Hi, since java bytecode can be created without using a compiler and it runs on a z/OS JVM, its no problem to run Scala on the z/OS JVM. Furthermore, since a scala class to the JVM looks just like a Java Class, you can also call Scala classes using JNI from COBOL, PL/I or C. That also means you can submit Scala Spark jobs using JNI from COBOL, PL/I or C. I have not tested JRuby, Clojure, Apache Groovy or Jyton, but any language the runs on a JVM, should work, if it does not have bugs like the following: In the current implementations of Scala (the one that are packaged with Spark for z/OS) has a bug, which requires to set the JVM Option on z/OS -Dfile.encoding=ISO8859-1. This problem is marked as fixed as of Scala 2.12.0-M5 which is not yet packaged with Spark for z/OS. Hope that helps. Denis. -----Original Message----- From: David Crayford <dcrayf...@gmail.com> To: IBM-MAIN <IBM-MAIN@LISTSERV.UA.EDU> Sent: Wed, Sep 7, 2016 2:41 pm Subject: Re: A Programing language called Scala?
Apache Spark is written in Scala and that runs on z/OS so I would suggest no probs! Scala is not for the neophyte. It's a fantastic language with so many plus points it makes Java look ordinary. However, beneath the succinct syntax lies hideous complexity that the average Java programmer cannot grok let alone your average COBOL coder. It twists the brain! 7/09/2016 7:57 AM, Paul Gilmartin wrote: > On Tue, 6 Sep 2016 16:03:56 -0400, Tony Harminc wrote: >>> http://www.scala-lang.org/download/ >>> http://www.ibm.com/systems/z/os/zos/tools/java/ >>> http://www.ibm.com/support/knowledgecenter/SSLTBW_2.2.0/com.ibm.zos.v2r2.azkd100/azk_getting_started.htm >> http://www-947.ibm.com/systems/support/machine_warranties/machine_code/aut.html >> >> One could argue that this precludes anything not actually written in >> the Java language. Or one could argue that it's Java bytecodes or >> class files that are or are not eligible, in which case Scala would be >> fine. But any argument that says Scala is eligible would work just as >> well to argue that COBOL compiled into a Java class would be eligible; >> something I expect IBM would not agree to. >> > Why not? One could similarly argue that HLASM code created > by the PL/S compiler is supported by HLASM, but HLASM code > created by the Dignus C compiler is not. The telling point is > whether the intermediate product conforms to the specifications > of the subsequent processor. > >> I suppose the bottom line, as always, is that IBM can decide to >> interpret this the way it wants at the time it wants. >> > Sort of. If HLASM fails PL/S output, IBM is free to modify HLASM > if they deem that easier than fixing PL/S. IBM probably wouldn't > do the same for Dignus C. But Dignus might object strongly if HLASM > is violating its own published specifications. > > -- gil > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
