Re: Major change proposal

On Mon, 2 Aug 2004 17:35:12 -0700 Mark Hahn <[email protected]> wrote:

> At OSCON last week, IronPython was announced. This is an implementation of
> Python running on the .net interpreter (CLI) from Microsoft.  The most
> surprising thing about the announcement was that it ran as fast as the
> normal Python interpreter.  Past attempts were dissapointing.
> 
> Paul Prescod repeated some of this IronPython presentation at the Vancouver
> Python workshop that he held (Prothon was a big hit there by the way
> <grin>).  Paul brought to my attention something that caused me to lose a
> lot of sleep last night.
> 
> If I were to put my current Prothon interpreter work aside for a while and
> switch to working on "IronProthon" (a Prothon to run on .Net), then I would
> have a fast reliable Prothon just as soon if not sooner, but more
> importantly, it would have the entire Microsoft .net library available to
> it.
> 
> It would also have the Unix/Linux world available through the open source
> Mono project from Novell.  This open source version of .net project is at
> the 1.0 stage and was developed by the author of Gnome, so it is to be
> taken seriously.
> 
> I looked seriously at using Parrot for an interpreter and almost everything
> wrong with Parrot is right with .net and mono.  It is stable and working.
> It has the library ready to use.
> 
> The thing that caused me to lose a lot of sleep is that while this would be
> a great thing for the promotion and adoption of Prothon, it would mean
> shelving code that I have been working on 7 days a week for 7 months.
> Prothon comes first though.  Getting users of the language is more
> important than anything else.
> 
> The nature of Prothon will change a bit.  It won't have the
> industrial-strength (scalable performance) advantage due to the unique
> engine.  IronPython will be identical in that regard.  
> 
> It will become Python3000.  In other words it will just become more of a
> direct competitior with Python (or to be more exact IronPython).  Our
> language features of Prototypes, Python fixes, enhancements, etc. will be
> our marketing points.
> 
> I welcome feedback and thoughts on this major change.
> 
> 
For me one of the big concerns with Prothon is performance. Calculating the
C3 resolution order for each attribute lookup is going to be expensive. So
I am curious as to how well Lieberman's caching scheme will work. I believe
the effectiveness of the cache, more than anything, will determine the ultimate
fate of Prothon. If prototypes turn out to be too expensive then they will
be regarded as a specialty tool. So will .Net support the cache?

Lenard Lindstrom
<[email protected]>