NOT ONLY IS there a huge internal debate at Sun Microsystems about difficulties with Java, but it has emerged that a bug in build 83 of 1.4 was only detected and fixed just before the final build of Solaris 9, according to an insider at the company.
I think that's just a bit of fearmongering myself. Testing is designed to catch bugs before release...
Our source at Sun said that memory hogging is a huge problem, with the JRE being huge, compared to similar runtime environments.
He said it can grow to as much as 900M, while Java 2 programs like "Hello World" hog 9M, Component Manager 160M and Together sometimes reaching 900M. He claimed that on an Ultra 10, Together could take a staggering four minutes to load and to start, while even the much vaunted administration console, SMC can take up to two minutes to start working. These problems all relate to the Java Runtime Environment, he claimed.
Sun Microsystems could not be reached for comment on the problems. \ufffd
A comparison to sizes of other REs with similar functionality would be helpful, rather than us having to take their word for it. But 900 M does sound a *tiny* bit excessive. :-/
I'd feel a lot better about these stories if they weren't traced to *one* person at Sun. But I don't want to shoot the messenger either.
FWIW.
Cheers,
Scott.