Python not a Very High-Level Language?

[Frank Sergeant]

Denys Duchier <Denys.Duchier at ps.uni-sb.de> writes:

> I apologize for the delay in responding to your question, but our
> newsfeed was out to lunch for about a week.

No apology needed!  Thank you for responding at all.  I am in no
rush, with too many other things to do, but I keep an eye out for
the "best" way to do GUI.  My opinion changes daily, but _today_
I lean toward using Python to generate HTML and let the web 
browser worry about the GUI.

> Drat!  Just the thing I know least about; how did you guess?  I am
> forwarding this message to Christian Schulte who was responsible for
 ...
> scheduled tomorrow where the current state of the GTK interface will
> be presented.  I am eager to learn about that :-)

Please, either of you, post the details as and if it is convenient to
do so.

> The Tk interface follows a fairly simple design: the Oz process starts
> a Tk subprocess and communicates with it via a socket.  

I like the idea of this.  It looks like it might be the simpler/better 
way to manage working with Tk from another language.  I gather Python
can't yet move to Tcl/Tk 8.3 (or anything past 8.05) due to the tighter
coupling between Python and Tk.  It looks like the socket coupling
approach might eliminate much of the maintenance effort in keeping up 
with new releases of Tk.

> On the Oz
> side, Tk widgets are represented by Oz objects.  The Oz/Tk interface
> is actually a very thin OO flavored layer.

> To be concrete, here is the functor (a bit like a Python module) for
 ...

Thanks for the example (I think I understood it) and the reference
to further documentation.  When I find the time ...


  -- Frank