[GEEK] Java Web Start

You might think it's a bad idea for me to complain about Java Web Start. Well the truth is that I really like it and so am really frustrated by it's failures. In my own small way I am hoping to wake up someone at Sun to see the importance of this little bit of technology to their grand Java strategy.

In my interview with ClientJava I again made some pointed comments about Java Web Start, and in the comment thread some annonymous person (coward) asked me to elaborate. Here is what I said.

"Why has WebStart been such a disappointment? I've written about this in my blog, here and here.
Basically, the promise of WebStart is great: a cross platform way to deploy, install and update Java based desktop apps. The disappointment is that it comes close to working but then fails in so many different ways. It's a tease.

  • There is no reliable cross platform way to check and then install a java environment. And then to add insult to injury, there's no user friendly URL to send users to that does it for them. Try looking at www.java.com. It's so sad. Lots of useless and confusing marketing and a little link in the corner to install Java.

  • Java Web Start is the very first experience my users get with BlogBridge. And it is the #1 cause for problems. I would say about 10% attempts don't work. People will only give an app like BlogBridge one chance so those are all lost sales. If you are really technical, you might be able to fight your way through and figure out how to install the jre or re-install it because it was corrupted.

  • Even when it works it does absurd things like telling the user that installation is strongly discouraged if there is no digital signature. This might sound like it makes sense but if you really think about it, a digital signature from BlogBridge Inc. would add no confidence whatsoever to the user that bad things wouldn't happen. All I need is $300 for a certificate and a postal address.

And there are lots more reasons. The thing that is hard to understand is why Sun doesn't see that if you want Java to take hold on the desktop, solving this problem, really solving it, is more important than probably any of the improvements in 1.5. And the only way to solve it is to have hard-core experts in each platform (Windows in all its flavors, Mac in all its flavors, Unix in all its flavors) building the solution. Becuase a working solution will be very difficult to build. But it is the keys to the kingdom IMHO. "