[Haskell-cafe] Re: [Haskell] ANNOUNCE: first Grapefruit release

[Wolfgang Jeltsch]

Am Montag, 16. Februar 2009 16:55 schrieben Sie:
> I would actually quite like to integrate Hieroglyph with Grapefruit,
> which would give you your Cairo support and give me a sensible way to
> implement events outside of my really rather broken model.

This sounds really interesting. Do you already have some concrete ideas?

The important thing for Grapefruit integration would be to provide signals of 
pictures (i.e., time-varying pictures) instead of just static pictures. In 
the classic Grapefruit version, we have a graphics package which defines a 
special graphic signal type. For the current version, I’d prefer to use 
Grapefruit’s standard signal types, applied to an ordinary picture type (type 
of static pictures). This would mean a more composable solution (Grapefruit 
does time-dependence and interaction and Hieroglyph does graphics).

However, there are situations where it cannot be ignored in Hieroglyph that 
pictures can be modified. I think of compound pictures and of lists. These 
should be updated incrementally but this wouldn’t work with ordinary lists 
etc. Grapefruit is expected to have support for incremental list signals 
soon.

By the way, I like your multi-backend approach (Cairo and OpenGL). If 
Hieroglyph should act as the standard graphics library to use with 
Grapefruit, I would very much like if Hieroglyph’s interface would be 
designed as backend-independent as possible. For example, there is also 2D 
graphics support in Qt. In my opinion, this would be worth supporting for use 
in Qt-based applications. What are your plans concerning Hieroglyph backends?

Best wishes,
Wolfgang