You might think that software portability does not affect you: your application software runs fine on your PC today and that's all you use. And that's true, right up until the time you want to consider a new or different system.
Say you're in the market to buy a new desktop system, and a friend shows you the video-editing, music, virus resistance, and digital picture capabilities of his Apple Macintosh. You consider switching to a Mac. Switching GUIs is a no-brainer—GUIs all do essentially the same things, and it only takes a day or two to re-train your fingers. The problem is the apps. You are faced with the “choice” of walking away from your investment in your existing PC-only software, or buying yet another Windows PC that has some compatibility with your previous system. You can easily switch hardware from Dell to H-P or IBM, but there's a software barrier to switching from Windows to something with fewer security problems, like Linux or MacOS. You've been locked in.
When your application programs are written in Java, you can upgrade OS and applications independently. You can try Linux and still use your familiar applications. You can move your existing Java programs to any new system, and carry on using them. This is why software portability matters to home users.
For businesses, the problem is worse and far more expensive. Even if your whole organization has standardized on, say, Microsoft Windows XP, there have been numerous releases over just the last decade and a bit: MS-DOS, Win 3.1, Win3.11, Win95A, Win95B, Win98, 98 SE, ME, NT 3.1, NT3.5, NT3.51, NT4, 2K, multiple service packs and required hot-fixes, XP, and Longhorn on the horizon. These platforms have subtle and different incompatibilities among them. Even applications running on a single platform have limited interoperability. Older versions of Microsoft Office cannot read files produced by default from the latest Microsoft Office, even when the files don't use any of the new features.
This is done deliberately, to force upgrades. If even one person in an office upgrades, everyone has to (or risk being cut off from reading new files).
Software portability is all about “future-proofing” your software investment. Rewrite it in Java, and that's the last port you'll ever need to do. Portability is the Holy Grail of the software industry. It has long been sought, but never before attained. Java brings the industry closer to true software portability than any previous system has.
Software portability has a wonderful side-effect: skills portability for programmers. Many companies today are out-sourcing programming jobs to countries with low direct labor costs. It's a short-term cost-saving that looks good on paper, until you look at the wider implications.
The beneficiaries of jobs exported from the West are Asian countries with poor performance in the annual global survey of corruption (see www.transparency.org). It's a risky long-term bet to move strategic expertise to countries that combine widespread corruption, unproven business privacy and intellectual property laws, with scant free market experience, and no effective environmental or workplace regulation. The bigger picture needs to be thought through when choosing to export jobs and assets from our home economies. As people with a stake in the computer industry, we have a duty to make our views (whatever they are) known to politicians.
Looking at this another way, China (a nation of 1.3 billion) has already standardized on Linux, and India (1 billion people) is reviewing it. Java offers the only viable way to create software targeted at both the West and at emerging markets.
If you're a programmer in the USA or Western Europe affected by jobs moving offshore, Java is a big plus for your career. Employers used to advertise for very specific hardware and OS experience (“must have 2 years of MVS on OS/390”) and ignore other resumes. Today, your Java experience gained on any OS is directly transferable to other hardware and jobs.
Java is in demand by employers. An April 2004 review of one of the US's largest job sites reflected these hiring needs. Microsoft's C# was mentioned in about 1400 postings, while C++ was a requirement in about 4000 postings. That's what you would expect. C++ is much more widespread than C#, and runs on many more computers. But Java was a surprise: more than 6800 postings sought Java skills. More employers wanted Java experience than those who wanted C++ and C# combined. This is just an anecdotal datapoint, but it is consistent with other surveys. The Software Development Times paper reported in December 2003 that nearly three-quarters of enterprise software development managers are using Java and another 11% plan to start in the next year. Investing some of your time in Java is good for your career.
Java portability poses a real threat to Microsoft's monopoly. Software that can run on any operating system has a larger market than software that is limited to Windows only. Over time ISVs will move their products away from Windows-only to Java—unless Java can somehow be spoiled or broken.
It is unfortunate for you, me, and all computer users that Microsoft uses its monopoly to try to undermine Java. At first, Microsoft introduced deliberate incompatibilities into the Java product it licensed from Sun. Microsoft paid $20 million to Sun to settle the resulting court case. In April 2004, Microsoft paid Sun a further $1.9 billion to settle Sun's litigation over other monopoly abuse.
The current Microsoft plan is to push the C# language, which is Microsoft's barely different copy of Java. But the core C# libraries will be only ever be available on Windows (there's an open source effort to duplicate some C# libraries on Linux, but few believe it will lead anywhere). The C# initiative will last only until Redmond wants to push the Next Incompatible Big Thing. Java is shaped by the computer industry as a whole, and will be around until industry reaches consensus that there is something better to replace Java.