, organizations, classes of computers and end-users is the right way to do IT.

“It brings in more points of view, more insight, more knowledge and skill, re-uses good code, and gets more people working in parallel — just what is needed to make great software,” Pogson added.In Other News: ‘Water Is Wet'”Of course, this comes as no surprise to those of us that use Linux day in and day out both personally and professionally,” agreed Linux Rants blogger Mike Stone.

“Further, to those of us that have both experience with Linux personally and professionally and Windows personally and professionally, this study ranks right up there with those that have determined the sky is blue and that water is wet,” Stone added.

“Only the most blind followers of Redmond’s poster child would ever dispute this study, but then I guess there are still those that believe the Earth is flat,” he pointed out.

‘This Should Not Surprise Anyone’

“I think this is something we have been saying for a long time, at least since ‘The Cathedral and the Bazaar,'” concurred Google+ blogger Kevin O’Brien.

“Many eyeballs make bugs shallow, and open code is more perfectible,” O’Brien added. “This should not surprise anyone.”

Indeed, “not news to me, but then again we have long passed the point where most of the interesting development needs to happen in userspace,” observed consultant and Slashdot blogger Gerhard Mack.

‘Thorough and Elaborate’

“Is Linux code the ‘benchmark of quality’? Well, it is very good, no doubt about that,” Google+ blogger Brett Legree told Linux Girl.

Still, Legree chose to take a closer look, downloading the results of the study and then reviewing the document.

“The research performed by Coverity seems to be thorough and fairly elaborate, as the study looked at hundreds of projects from both open source and proprietary code bases,” he explained. “The analysis covered code bases of various sizes, and the results were broken down by project size (lines of code) to compare defect densities between open source and proprietary software.”

The data presented on pages 9 and 10 of the report, however, highlighted some things Legree found particularly interesting.

‘As Good as Anything Else’

“First, the defect density of open source software was on par with proprietary software (per the projects considered by the 2012 study),” he pointed out. “Second, the project size (lines of code) seemed to have different effects on defect density for open source vs. proprietary software.”

Specifically, “small open source projects tended to have lower defect densities than proprietary software, but as the lines of code increased, proprietary software defect density dropped whereas it increased for open source projects,” he noted.

Still, “the main take-away point from the study should be that open source software, including Linux, is on par with proprietary software from a quality perspective,” Legree concluded. “So, Linux code could be considered a benchmark of quality — it is as good as anything else out there.”

‘It’s the Attitude’

Slashdot blogger hairyfeet didn’t see the significance.

“The problem with Linux has NEVER been the quality of the code itself — that has always been fine, top-notch even,” hairyfeet suggested.

Rather, “it’s the attitude,” he opined. “Nobody talks to anybody else, nobody thinks about anybody else, it’s all ‘works for me!’ while ignoring the fact that if it works on the teeny tiny subset of hardware your typical dev has means exactly jack and squat.”

At the end of the day, “nobody other than a handful of geeks care about how pretty or well-written the code is; heck, some of the biggest programs in history were buggy as hell and nobody cared,” hairyfeet concluded. “What they care about is, ‘Will my software/hardware work and will it keep working when I update?’ and nobody seems to care about that part, not a bit.”

‘Happy Future, GNU/Linux’

Not everyone saw it that way, however.

“FLOSS people work hard, not just for a hobby, but to deliver their best,” Google+ blogger Gonzalo Velasco C. told Linux Girl.

“Now, after 20 years of GNU/Linux, with all the modern distributions and the larger number of users (corporate, scientific and home communities), it just became evident,” he concluded. “Happy future, GNU/Linux. You deserve all the compliments.”

Katherine Noyes has been writing from behind Linux Girl's cape since late 2007, but she knows how to be a reporter in real life, too. She's particularly interested in space, science, open source software and geeky things in general. You can also find her on Twitter and Google+.

2 Comments

  • While I understand you may feel the need to add Hairyfeet’s comments in the interests of "Fair and Balanced", which seems to be something very American, that always seems anything but. It needs to be pointed out that his comments are almost invariably a load of horse manure… no I take that back horse manure grows good roses.

    His comment "hairyfeet suggested.

    Rather, "it’s the attitude," he opined. "Nobody talks to anybody else, nobody thinks about anybody else, it’s all ‘works for me!’ while ignoring the fact that if it works on the teeny tiny subset of hardware your typical dev has means exactly jack and squat."" is rubbish.

    I AM NOT a Linux dev by any definition of developer, up until I retired in 2010 I was a Windows developer (for about 20 years, that’s going back to pre Windows 95), before that I was a mainframe developer.

    I’ve used various Linux based operating Systems since 2000 (1999 if you count typing some commands into a SuSE Server for it’s Administrator). Over the years I’ve come to know dozens, maybe several hundred, if you count people I’ve had conversations with about Linux Desktop Systems, who are NOT devs, they are ordinary users, some of those people, who number in the dozens, I’ve set up with one or another Linux based operating system.

    On only 2 or 3 occasions have I ever encountered anything approaching anything close to the issues that Hairyfeet constantly complains about, in all occasions they were on cheap hardware, where the manufacturers provided poor support for Linux, and early on in my time as a Linux Desktop user.

    My personal usage, and that of the people I’ve met and spoken to who use a Linux based operating system represent a very wide range of hardware, probably including all of the OEMs and a reasonable subset of their products.

    So in conclusion, I really have to wonder, does Hairyfeet set out to locate the most Linux unfriendly hardware, or perhaps the most obscure hardware, or does he have some other axe to grind?

    • Riiight, blame the victim. Are you gonna honestly claim that Dell, one of the biggest OEMS ON THE PLANET, don’t know anything about hardware either?

      Go to The Inquirer and look up "Ubuntu ‘broken’ on Dell Inspiron Mini 10" to see how that major OEM had to RUN THEIR OWN REPO because even on the teeny tiny set of hardware they were offering Linux on updates BROKE DRIVERS CONSTANTLY.

      But of course I’m sure you’ll counter with "they should have used Distro X!" with X being any one of the over 400 distros, which is why I no longer do "The hairyfeet challenge" because after picking random bog standard Dell, HP, and eMachine systems off the shelves and showing that time and time again Linux would puke on its own drivers all I would get is "Use distro X!" which if I installed nothing but Linux distros for 8 hours a day would take around 9 years just to go through the ones on distrowatch, of course by the time I got even halfway done I’d get the old "Oh that was the OLD version, the new version of distro X is better!"

      Sorry, you can give your little anecdotes all you want, I can counter with major OEMs having the exact same problems, and if even the OEMs can’t get stable drivers and decent QA, what chance does Joe average have?

Leave a Comment

Please sign in to post or reply to a comment. New users create a free account.

LinuxInsider Channels