Columns

Is Linux the answer?

• By Michael W. Bucken
• June 18, 2001

Once again, we're nearing the time when we won't be dependent on a single operating system supplier. Linux is now taking its place as the next great corporate operating system. The seven-year-old Unix derivative, available freely over the Internet, is attracting the attention of industry pundits. A primary argument from many of these backers is that Linux is a potential alternative to Microsoft. But to date, its path to the heart of corporate managers sounds eerily familiar to earlier failed efforts to create the consummate I/S operating system.

Just during the last decade, experts have crowed about IBM's plan to join its myriad operating systems under the Systems Application Architecture (SAA) umbrella. High hopes also followed attempts by the Open Software Foundation (and Unix International) to merge multiple Unix implementations. Today, SAA is just a memory and IBM still sells computers running multiple, incompatible operating systems. Meanwhile, UI is long gone and the OSF has forsaken operating system research as part of The Open Group. Also, top computer makers like IBM, Sun and Hewlett-Packard still sell proprietary Unix software.

Over the last few months, Linux has been getting the attention of the (trade and general) press and analyst community. Two top venture capital firms, as well as Intel Corp. and Netscape Communications, have invested significant sums in a Linux distributor, Red Hat Software. Informix, Oracle, Sybase, IBM and Computer Associates are porting databases to the software. Even Microsoft is using Linux as an example of the small investment needed to get into the operating system business during its anti-trust trial in Washington.

What does this all mean? So far, not much. The source code is free, but the software lacks the strong support needed by corporate users. And while the database suppliers have jumped on board, the makers of top development tools and applications software remain on the sidelines. Linux may some- day become a widespread corporate computing platform, but it won't happen for many years to come. In the meantime, remember the past. Promise doesn't usually become reality -- especially in this business.

In this issue of ADT, we take a close look at the swelling need for automated testing as corporations install complex packaged applications. Our Special Report on testing also examines potential solutions to the complex problem of managing event-driven, distributed applications across diverse environments.