Atze Dijkstra wrote:
> Heinrich Apfelmus wrote:
>> Personally, I'm betting on browser-based GUIs, and in particular Chris 
>> Done's "ji" project
>> It's not on hackage and I get the impression that Chris has abandoned 
>> his little experiment, but he has still accepted a patch I've sent. A 
>> small experiment on my part
>> has actually convinced me that implementing GUIs in this way is both 
>> viable and useful. If someone else were to take an interest in Ji as 
>> well, I'd be very happy to contribute.
> Although I think many GUI's will (eventually) run via a browser,
> I also think there is a place for standalone (i.e. non client/server)
> apps  with a responsive interface with many bells & whistles.
> In other words, until the time browser based GUI's like ji
> are a reality wxHaskell (and/or gtk2hs) will have to do the job.

Oh, I agree completely that desktop applications are still relevant. 
What I mean to say is just that the Haskell ecosystem currently does not 
really offer a low-cost and portable way to make simple GUI applications 
and I think that the browser could fill that niche. In other words, I'm 
thinking about the "I just want a button" desire, I don't really care 
that it's in the browser.

Both WxHaskell and Gtk2Hs are barely maintained and hard to install. I 
have gotten so many emails from people that tried and failed to install 
wxHaskell in order to check out my reactive-banana-wx package. Another 
example is the  hp2any  suite for analyzing GHC performance profiles. 
Unfortunately, I can't use it, because I didn't manage to install Gtk2Hs 
on my OS X machine. Also, Conal Elliott once remarked that he has 
stopped working on his legendary graphical tool ideas because the GUI 
library situation was rather bleak.

I think there is a considerable market for a low effort GUI thing that 
works everywhere. It doesn't have to be great, but I think we're missing 
out on a lot of brilliant GUI ideas that can't come to fruition because 
Haskell doesn't have a very simple GUI framework.

> For some more ji like experiments see also:
> (using UHC)
> The idea in the latter was to offer a wxHaskell implementation
> minimally using code behind a FFI, therefore avoiding as much as
> possible maintenance of a non-Haskell codebase. If not done that (or
> similar) way we will still end up in a similar situation as is the case
> with wxHaskell, i.e. reliance on C++ (or Javascript) code.

Nice, these projects look promising and definitely more mature than Ji. 
What happened to Modular-Haskell-GUI? There's a paper draft in the 
repository, but I haven't seen it published anywhere.

Concerning the non-Haskell code base, Ji uses a quick and dirty 
interpreter that can execute arbitrary JavaScript commands, so not too 
much code is spent in JS. This way, the GUI programs can still be 
written with GHC.

Best regards,
Heinrich Apfelmus


Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
wxhaskell-devel mailing list

Reply via email to