Saturday, July 07, 2007
I've finally figured out how to get tesseract to give me the coordinate data that I need. It was quite simple once I found my way around the codebase. And figured out how to use CMake. And KDE4. Google code has it all in svn. Look at the plugins/scan directory.
It's all very basic right now. I've got to figure out how parse the data and draw boxes around the text. Then some analysis of the data can start. Interface stuff can wait.
There seems to be quite a few people interested in KDE applications for windows. Hopefully they realize that if they want them, they'll have to write them or port them.
Any moderately complex application needs support from libraries. In my case, off the top of my head, scanning, ocr, database, and probably more as time goes on. I'm finding that in most cases there is work that needs to be done to bring the secondary libraries up to speed. They lack either the interfaces or the features needed by other developers. For example, Sane works very well, but doesn't (as far as I know) support the buttons on scanners. In other words, a batch type scanning operation would be a pain in the neck. Someone has written a small application that watches the usb bus for button action, but lacks support for many sane supported scanners, and it would be a prime candidate for d-bus notification. I, and others working away on things, improve the platform. When I need something, I may need to write and submit a patch, and I depend on others doing the same thing.
When the focus moves away from the platform, the platform stops improving. If I was interested in just getting it done quickly, I probably would do best with some Microsoft development platform, probably finding most everything available to do what I want. But why would I do that? I dislike closed environments. As frustrating it is to figure out the code in Tesseract, it is far more satisfying compared to bumping into and working around bugs in some closed libary. Been there done that, and never again.
People can do what they like, work on whatever they want. If someone wants to port something to windows, fine. I long ago gave up believing that things have no cost. The cost here will be born by the free platform.
Subscribe to Posts [Atom]