Makes sense. Have you considered porting Squeak itself to JVM? This would get the JVM benefits, and the similar "take advantage of existing stuff" from the Smalltalk side.

Just to be clear: I'm not in any way invested in Squeak or Pharo, just curious about your approach. Well, this and also perhaps a bit skeptical about implementing new stuff when similar stuff exists.

I wrote tool that can export classes from pharo/squeak in the syntax that redline expects. My original idea with this was to aid people who want to develop in Pharo and deploy to redline. In order to make that happen, there would have to be a lot more work done.

There have been several discussions about using that tool to export large chunks of pharo and using that as the basis for the redline class library. I personally have tended to resist that (and swayed many people) because it brings along with it a ton of cruft. That said, we all have full time jobs and I'm starting to think that it would be better to start with cruft and remove rather than starting clean and build up. Which side of that I fall on, really varies from day to day. I'm thinking of taking a chunk of pharo and seeing how hard an export of modify would be vs. doing the implementation from scratch.

https://github.com/redline-smalltalk/redline-smalltalk/tree/...

Well, Dan Ingalls did something a like that. He implemented a Squeak VM in Java. It's six years old, and even then it didn't run a current Squeak image, but it works pretty well.

Announcement here:

http://lists.squeakfoundation.org/pipermail/squeak-dev/2006-...

JAR file here:

http://weather-dimensions.com/Dan/SqueakOnJava.jar

Thanks for this. I didn't know that existed.