Re: Toolkit Selection

[Jan Jona Javorsek <jan.javorsek@guest.arnes.si>]

When I started the "toolkit" discussion I was not trying to start any kind
of religious war: I don't have any beliefs in this field. I have been
using different tool kits, mostly from perl, and noticed that that tk
(from tcl/tk) was performing very slowly on my machine. I recognise its
higher portability (Win and Mac, I beleive) and excellent canvas, but I
have oftne used Sx and gtk to speed things up.

In fact, a general examination of available tool kits and their
suitability for Xconq is a better result than anything I expected and I
want to thank Mark Russell Anderson for moving this debate even further in
this direction.

The purpuse, in the end, is to make a wanderful game even better, not make
one gui win over another.

Jan Javorsek


On Tue, 2 Jun 1998, Mark Russell Anderson wrote:

> >>>>> "Stan" == Stan Shebs <shebs@cygnus.com> writes:
> 
>     Stan> OK, so here are the goals and constraints.  I'm not
>     Stan> particularly religious about toolkits, so will consider
>     Stan> anything that meets the requirements.
> 
>     Stan> * My ideal is to have graphical versions of Xconq on
>     Stan> Unix/X11, Mac, and Windows 95/NT.  I would accept interfaces
>     Stan> for other platforms, but wouldn't put much personal effort
>     Stan> into them.
> 
> I'm not very aware of _all_ the possibilities, but when I was checking
> it out, it seemed like wxWindows is a good match here.
> 
> wxWindows has a Mac port. I believe they offer a common interface.
> 
> wxWindows:
> http://web.ukonline.co.uk/julian.smart/wxwin/
> wxMacintosh:
> http://www.aiai.ed.ac.uk/~jonto/wxmac/
> 
> There is also Qt, or Amulet.	
> 
> Qt: (everything but Mac)
> http://www.troll.no/qtinfo.html
> 
> Amulet:
> http://www.cs.cmu.edu/~amulet/
> 
> Others:
> ????
> 
>     Stan> * Xconq should have a uniform appearance on all platforms -
>     Stan> maps should have the same images, colors, magnifications,
>     Stan> etc.
> 
>     Stan> * The interaction machinery should have some common
>     Stan> elements, such as keyboard commands, but other elements,
>     Stan> such as menus or popups, may be different for each platform.
>     Stan> These should be able to use as much common code as possible,
>     Stan> however.
> 
>     Stan> * Any toolkits must either be available with the programming
>     Stan> environment (MFC with VC++) or free software (tcl/tk).
> 
> wxWin, Amulet
> 
>     Stan> * A toolkit must support basic animation.
> 
>     Stan> * A toolkit must be either be optimizable, or allow access
>     Stan> to the fundamental API (X11 under Xaw for instance), for
>     Stan> speed-critical display code.
> 
> I don't know about this. 
> 
>     Stan> * A toolkit must be able to handle the display of very
>     Stan> complex maps, such as earth-50km (200,000+ cells, 3,000+
>     Stan> units).
> 
>     Stan> * A toolkit must allow for tuning of the details of
>     Stan> interface elements so as to look "professional" for the
>     Stan> platforms supported.
> 
>     Stan> * A toolkit must support the construction of interfaces that
>     Stan> look "native" for each platform (tcl 8 not tcl 7).
> 
> At least wxWin and Amulet do this, the native part, that is.
> 
>     Stan> * I'm generally happy with the existing Mac interface, and
>     Stan> generally unhappy with the existing X11 interface, and want
>     Stan> to rewrite all its toolkit-level code (low-level drawing
>     Stan> code is OK).
> 
> 
> kraM A.
> krama@ee.byu.edu                     http://www.et.byu.edu/~krama/
> Mark Anderson
> Reconfigurable Logic Lab: Brigham Young University
>