| > I'm not against sexy, interesting designs built by one person -- far from it | > -- but I desperately want One Thing that is multiply supported, even if the | > One Thing is not very sexy. | | Well, people probably won't use the interesting alternative libraries because | they have a standard platform. The problem I see is that we get a not so well | designed standard interface and better approaches just don't have a chance to | get widely used. The problem I'm addressing is that if there is no standard GUI platform for Haskell, people won't use Haskell at all (let alone the interesting GUI alternatives!). Surely "better approaches" have had every chance to be widely used! We've tried the "let a thousand flowers bloom" approach for about ten years, hoping that one or more GUI libraries will emerge as widely-accepted alternative, and it just hasn't happened. That's why I advocate the boring "agree a non-sexy standard platform" approach. I should have said in my earlier message that I am, of course, only expressing my own view, partly in the hope of stimulating debate. Perhaps one reason for lack of convergence is lack of agreement about the main goal. If so, we should clear that up before getting into detailed design. Simon PS: A standard interface does not necessarily have to be "not so well designed". It can be well designed, but perhaps not as rich as the more interesting alternatives.