GUI libs

[Grant Edwards]

In article <b8camh$g65$05$1 at news.t-online.com>, Martin v. Löwis wrote:
> Grant Edwards wrote:
>> But those C runtime library routines aren't redundant. they're
>> providing functionality that Python doesn't implement. 
> 
> They are *also* providing functionality that Python doesn't
> implement. There is quite some redundancy, too. E.g. a Posix C
> library supports regular expressions, which Python also does,
> and Python doesn't use the C implementation.

If Python doesn't use a particular library module, then it
won't get linked in.  The C run-time library isn't a single
monolithic object.

> > OTOH, having both Tcl and Python interpreters seems a bit
> > redundant.
> 
> "A bit" redundant would be fine, I think.
> 
>> I have used Scheme with Tk bindings, and that didn't use Tcl.
>> The Scheme interpreter was bound directly to the Tk widgets.
>> That results in several advantages to the application
>> programmer, but disadvantages to the binding maintainer. 
> 
> What are those advantages?

The big advantage was that Tk "understands" Scheme variables.
If you hook a Scheme variable to a Tk widget, then changes to
that variable are reflected in the widget, and changes to the
widget also change the variable.  IOW you don't have to use
"wrapped" string/int/etc. objects and then use accessor
methods.

Another advantage was that parameters passed to and and return
values from Tk methods all use native Scheme data types.

> Aren't they also obtained by exposing the Tcl object interface?

Don't know.

-- 
Grant Edwards                   grante             Yow!  Just imagine you're
                                  at               entering a state-of-the-art
                               visi.com            CAR WASH!!