Fwd: Java SE general and spec comments.

Brian Goetz brian.goetz at oracle.com
Wed May 2 16:25:21 UTC 2018

This was received on the java-se-spec-comments list.  

None of these are actually comments on the SE spec.  

For (1), the poster should inquire as to the status of JEP 306 on its associated mailing list.  JEPs are targeted when they are ready; they are not back ported to old release trains.  

For (2), this has come up many times and been rejected as having a poor return on complexity.  However, more general functionality surrounding numeric library types may be considered as part of Project Valhalla.  The poster may wish to follow that discussion.  

For (3) - (4), ideas for features can be socialized on the appropriate OpenJDK mailing lists, and if there is interest and someone is interested in signing up to do the work, can propose a JEP.  

> Begin forwarded message:
> From: A Z <poweruserm at live.com.au <mailto:poweruserm at live.com.au>>
> Subject: Java SE general and spec comments.
> Date: May 2, 2018 at 8:18:17 AM GMT+1
> To: "java-se-spec-comments at openjdk.java.net <mailto:java-se-spec-comments at openjdk.java.net>" <java-se-spec-comments at openjdk.java.net <mailto:java-se-spec-comments at openjdk.java.net>>
> To who it does concern,
> 1)  I notice that neither the Java SE 8, 10 or 11 specs
> do not yet include JEP 306.  I believe that it should be included,
> certainly for version 11 Java.
> Java floating point probably needs to be enhanced in place.
> Certainly given library code that presumes a lack of
> floating point underflow or underflow;
> javax.vecmath.Vector3d.cross
> is an example in case.
> 2)  None of the JEPS seem to include anything about adding
> at least operator syntax support for
> any classes or types that are as BigInteger or BigDecimal.
> 3) There isn't an algebraic transcendental function library,
> equivalent to StrictMath, for BigDecimal
> (and therefore BigInteger too)  supporting operators
> upon and thereby those types, under Abritrary precision.
> 4) BigDecimal and BigInteger aren't truly Arbitrary Precision,
> pro-extensible, in terms of precision subject only
> to jvm accessible memory.
> For the sake of Java, the JCP, robust, clear and useful coding
> everywhere,
> poweruserm
> (Anonymous University Degree Graduate Java Programmer.)

More information about the java-se-spec-experts mailing list