IMHO: The Day the Software Died

Commentary
Typography
  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times
Not all that long ago, I could have easily written a very flattering article on using WebSphere and VisualAge for Java (VAJ) to develop powerful Web-based applications for the new low-end iSeries computers. I've been steadily expanding my little set of tools into a fully mature product that allows existing legacy systems to be accessed via graphical interfaces, and WebSphere and VAJ provided a productive and inexpensive way to do that. Run on a very reasonably priced Model 270, this environment had all the makings of a revolution in business application deployment.

Unfortunately, some of IBM's tactics have shaken me to the core. For example, consider these excerpts taken from the IBM Web site last summer:
· "Version 4.0 brings together the functions of Version 3 Standard Edition and Enterprise Edition into a single product." (www-4.ibm.com/software/webservers/appserv/about.html)
· "WebSphere Application Server, V4.0--available beginning June 30, for $8,000 (single server) $12,000 (multi-server)." (www-4.ibm.com/software/webservers/appserv/pr_version4.html)

Subsequently, I received a call from John Quarantello of the iSeries e-business marketing department. John verified these two pieces of information, but wanted to assure me that there would be a free alternative (which has since been revealed to be Apache and Tomcat running natively under OS/400). But this assurance from IBM only came in reaction to the furor caused by the other two statements.

Taken individually, neither quote is particularly earth-shattering, but together they may epitomize the death of software distribution as we know it. Alarmist? Perhaps, but let's review this in a little more detail. IBM gave us WebSphere. IBM told us to use WebSphere. And now IBM is going to start charging for WebSphere? Why would they do that? Call me crazy, but I've been seeing a trend lately.

The trend is toward "application servers," rent-a-program devices similar to ATMs. You download the software, pay a transaction fee of some type, and when you need it again, you pay for it again. As this technology matures, you'll simply send a message to a server and be charged for an answer. A software vendor's dream--getting a royalty every time someone uses the software! And once someone is locked into your server, there's no need to worry about silly things like consumer needs or quality issues. Take a page from the airlines or power companies. Flight cancellations? Cattle car seating? Rolling blackouts? Ridiculous gas prices? No matter! We're the monopoly, and you have no choice!

But there's something slowing this trend--the custom software shop. As long as it's economically feasible to purchase custom software that performs better than off-the-shelf shrink-wrapped "solutions," smart businesses will pay for that software to be developed. Companies run better on software that reflects their particular business rules, rather than someone's generic idea of good business practices. Because of this, the custom software shop has always been the primary developer of good business software, as well as the thorn in the side of those who would shrink-wrap business systems.

So how to get rid of those pesky custom software developers? Here's how: Provide free tools to build software for your proprietary model. Offer the servers for that model free of charge. Make it more expensive to develop programs from scratch and to maintain existing legacy systems than it is to use the pieces from your servers. And once you have locked the developers in, start charging!

See any software packages that fit this particular pattern?

I just keep thinking of razors and razor blades. The razors are no good without blades, and these "Web service applications" are no good without servers. The race now is to become the major provider of servers. And the people using the software will be at the mercy of service providers who will, by definition, be in it only for the highest profit margin.

At least with custom software you have your own version. It encapsulates your way of doing business, and you control what it does. You can change it, rearrange it, and do whatever is necessary to make sure it helps you run your business more efficiently. When you get your software from the Internet spigot, you are stuck with whatever comes out of the pipe--and seeing what passes for quality in software these days, it's plain to me that what comes out of that pipe will be pretty toxic.

I don't mean to focus on IBM here. The other vendors will each push their own version of this model. WebSphere is personally painful to me because I took the bait and spent the last year or so developing tools for the platform with the mistaken notion that it would continue to be a free component of the operating system. But it's becoming clearer to me that the major computer companies no longer want your business--they want you as a captive audience.

Now, lest you think I'm a wild-eyed conspiracy theorist, there's another possible answer: IBM just isn't sure how it wants to handle the Web. Machines that use 5250 screens provide a lot of revenue through the interactive feature, so an option that would provide Web access to legacy programs running in batch would make the iSeries much more attractive, but at the same time, IBM would lose that revenue. So IBM gives us WebFacing, which sort of Web-enables applications but still requires the interactive feature, in addition to WebSphere! Talk about half a solution.

This second theory became more credible when I found out at COMMON in Minneapolis that VAJ is being phased out. A new tool suite, based on the open-source Eclipse development environment, is the new direction. It will incorporate many of the features of VAJ (and new ones besides), but that's not the point. The point is, we were told that WebSphere and VAJ were forever, and now it looks as though it's Apache and Tomcat and whatever the new development suite will be, unless you want to pay the extra money for WebSphere 4.0.

So, is it a calculated plot to lock us into the pricing model that makes the most profit for IBM, or is it a mad scramble to try to develop a rational development environment? In either case, until the dust settles, it's hard to know which route to take to move to the Web, and in the long run, you and I are the ones who are going to pay for it.

And if you don't believe me, then you probably thought that ATMs were going to be free forever too.

Joe Pluta is president of Pluta Brothers Design (www.plutabrothers.com) and the author of eDeployment: The Fastest Path to the Web. You can email Joe at This email address is being protected from spambots. You need JavaScript enabled to view it..

BLOG COMMENTS POWERED BY DISQUS