bowerbird said:
>and maybe carel will say that this is exactly what she has meant,
>and we've been doing a semantic dance around the same thought.
Yes, we have been doing a semantic dance around the same thought. My tools would interactively take a human through the process of doing changes. I do not believe in blind changes (other than the simple things that _can_ be done blindly (like some spacing issues within paragraphs, etc.)). In many respects, it would be nothing more than what can already be done using various word processing programs, etc. It would, I hope 'learn' enough or be upgraded enough over time to be more intelligent about the process to better assist the human doing the processing. It would also err on the side of caution.
I just feel that a final 'proofing' stage before release would assist in locating errors that were either missed or introduced by the processing. A human will do the processing and humans can make mistakes and some of the mistakes that could be made in what would be both error and formatting processes could be quite grand. I feel that a second set of eyes can never be a bad thing when it comes to something like this. And, those with less experience (or no experience) in shaping the output of a text may feel more confident about doing the process if they know someone else will provide a checksum for their work before it goes public.
With or without a 'proofing round' (at the discretion of the human doing the processing), the text would be released to PG and then should be placed in some environment that allows for editing it to 'perfection.' This could be DP or wherever, and I would be willing to host texts that have no other home for corrections.
I believe that the people who submit texts (or adopt ones I've introduced into the system) are sentient enough to decide what is best for a project and should just be given a good assortment of tools and options so that they can make the text ready for release. I don't believe in controlling the flow beyond requiring a text to meet certain formatting and error checks if they wish to release the text through _my_ system (to comply with the requirements of PG). Different texts have different needs and forcing a simple novel with some scannos, etc. and some formatting issues through the same process as a book that could not be scanned well or that has complex formatting needs (dictionary, math book, etc.) makes little sense to me. In fact, initially my software will only handle books that require minimal markup so that I can get that process working well before dealing with the more advanced formatting issues.
I'll provide examples when I get a chance to code some. In the meantime, you'll just have to take my word that we are pretty much on the same page when it comes to human processing and 'automated' processing. :)
Back to my school work....
Carel