Quote:
|
Originally Posted by scotty1024
So far the Quantum DJVU file is tiding me over. But I must admit I feel weird building a viewer for a file format I have but a single instance of! You all really are going to owe me for this one.  |
There is also some String Theory available, but I am sure you will prefer a trustyworthy theory
Believe me there is some more instances. But it is true, DJVU is about some different page description formats packaged as if it were a single file, so a single instance is risky. Basically DJVu has three layers: background, foreground and OCR. The latter one can be empty, as well as background. Background admits a couple of compresion/smoothing formats, while foreground is about the real trick, a compression able to locate letter shapes and use them as patterns of compression. DjVuPhoto is, I guess, one of the allowed formats for background.
Basically, this means that a serious testing would need of about half dozen of different files: with and without OCR, with and without background, and some different backgrounds. Most people having DJVU files usualy ignores the fine details of each file, somehow they are not surprised about being able to search for text in an scanned document! (btw, I do not think we need search tools for the moment).
As I told elsewere, I see no problem about doing it microsoft-way: first release the program, then wait for users to tell about the bugs.