Web Informant #87, 9 October 1997
In re Sun v. Microsoft: Countersue!

http://www.strom.com/awards/87.html

The big news this week is Sun's lawsuit about the supposed lack of Java implementation in Microsoft's Internet Explorer v4. And while I hold no thrill for more ligitation, perhaps the real action should be a countersuit from Microsoft to Sun. Here to tell you more about this is a view from an Internet Old Hand and manager at an Internet-related software company who prefers to go by the nom de net of "Poor Richard:"

David, I realize that penning an anonymous note leaves you at a disadvantage, but I want to first introduce myself. I have developed exclusively on UNIX since the early '80s. Today, I use two computers: a SPARCstation as my primary workstation and a laptop running Windows 95. I have a fondness for Sun, and no love for Microsoft. I have respect for both, but loyalty towards neither. (More on this later.)

In its filing with the Court, Sun claims that it is acting as steward of Java, but after some careful consideration, Microsoft appears to be doing a far better job than Sun of promoting Java's good name! Indeed, I think that Microsoft might even be a better steward of Java than Sun.

I run a small group of Java developers. They build highly interactive applets that are designed to occupy a few kilobytes on disk. My programmers are each free to use whatever development suits them best, on whatever hardware and operating system platform they prefer. All have chosen to develop using Microsoft's Java tools and under NT. The reason is simple: they get their work done faster with fewer problems. Microsoft's development environment is easier to use and demonstrably faster by a wide margin than Sun's offering. The only competitor in the area is Visual Cafe.

And Microsoft's Java virtual machine in IE is also faster than any other JVM I've run across. Whenever I need to demonstrate our work performing at its peak, I always make sure that I'm using IE: it just runs Java applets faster.

In the lawsuit, Sun argues that IE doesn't pass certain portions of Sun's conformance tests for Java. Sun contends that passing these tests are fundamental to ensure the "write once, run everywhere" essence of Java. I've got news for my good friends at Sun: there are lots of products with the "100% pure Java" mark that behave inconsistently. In my development shop, as much as we like Java (which is a lot), we refer to Java as "write once, debug everywhere".

When we produce an applet, we have an assurance engineer rigorously test the applet on every combination of operating system, browser, and platform available to us (about 15 at present). Variations in implementations of the Java Abstract Windowing Toolkit are sufficiently large that we wrote our own windowing package in Java just to make the behavior consistent. Variations in implementations of Java threads are sufficiently large that we developed our own specialized Java coding methodology to avoid aberrant behavior on different platforms.

To make things even more frustrating, these and other variations can not be tested for in an automated fashion. So our assurance engineer sits in on the entire creative and development process, in an effort to ensure not only faithfulness of the implementation to the creative process, but also to gain confidence that the applet will behave the same on all Java-enabled desktops.

From my perspective, Sun's conformance program is a technical failure. It is shameful that Sun should continue to suggest that conformance tests will eventually achieve this marketing mantra, and then appear to selectively enforce the results of the testing.

It is perhaps impish to suggest that Microsoft is a better steward of Java than Sun. However, in my view from the trenches that is exactly the case, and others may also agree. Of course, for Microsoft to demonstrate its suitability as a steward, it needs to start delivering Java products on other platforms. I'm still waiting for a Solaris version of IE, which until this point has only appeared in press releases.

Here lies the distinction between respect and loyalty in the Internet: both Sun and Microsoft are formidable companies who have made significant contributions to enrich our community -- as a consequence, we respec them; however, loyalty in the Internet is only as good as your last download. Both Microsoft and Sun should take notice of this!

Site keeping and self-promotions dep't

Thanks, "Richard." I appreciate you sharing your thoughts. As usual, I have been busy writing. This week you can see in ComputerWorld my review of IBM's Host on Demand 3270 Java software, entitled, "IBM host gateway tool still trails rival".

And a review of the web server marketplace for CMP's NetInsider.Com entitled "Apache dominates the web server market".

David Strom
david@strom.com
+1 (516) 944-3407
back issues
entire contents copyright 1997 by David Strom, Inc.
Web Informant is ® U.S. Patent and Trademark Office