[Info-vax] Python 2 support ends 1-Jan-2020

[Arne Vajhøj]

On 1/4/2019 6:26 PM, Arne Vajhøj wrote:
> On 1/4/2019 4:42 PM, Dave Froble wrote:
>  > For something like java, the concept of a Java engine that once
>  > implemented in any environment would then execute Java code appears to
>  > be a good idea.  The reality might not live up to the goal.
> 
> It seems to work.
> 
> My guess is that >50% of all Java code is:
> - written, build and unit tested on Windows
> - tested and run in production on Linux
> 
> No problems.
> 
> Platform specific Java code is extremely rare.

Bjarne Stoustrup once said that "Java isn't platform
independent; it is a platform."

In many ways I think he is right about that.

But that doesn't really change anything.

The correct phrase is just: Java code that is specific
for the OS platform below the Java platform is extremely
rare.

Note that this situation has not happened by magic. But
via a deliberate effort.

To get there you need:

1) A very large standard runtime library.

Java 8 has almost 111000 public methods.

2) Enforce the standard.

Java has a TCK that any Java implementation must pass.
And SUN/Oracle will sue if something non-compliant is
sold as Java. SUN took 20 M$ from Microsoft on that
account. I suspect that Oracle will ask for 5-10 B$.

3) Not having any OS specific modules.

There is no java.unix or java.win32 packages.

4) Design standard API and various guides to
    steer developers towards generic code.

Java does that. It is almost impossible to find
Java code that has a reference to a (relational) database specific
class. JDBC API and JDBC documentation got everyone
to use the generic API.

Arne