copying the cocoon folks since we are getting pretty serious with continuations overthere (we implement them using a modified version of Mozilla Rhino, a javascript engine written in java)
on 6/26/03 3:15 PM Ask Bjoern Hansen wrote: > On Thu, 26 Jun 2003, Santiago Gala wrote: > > [...] > >>I still feel shocked when I (rarely) see a JavaVM crash with a seg fault >>(out of memory always, maybe some beta JDK at times). The safety of the >>JavaVM contrasts a lot with the dangers of C/C++ environments, and makes >>it compelling to write a java alternative even when good native >>libraries do exist. This is the viral character of it. I wonder >>is "parrot" will do the same for perl/python/ruby. > > > I think it has always been true for those languages that people > often choose to reimplement something in that language instead of > wrapping a C module. > > Some of the big things that Parrot hopefully will bring are: > easy interop between languages that are targeted to the parrot VM > (use Python and Perl libraries in your BASIC program) > > Shared VM development between the various dynamic languages > instead of having everyone roll their own as it happens now. yeah, this is a cool thing and it's impressive to note that almost all modern programming languages are moving (one way or another) the VMS way of bytecode compilation for a virtual machine. (even XSLT stylesheets are being compiled into bytecode) > Dan Sugalski wrote an article about why we can't just run Perl, > Python or Ruby on the JVM or CLI: > http://www.sidhe.org/~dan/blog/archives/000151.html > > - ask > > ps. http://oreilly.com/parrot/ was the April Fools joke, > http://www.parrotcode.org/ is not. :-) > Wow, a VM with native continuations, very interesting. Question: do you think it would be possible to compile java source code into parrot bytecode? how would the limited Perl typing capabilities would impact that? I feel like crosspollinating these days ;-) -- Stefano.
