[Edu-sig] re : Python Shell - was: RE: IDLE's save-before-run r equirement

[Arthur_Siegel@rsmi.com]

>Well, the current IDLE is 0.8. So any difference in the Shell capability
>between 0.5 and 0.8 are going to show up. And if VPython remains at 0.5 this
>will become more of an issue over time. And if the Edu sig decides to do a
>IDLE Tutorial fork then we are just making things worse, imo. I just can't
>help feeling that these IDLE variants are not a very good thing because they
>are complete forks and not just add-ons or plug-ins. On the other hand, you
>are correct that the differences right now between the IDLE and VPython
>shells is minor or possibly even nonexistent.

I think that one right IDLE should certainly be the goal. As mentioned in the
thread a number of times before, the only fundamental issue seems to be
 the 'single process' in the 'official' IDLE which makes it unusable for 
something like
VPython.  VPython bills its IDLE as having a number of "important
enhancements"
over the standard IDLE.  Other than the processs issue , I for one cannot 
identify 
what they are, and don't know whether I would agree with their billing as 
'important', or even enhancements if I did.  We are perhaps at that point into

a matter 
of taste, style and specialization for a specifc use.  VPython's IDLE, for 
example, color
syntaxes (I believe) VPython key words - certainly not appropriate for a 
general purpose IDLE.

>From my perspective,  anything IDLEish, that can spawn a new process, is a 
suitable environment for creating and running scripts - VPython or otherwise. 

The variants of different flavors of IDLE distributed under one name, is the 
real
problem and confusion.  Seems to me Gudio in effect owns that name.  Guido 
should decide which code base gets to use that designation, through some
mechanism.

If there is a ViDLE distributed with VPython with specialized syntax 
highlighting and
other modifications - that's fine. Among the beauties of IDLE is in providing
the VPythons a code base for a custom IDE.  Played with the concept myself 
on a much less ambitious scale with PyGeo at one point.   But there is
no reason I can  see *not* to be avoiding the confusion of distributing 
something 
different under the same name. 

We are close to being there.  The current VPython puts its IDLE into its own 
directory,
rather than replacing tools/IDLE.

Guido sounds like he is close to checking in the code that would make *the* 
IDLE useable 
for something like VPython.

Maybe then Guido just needs to assert some naming  conventions for IDEs 
derivative of the
IDLE code base.  And we're home.
    
ART