Quote:
Originally Posted by alex_d
processing folders of images as input to rasterfarian (and also some special features for scanned docs) is a big item on the rasterfarian todo.
as for the rasterfarian code: the core conversion code is pretty neatly contained in kernel-processing.cmd. All the other logic is pretty useless to look at because the same stuff can be done much more elegantly in a better programming language. E.g. a language that, incredibly, has arrays. I won't even mention structures. Godforbid classes. Reading source code before writing your own isn't usually a good idea unless your goal is nothing more than porting.
Anyway, i've started doing a little bit in c# already. Definately fully OO. This is in part to practice coding c# itself, but I also like the new features like interfaces and properties. I know it's nothing that classes aren't able to accomplish. But don't you hate the people who say C is just as good for doing OO as C++? Syntax does matter.
I'd like to also see modularity in the way rasterfarian works, to make it easier to add input or output formats (or the same format but alternate utility) or processing filters.
Re a gui. I'm not able to do it yet, but it would great if there was a nice workflow for picking input docs, manually cropping them, choosing filters and setting parameters, and having a dynamic preview for all that. This would be the real point of c#.
|
I think that making RasterFarian modular would go a long way in making features easy to add. I was at the same stage with PDFRead a while ago, and I made the same decision to re-write PDFRead for more maintainability (I think we discussed this in another thread). After that, adding support for new formats was a breeze (see
this and
this) -- even adding more steps to the pipeline was simplified to a great extent, I didn't have to go hunting all over the code. I hope you find the time to do that!