From: Maxim Shemanarev (mcseem_at_[hidden])
Date: 2002-04-29 19:19:42
> Yes, yes, yes! But people may and will have different requirements on
> internals representing the polygons or raster graphics. Perhaps,
> a rasterizer math algos from underlying data formats would be a good start
> make it more generic and acceptable for boost.
OK, I'd like to make it more clear. It actually is decoupled. The rasterizer
generates a number of scanlines that don't depend on any particular pixel
format or even the type of the color space. How it's being used completely
depends on the implementation of the scanline renderer class. These classes,
such as agg::render_bgr24_solid know the exact pixel structure, but they are
very simple. I'm planning to implement these classes for all popular
pixel formats and color spaces. There's also a way to implement a universal
scanline renderer using classical polymorhism via the virtual calls. The
only requirement is that the scanline renderer must support a certain
interface (for now it's only one function "render"). Similar approach is
used on higher levels of the library, for example, conversion pipelines.
----- Original Message -----
From: "Petr Kocmid" <pkocmid_at_[hidden]>
Sent: Friday, April 26, 2002 8:04 PM
Subject: RE: Re: Vector graphics: Anti-Grain Geometry, second round
> Hi Maxim,
> > Once again, the question: Is BOOST community interested in a
> > lightweight and
> > high quality vector graphic library of common use?
> Suddenly, I found AGG itself as a perfect drop-in library to my current
> project, but it may be not as easy for others.
> Petr Kocmid
> Unsubscribe & other changes:
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk