From: "Yemi I. D. Bedu" Date: 2008-04-04T08:00:39+09:00 Subject: Re: RUBY_IMPLEMENTATION Hello, What do you feel about RUNTIME. IN both your examples it would be "JRuby". A variable of PLATFORM would handle the Windows/OpenVMS information. It would not matter to the code what compiler was used but if so, then VM could be used to specify the machine code compile type "Rubinius/YARV/JIT)/(Interpreted)". Then IMPLEMENTATION could equate to PLATORM + RUNTIME + VM. Would that be sensible? Maybe the ordering is not should be changed. Good day. Yemi Bedu P&R Fasteners, Inc. P&R Castings, LLC. 325 Pierce St Somerset, NJ 08873 (T) 732-302-3600 (F) 732-302-3636 -----Original Message----- From: Charles Oliver Nutter [mailto:charles.nutter@sun.com] Sent: Thursday, April 03, 2008 6:22 PM To: ruby-core@ruby-lang.org Subject: Re: RUBY_IMPLEMENTATION Ezra Zygmuntowicz wrote: > +1 for RUBY_ENGINE Here's a wrinkle for you...JRuby has partial support for both Rubinius and YARV bytecode. And in the future, we'll have *full* support for both. So what would RUBY_ENGINE be if running Rubinius bytecode under JRuby on JVM on Windows? How about YARV bytecode on JRuby on OpenVMS? I think we need to take a step back and determine what we're actually trying to track here. Is it simply what codebase the code is running against? Is it a set of features we expect to be available? It is a host environment? I'd argue we're trying to specify a set of features we expect to be available, and so indicating the "engine" we're running on is the wrong way to go. Perhaps something more like "environment" or "profile" is the way to go for this, and let the implementation be characterized elsewhere. - Charlie