Originally Posted by Jellby
If I were to write such an app, I'd try to implement some kind of heuristic that would parse the metadata (<meta> tags in particular) looking for anything that looks like "series" or "number" (maybe some kind of regex on the "name" attribute). That should catch calibre's metadata, but also other possible schemes.
That would certainly work. I used a similar process when I wrote my (very basic) program to convert the html files I had created into ePub...it used those meta tags to do everything from automatically naming the file and placing the file in the appropriate subfolder to creating and then putting the appropriate info in the toc/ncx/opf.
It wasn't too difficult - although I made it easier on myself by being consistent with the tags I put in the html in the first place.
I'm not sure how this particular developer - Kris is his name - will go about doing that...or even if it's specifically Calibre's set of meta tags...I can't recall if he specifically said calibre, or if he just said "series/#" and I assumed it was calibre because that is what ePubMetadataEditor uses...I'll have to go back through that thread and see. If you take a look at what he has already accomplished with Marvin, he is WAY more advanced than me in his programming skills and shouldn't have any problem getting it to work.