I’m tired of IBM sticking its cross-platform smiley face on all its marketing hype. After years of pushing the proprietary technologies of the S/38 and the AS/400, IBM began pushing cross-platform strategies a couple of years ago. Fine, I’m in favor of cross- platform technologies. The theory behind them is that vendors develop products that conform to standard APIs. Vendors compete by writing the “best” (i.e., fastest, least error- prone, and most scalable) implementation of that cross-platform API. No proprietary technologies—supposedly, anyway.

IBM is leading the charge for cross-platform APIs; the company has more Java developers than Sun does. Many of us have already bought into the concept of cross- platform strategies, so we naturally start to develop with IBM’s implementation of those cross-platform APIs. But, after committing to IBM’s cross-platform products, you’ll find out that IBM has inserted its own hooks (read: APIs). These hooks may be cross-platform, but they are not cross-vendor or cross-product. Sure, you can use IBM’s products without these vendor-specific features, but IBM makes it all too easy to use its private hooks.

Several of IBM’s Redbooks, for instance, show how to implement cross-platform e-business applications. But those examples use IBM-specific hooks available only with WebSphere. Those private hooks could easily be encapsulated using well-known, simple object-oriented design strategies, but IBM doesn’t mention that.

WebSphere

I recently downloaded the “Home Bank” examples from IBM’s Redbook Developing an e- business Application for the IBM WebSphere Application Server (SG24-5423-00) from an IBM FTP site. I was able to quickly install the examples on Midrange Computing’s AS/400. Home Bank is a really cool Internet application. But when I began to look at the code, IBM’s propriety hooks began to show. First, a good bit of the Java code was hidden. Where was it hidden? In a .dat file. And .dat files are IBM’s VisualAge for Java (VAJ) repositories for all the code for your Java applications. You can’t work with IBM’s examples without VAJ. I have VAJ, so I was able to cull out the Java source code from the
.dat repository file; I exported the classes from VAJ into Symantec’s VisualCafé—just to spite IBM. Anyway, I started digging into the example source only to find that it uses WebSphere-specific servlet classes (read: private hooks). Those classes are available only if you have WebSphere on your machine.

VisualAge for Java

WebSphere Version 3.0 now supports Enterprise Java-Beans (EJB). EJB is a good thing; the bad thing is that to use existing databases with WebSphere’s EJB engine, you need to use utilities available only with the Enterprise Edition of VAJ. Other vendors of Web application servers that support EJB (such as BEA’s WebLogic or Bluestone’s Sapphire/Web) don’t have any IDE requirements.

I recently spoke with a project leader of a major corporation who is pushing the use of cross-platform Java for his company’s e-business applications. He said the first thing he needs to do is get VAJ for his developers, and he knows that means they all need larger machines (VAJ requires lots of disk, lots of memory, and a fast Pentium). “Wait a minute,” I said to him. “You can develop Java applications for the AS/400 using any Java IDE you want. You don’t have to use VAJ. Many Java IDEs run on small to midsize Pentiums, whereas VAJ requires pumped up machines.”

“Oh,” he said. “I thought...” “That’s right,” I said. “That’s what IBM wants you to think.”

Smiley-face Stickers

IBM, after years of receiving Big Blue Bashing, is very happy to point at Microsoft and cry foul. That’s why IBM is smiling: For once, it looks like the good guy. But I’m ripping that smiley face off IBM. Underneath that yellow smiley-face sticker, you’ll see the same ugly IBM of 10 years ago, only now it has no eyebrows; they’re stuck to the smiley-face- sticker. Nevertheless, the IBM face, no matter how ugly, is still smilin’.