Another reason to love open source
During a conversation with Jon today, I added another entry to my list of "rules of thumb for Java programming":
Anything that can be deployed on JRun is going to suck.
Several years ago, JRun was a poor-quality but mysteriously popular J2EE appserver. In the days that WebLogic sold briskly for $10,000 per CPU, JRun was closer to $1,000, so it may have seemed a moderately priced alternative to something you could otherwise have gotten for free with JBoss or Tomcat. At the time it had a well-deserved reputation for being buggy, slow, and only vaguely spec-compliant.
In any case, I was curious about what became of JRun, having not heard that name in the last four years. I knew that the Allaire (the developer) had been bought by Macromedia during the short period when the J2EE container market shrank from 35 (!) vendors to less than 10 serious contenders. I looked it up.
This is what I discovered:
Macromedia (now Adobe) still sells JRun. They still have a shiny product marketing page touting its features and proudly noting "Certified J2EE Compatible. One of the first application servers to be certified J2EE 1.3 compatible, demonstrating the Macromedia commitment to J2EE." There is a Buy Now page which will happily charge you $900/CPU for a license.
This is what they don't point out:
- J2EE 1.3 is obsolete. It was officially released (and supported, at least unofficially, by most vendors) in 2001. As I write this, J2EE 1.5 is nearly final.
- The last press release is from January 2003.
- The last major version was shipped in May 2002.
What Macromedia's marketing won't tell you is that JRun is a dead product. It's likely that most (if not all) of the development team has been disbanded and the product has been put in support-only mode. Buying a copy of JRun is buying into a long-obsolete and already dead product tree, all for the very reasonable price of $900/cpu.
I would argue that this is a kind of fraud.
Development tools are not like products that you buy and use until they wear out. You're investing in a platform that you expect to grow with your business in the future; your purchasing decision is based very heavily on what you expect the future of that platform to be. This is why new platforms have difficulty finding critical mass - nobody wants to mate their product to Homo neanderthalensis. Having myself written a large application in J++ many years ago, I can tell you the fear is visceral.
The fraud that is being perpetrated here is that JRun has a shiny future. It doesn't.
It's not just Macromedia. It's very common for commercial, closed-source companies to sell a product long beyond the time when everyone inside the organization except the sales staff has given up on it. I've worked for companies like that; I suspect only the very lucky haven't.
This is where the transparency of open source shines. When evaluating a new development tool, I don't necessarily look at the source code. I look for recent commit logs and traffic on the developer and user mailing lists. I look at the date of the last release and the frequency of releases. I look at the number of bugs recently opened or closed in the tracking system.
With these metrics, it's almost always trivially easy to tell a living, breathing project from a dead-end. On a closed-source marketing page, it's almost always impossible.
