RE: Major change proposal

Samir Patel wrote:

>It is good to run prothon under .net  but IMHO you will be better of 
>continuing your current task of releasing prothon 1.0. Here 
>are my reasons.
>- You have already spent more than half a year coding prothon 
>interpreter 
>which works.
>- Almost all of your 1.0 specification is done and you can 
>release 1.0 by end 
>of this year.
>- If you built it, they will come and you know you can built 
>1.0 without a 
>doubt.

It is not clear they will come unless there are libraries.  Many people on
comp.lang.python said they will not use it unless it has the Python
libraries.

>- By having 1.0 under your belt, there will be lot more people 
>who will create 
>useful libraries and tool around it. 

It is kind of a catch-22.  If they don't use it they won't create libraries
and if there aren't libraries they won't use it.

> 1.0 will also give you 
>more credits to 
>you work (like Jim Huggins for creating Jython in the past).
>- By this time Jim Huggins, working with microsoft, will shed 
>more lights on 
>integrating dynamic language  on .net platform. You will know 
>what will be 
>major hurdles you need to cross, plus remember he is working 
>for microsoft so 
>he will get inside into .net which you can only dream of. You 
>will have more 
>time learn from his work (just like python and stackless).
>- 1.0 will give you all implementation details you need to 
>know to jumpstart  
>on .net platform if you choose to do so (provided Jim is 
>successful with his 
>IronPython).

He appears to already be successful with IronPython.  I have IronPython
running on my desktop with all the source code.  I am currently studying how
he made it work.

>- And if you choose to do so, your work effort will be 
>multiply by communitys' 
>work on libraries and tools in parallel which you can take it with you 
>under .net.

The work on libraries for my VM would not port to .Net, just as the current
CPython libraries will not port to .Net.  The current trend is towards VMs
with multiple languages like the Java VM with Jython and .Net with
IronPython.  People want to work on libraries for those VMs because they
know the work will be used with multiple languages.  There is a very real
chance that I will work 6 more months on my VM and no one would ever help me
with libraries.

My doing a custom VM for Prothon is old-fashioned.  That is what I learned
from talking to people at the Python workshop.  It is also common sense.

>- Community will feel more involved since they have a 
>programming language to 
>work with. They might even start using prothon as soon as you 
>have alpha 
>release.

A community will feel involved with Prothon using the .Net libraries also
and they will be able to do it sooner.  I will probably get a useful
language quicker with a .Net port than with finishing my VM.  A useful
library will definitely be much sooner with .Net.

I appreciate your input, but I have given this much thought and lost sleep
over this.  I really believe I am doing the right thing.