New to ocropus and I've got a few questions and issues:

[Galen Brownsmith]

Is there a canonical or preferred location for documentation and support? The page on code.google.com is lacking (for instance, there are 30+ defects dating back 9 months that are all in the new/un-reviewed state; only the most recent few are ones I submitted).

Also, the documentation is a little sparse and aimed at people with some measure of field expertise. (-k is the k value for Sauvola Binarization. What does that mean?  How does changing it impact recognition, if I were using it rather than nlbin.  How does costing work?  Is cost a sum of the elemnts, a product of them?  I have a decent grasp on confidence rates in computer learning; how does that relate to confusion rates?) Is there anywhere there is a primer on how OCRopus works?  I don't need specific details, but enough to understand how it works in general.

Is there a preferred means of submitting potential patches?  The inconsistencies in the ocropus command line scripts are driving me a bit nuts and I'm working on that.  (One uses "--parameter=value" format; the rest are "--parameter value", most of the time "-Q FOO" means 'run in parallel using FOO separate processes', except when it means 'nocheck' or 'execute this (arbitrary python?) before anything else' -- but that's usually '-X'; '-V' is usually not verbose, except the one time it is, etc. .)


-- ggb