On Fri, Jan 24, 2003 at 03:26:35AM -0800, John Meacham wrote:
if people havnt already seen it: http://www.t4p.com/xplat.htm describes some of the issues relating to any endevour such as ours.
I have read the article. And although I do the gtk2hs binding and think it's the right way to go, I know it's not perfect: - Gtk will inevitably impose its view of UIs on Windows users. But these differences are small since Gtk applications look quite similar to Windows applications. - Gtk will not be ported to Aqua since Aqua is so different. There is only one menu bar, dialogs pop up within main windows, etc. but it will feel ok under XonX (XWindows on Macs) Maybe you can define a GUI library that looks and feels native on every platform without compromising the native style guide (look and feel). But the lack of *any* commercial or free toolkit that fulfills these requirements makes me wonder why a small number of people with a nice language can do it (Wolfgang said that even Java doesn't get it right). Maybe we can, but on what time scale? I do not want another "it's already usable" prototype. Peter, Krasimir, Andrey - when would we be able to go to a company and say "Use Haskell to program your applications!" and would feel comfortable about our (GUI) libraries? 5 years? 10 years? The Gtk library has been under development for over 6 years and they only took an average of all GUIs available and didn't try to combine three huge, constantly changing APIs. I'd rather have a GUI library which runs on all platforms, sacrifices some look-and-feel but is available in a couple of month. I would be more than happy to provide an Object I/O like view of my library, but I know it will look different from the one you implemented in the IFL paper, simply it's geared towards Gtk not Win32. If you think you can give me a specification *today* how a portable Object I/O layer would look like when you finished your approach, I would be more than happy to implement it. Axel.