It is apparent that, in the Macintosh world, the Java version in which an application is implemented affects the size of the potential market for the application. I'm looking for sources of data (preferably published, but I'll take rumors and gut feelings) that would help me quantify this fact.
One thing you can do is release free "demo" versions of your product
compiled for each Java versions and track which ones get downloaded. My
company does this, but it's not OS X release dependent but JVM
dependent. In any case, OS X versions of FastPak (our product) get
downloaded for Java version 1.5 34.3% times more than they do for 1.4,
so on OS X Java 1.4 is still popular. Interestingly, data from Windows
and Linux is much more radical with most people opting for 1.6 WAAAAAY
ahead of 1.4.
For example, it seems that implementing in Java 1.4 would result in a potential market that includes users of OS X 10.2 and greater, while implementing in Java 1.6 would result in a very small market right now.
All of our code is written to be 1.4 compatible and then the code is
recompiled with the other JDKs. If you're looking at other markets
(Windows, Linux) you might very well have it reversed (i.e. the BIG
market would be for 1.6, with smaller shares for 1.4. and 1.5). Never
the less, these are still customers so I'd go after them if I were you.
I would appreciate any information that would suggest how many Mac users might still be using earlier versions of OS X. I'd also like to find information on the PPC/Intel ratio of Mac users, because we have an application that might be of special interest to users who still run Classic applications.
We can't tell, but once again, if you released a demo specified for the
OS version AND the Java version you could collect stats on it.
On the more technical side, are there any significant "gotchas" that developers should watch for when deploying Java 1.4 code on 10.2, 10.3, 10.4, and 10.5?
We had a lot of problems with FastPak on the OS X Versions from 10.2
(and before) with JDK and JVMs 1.4. One is that the application, if not
run as a "free standing" Java application outside the OS X menu bar
wouldn't die even after it was killed (you had to manually click on the
bottom bar and force-kill it....I'm not really sure if that's considered
an OS bug or a Java bug). This is not the "termination by thread
expiration" used in FastPak, it's hanging around even with calls to
System.exit(0) which shuts down the JVM which to me implies it's an OS
problem. We also had problems with static values being shared between
isolated classes, but this problem is obscure for most applications.
Additionally we had some problems with buffering large amounts of data
seeming to get "lost".
All of these problems have long since been fixed by Apple. The last time
I saw any of them was a few years ago...of course I'm sure there are
others. I think others on this list (as well as the bug lists on Apples
stie) might be of more help to you.
Best Wishes,
Bill Wagner
www.scsc-online.com
Thanks.
Gary Ford
Plaid Flannel Software
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Java-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/java-dev/email@hidden
