Re: Major change proposal

Joe Mason wrote:
>...
> Yes, I thought the prevailing wisdom was the the .NET architecture
> wasn't good for dynamic languages - hence Parrot.  I wounder how
> IronPython managed it.

It's safe to say that .NET wasn't _designed_ for dynamic languages but 
too much to say that it isn't _good_ for dynamic languages. ActiveState 
did an experiment with a pre-release version of the runtime and given 
their budget at the time they could not make it run quickly enough. This 
translated into common wisdom that the runtime was poor for dynamic 
languages. Here's what ActiveState said:

"Only a small amount of effort has gone into analyzing the performance 
of the runtime, mainly due to the lack of performance analysis tools 
available for .NET. Without such tools, making performance related 
changes is fruitless, as the effectiveness is difficult to measure."

Jim Hugunin had a few advantages like working with the mature version of 
the CLR, having already done Jython, etc.

  Paul Prescod