Subject RE: [IB-Architect] UDF replacement: native shared libraries vs. J ava
Author Jim Starkey
At 11:59 AM 4/18/00 +0200, Kim-Bo.Madsen@... wrote:
>Not to be argumentative.... But :)
>Why is it that the JVM gives Access Violations from time to time using plain
>Java applets..... If the sandbox model is waterproof as is claimed.... A/V's
>would never happen.

"the JVM"? Are you talking about the one true JVM or just one that
showed up on your system. I presume that it access violates because
somebody forgot to debug it.

>The current JVM's are _not_ waterproof. Its been proven time and time again.
>Perhaps they will be at some time... perhaps not. But currently you can
>forget about fixing the bug in JVM yourself. If it A/V's its just bad luck.
>Using native 3rd gen. languages you have a chance to solve the problems if
>you encounter them... using Java... no chance.

I presume Interbase will latch onto an Open Source JVM. If it has
bugs, we'll fix them.

>Further... Im not terribly sure that the argument that programmers are more
>effecient in Java than in another language is true. IMO its not allways the
>case developing ordinary applications, since so many things goes deprecated
>in the Javaworld from one minor release to the next. The basic class sets
>are often also quite complicated to use. Perhaps this does not apply to UDF

Learn to differentiate between Java the language and the set of base
classes. Java the language is completely stable. The base classes
are prone to wander. Since the core set of base classes (Object,
Class, System, Thread, etc) are closely tied to the VM they will
be managed like other pieces of Interbase source.

>In an ideal world, I would be inclined to agree with you that a sandbox
>model is really a benefit. Thus you can leave UDF programming up to people
>not really capable of programming.

We do it every day. Some programmer are better than others. The
trick is keep the the bad ones from taking down the system.

Jim Starkey