• Investment World
• eWorld
• Catalyst
• Mentor
• Life
• Canvas
• Praxis
• Urban Pulse
• Brand Quest

Radha Subramanian

It takes little to provoke a `Linux vs Microsoft' debate. This time around, what sparks an exchange is a presentation critical of Linux. So who scores?

LINUX versus Windows is an unremitting source of debate in the IT world. Neither Linux nor Windows advocates will pass up an opportunity to champion their cause. So it should not have come as a surprise when a presentation by Stephen R. Schach of Vanderbilt University highlighting some of the deficiencies in open source software engineering stepped on more than a few toes.

Linux, possibly the biggest name in open source software, has been increasingly accepted by businesses on account of its reported stability, security and low cost of operation. Supporters of the open source movement often believe that the unparalleled success of several open source products is found in the fact that there are so many individuals examining the product. Open source developers are divided into two categories: the core (the small group of individuals who have the privileges to modify the software) and the periphery (the mass users who are theoretically supposed to report problems and recommend fixes).

The question is: Is the periphery actually fulfilling its unofficial role of correctly identifying faults and suggesting fixes, as the model proposes? Furthermore, do members of the periphery have the debugging skills or the motivation to function as effective developers? Schach's research concluded that it is, in fact, the small group of core users who are fixing 70-90 per cent of the faults reported.

Linux supporters beg to differ on this point. Says Mohan Sundaram of Net Devices, a system level design company operating on Linux, "You would be surprised how fragile testing in closed source development is. An ideal testing ground is one where the testing team does not identify itself as owners of the software and hence does not overlook anything. Working for the same company, this is at crossroads to the main purpose of bringing out a product on time. Closed source developers cannot hope to build or employ the mass of varying skill sets and intellect that an open source development innately does. There is only so much one entity can do."

The Microsoft camp disputes the efficacy of periphery developers. Satya of Vortex Technologies Ltd, a Microsoft partner solutions company, says, "People with a passing interest cannot contribute to a quality process even if they are extremely capable. Software is not an art - it is an engineering process."

Schach's preliminary research also shows that it is primarily the core users who are making the enhancement decisions as well in what is supposedly a democratic user system.

Javed Tapia, director of Red Hat India, counters, "The open source community is extremely democratic. The identification of code flaws by a large community is what sets open source apart from proprietary software. The core is who the community trusts to improve the code based on the reputation of the technologist."

The next point brought to the fore was the issue of maintenance. Schach's research concluded that Linux's overuse of global variables and the resulting high instance of common coupling has deleterious effects on maintenance. Because the frequency of common coupling increases with version number, Schach's research deduced that as Linux grows in size, it will become increasingly unmaintainable.

Abhijit Das of Microsoft Corporation foresees maintenance challenges for Linux, particularly in India. "In such a geographically vast country, the availability of technical support from commercial Linux distributors in some of the smaller cities becomes a problem. If the computer operating system is expected to grow with a national organisation, the skilled manpower providing on site maintenance support must also be available. Compare that with the service provided by Microsoft. We have a highly trained team located India-wide to cater to the needs of all of our customers."

The severity of the maintenance issue is exaggerated, according to Mohan Sundaram of Net Devices. "It is true that software reliability decreases with number of lines of code while features increase. But the proof of the pudding is the user base and the success of Linux. We have bright minds that take care of different parts of Linux and I am convinced maintainability of Linux is far lesser an issue than it is made out to be," Sundaram says.

The third unexpected result of Schach's research is with reference to the Lientz Swanson Tompkins (LST) survey on software maintenance (1978). This survey revealed that 60 per cent of maintenance is perfective and only 17 per cent is corrective. Schach's data, however, found that a whopping 40 per cent of maintenance is corrective, more than three times the LST value! Schach claims that this discrepancy is due to the abundance of inaccurate estimates in the LST data.

According to Sundaram, a high instance of corrective maintenance can only be expected with open source software. "The philosophy in a successful open source project is "release fast, release early". Thus we have a concept of unstable and nightly builds in open source. Software released may have bugs, but is released early to allow for testing. The key is response time in fixing bugs. In open source software, this is astonishingly fast for projects that are nascent," says Sundaram.

So what does all this mean in the end? Will Schach's findings leave Microsoft with a rush of converts who have finally "seen the light"? Is the end nearing for open source software? It does not look like that is the case. Linux and Windows fundamentalists treat their operating system choice as a religion. They are steadfast in their beliefs and committed to propagating the "truth" about their cause. Regardless of how much negative publicity they endure, they will continue to preach the virtues of their respective IT doctrines to win the votes of fresh unbiased minds. There are indeed at least two sides to every story, and from the looks of it we will continue to hear them for years to come.

eworld@thehindu.co.in

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
The other way

Copyright © 2004, The Hindu Business Line. Republication or redissemination of the contents of this screen are expressly prohibited without the written consent of The Hindu Business Line