Ruby and .NET - how will they taste together?

Wow, big news to start off my new blog. The Gardens Point GP Ruby .NET development team has announced an initial beta version of their system.

At this point, they claim that it can compile Ruby source into verifiable .Net v2.0 assembly, or it can run Ruby code directly in a compile, load and execute cycle. They do warn that their implementation is not yet complete, although it does pass everything in samples/test.rb (I wonder if they're using the Rubicon/Rubytests stuff for further testing?).

Source code for their this release is available here. This software is being released as free software, so there should be ample opportunity to work with other development groups.

All this prompts an interesting set of questions though:

  • There are now several groups working on Ruby implementations (JRuby, YARV, metaruby, etc.), how much of this is wasted duplication and how can the various groups help one another?
  • Which of these is going to gain widespread traction in the Ruby community, and how will that affect the projects that don't?
  • Some of the projects working in this space seem to suffer from the Not Invented Here syndrome, what needs to be done to better communicate the goals and progress of all these projects to the comunity to help reduce needless duplication?




Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

More Ruby

Megatux's picture

Don't forget other future implementations, in hlvm ( ) and even in the NekoVM


Pat Eyler's picture

thanks for pointing these out. I'd not seen them before, but they're certainly something I'll keep an eye on going forward.

I'd be interested in seeing how much they interact with the other Ruby implmentations, especially the main C based implementation.


Unique Perspectives

Charles Oliver Nutter's picture

Greetings from the JRuby team! We believe that all these different implementations can only help the Ruby world. YARV is intended to become Ruby 2.0, and its implementation challenges have led to many questionable features of Ruby 1.8 potentially being cut. metaruby is an attempt to write Ruby in Ruby, but it is mostly dependent on a Ruby-to-C compiler which cannot support core interpreter features as yet. There are also several other .NET-related projects at varying stages of completion. We welcome all comers.

JRuby is becoming very mature at this point. We can run a number of core Ruby apps out-of-the-box. We can run large parts of Rails, and more complicated apps every day. We are working toward speed parity with C Ruby before we start working on compilation...and all bets are off afterwards. JRuby is without a doubt the furthest along in creating a complete Ruby runtime, and we believe both the Ruby and Java worlds will benefit from it. The other important detail of all these projects is that they are open source; we keep an eye on all projects to see if they have creative ways to solve problems we haven't tackled yet. Some projects are tackling compilation early on, which we have delayed until we know JRuby is correct. Some projects just bridge the C implementation, binding types in interesting ways. All viewpoints are useful.

What would help out all these projects most is hiring on full-time developers to improve them. The potential for Ruby on Java or .NET is tremendous, and yet very few of these project have any official backing or funding. Hopefully that will change in the future.

Another JRuby comment

Thomas E Enebo's picture

Another comment from JRuby-land. As an alternate implementation of Ruby on the JVM, JRuby also adds nice Java integration features. This allows the mixture of Java and Ruby in various interesting ways (like write a DSL in Ruby and call it from your legacy Java application). We see this as a value add for Ruby programmers and it gives Java programmers another avenue to solve a problem. I expect the .NET implementation has the same benefits.

Duplication of effort only makes sense if we had any plans to work on another implementation. I am not .NET guy and I think the C Ruby implementation has enough hands in it already. Creating a second implementation of a language also helps find bugs in the first implementation. We have found a few bugs in Ruby by working on JRuby. Plus if we have three (or more) impls of Ruby, then perhaps it may even generate a formal language specification...