Java applets...

Recently re-investigating applets, and coming to the conclusion that Sun probably doesn’t want anyone to write them any more - perhaps they’re a bit of an embarassment, giving java a bad rap, getting confused with javascript and all that. If I were a cynical strategy person at Sun that’s probably what I’d do ;D.

I see java still doesn’t work with Mozilla and Firefox (which are rapidly moving towards 30% market share on Windows alone; not exactly a small number of people):

http://plugindoc.mozdev.org/faqs/java.html

“On some systems, JRE 1.4.2 and later do not work. In this case, you can use JRE 1.4.1_07”

Um, no. 1.4.1 is known to be so badly broken as to be unusuable in many areas. 1.4.1 is not an option. Irrespective of the difficulty of getting it running in parallel with a pre-installed higher verison (1.4.2), which no-one wants to explain to anyone. Joe Bloggs hasn’t a chance, really.

“Always make a symbolic link, as shown above, instead of copying the plugin. If you copy the plugin, your browser will crash every time you open a page containing a Java applet. You have been warned!”

WTF? Allegedly, this is Sun’s decision to hard-code the plugin to only search in it’s current directory, and to totally ignore the environment etc (although from personal experience I can imagine that load-ordering issues mean that such data isn’t available at the point when the plugin is started, it’s still possible to workaround). Google a bit and you’ll find that many of the lucky people who can use java as a plugin in Mozilla (note the unexplained note that this is not everyone) do the obvious: they copy the plugin, and immediately cause Mozilla to crash big time on every java-enabled page.

That, surely, is Mozilla’s fault (and IIRC this bug has been around for a year or more) Not exactly a welcome sign to people wanting to try out java applets is it really? :frowning:

Sigh.

funky applet on your entry page!
eye catching even with my Firefox

I did originally write this applet in swing - but it didn’t work too well simply because of a bug in Sun’s JVM where it didn’t initialize swing applets properly (as many as 6 times out of 10 it would assign 0 graphics memory, and all your Swing components would report witdth, hiehgt of -1, -1 (!)). IIRC they fixed this in 1.4.2 (I remember being later asked by Sun to check if 1.4.2 had fixed it).

It was dependent, IIRC, on the timing between particular AWT calls in your source code. i.e. adding or removing no-ops would change the number of times out of 10 that the applet randomly wouldn’t start.

c.f. my comments above on how using 1.4.1 for applets is simply not a realistic option :(…