On Thursday 16 November 2000 18:23, John Califf wrote: > The situation is this. Krayon does NOT use the normal QPixmap and > QImage and QPainter classes for its drawing and image processing. > Instead, it uses custom classes and routines, all written at a low > level. Essentially, it's binary reading and writing of scan lines in a > format similar to, but not exactly like, QImage. But these scan lines > are broken up by a number of tiles, which are updated with a timer > which runs continuously. Sounds complicated, and it is. Another > difference is that Krayon access binary data by channel, like RGBA. On= e > for each color, basically. This is well encapsulated and you can access > the data in terms of x,y coordinates and translate to QImage and QPixma= p > formats, regardless of the tiles, but writing TO the data using Qt is > nearly impossible. Each tile is like a piece of paper with writing on > it, wrapping lines. And there are hundreds of such tiles for a large > image, like laying sheets of typewritten paper out on a desktop. Scan > lines are within each tile, and do not run continuously across the > entire image as in almost every other image format. Think of a 1MB JPG images that uses say 40MB uncompressed when in memory.= You=20 can't handle images of that size with QPainter drawing routines. You _nee= d_=20 the tile system to keep only the tiles you currently work on in memory. QImage also can't handle CYMK data. The abstract Krayon image class can (= or=20 can be extended to support it). There are a bazillion reasons why you want to keep the tile system. Greetings, Matthias -- Matthias Elter me@kde.org me@caldera.de _______________________________________________ Koffice-devel mailing list Koffice-devel@master.kde.org http://master.kde.org/mailman/listinfo/koffice-devel