[Katsunami]

Originally Posted by DiapDealer

The ability for users to use their own pre-existing, personalized directory structures and filenames (without duplication upon import to library management software) is one that some people seem to find very important. Have for quite a while. I don't share that need (and have no idea how big of a demographic they represent), but the desire for that ability is real nonetheless. And for better or worse, it IS quite clear that calibre is probably not going to move in that direction. That's not a judgement, that's an observation.

For me, it depends on the program and the media in it. Take music for example. I have all of it in a directory structure:

\Music\Artist\Album [Genre]\Disk 1-5\01 - Title.flac.

The stuff is tagged correctly to the last bit (as far as I can see). I use Foobar to play that music. The only thing I do is to point Foobar to the \Music folder. It reads the folder, and creates a database with tags. If I find a song and want to play it, Foobar plays it, and that's that. Should I need or want to have some metadata inside the song changed, I change it (inside the song), and have Foobar update its database. (So it works exactly the other way around as compared to Calibre.) Same if I have a new album. I rip it to FLAC, tag it, drop it into the folder, and have Foobar take a look at it for the metadata. Done.

However, if one wants to have a library of FLAC files (for the audio system), MP3's (for in the car) and AAC (for the iPod), Foobar's way of working becomes very cumbersome: if you change metadata, you need to change it in three places.

Calibre does exactly that: it needs to keep track of several different files, all having the same metadata. It needs to be able to convert from one format to another. It needs to be able to keep those conversions somewhere. If the metadata changes, it must not only change for one file, but possibly also for 15 conversions. If that happens, would Calibre need to change the metadata in the base files, and then reconvert? That would be inefficient. It's much better that it stores all files by itself and manages them in the background, and keeping a database with all information needed. It's more efficient.

That said, I don't want to be 100% dependent on Calibre. First, I also save the "raw" de-DRM'd original file. Second, I now have a workflow that provides me with a fixed/optimized version of every book I acquire, seperate from the one in the Calibre library.

If you are interested in that workflow...

Spoiler:

1. I have a Calibre library called "Workplace". All books I acquire go there. Before doing anything, I save them to disk, so I have a de-DRM'd version. I stash that away safely.

2. I use Modify EPUB to take out all stuff like DRM artifacts, destroy built-in fonts (which I hate), remove any non-DC metadata, and so on. I then consider the file to be clean.

3. I fix the tags, authors, titles, series, series index, publication date and so on... all the basic stuff I want to have correct. After that is correct, I use Polish Books to update the metadata in the books.

4. I save the books to a folder on the hard drive. ("Update Metadata on export" is turned off. I don't want Calibre doing that behind my back.) Everything is also renamed to:

"\Lastname, Firstname\Series [Index]\Lastname, Firstname - Series [Index] - Title.epub"

5. In this folder, I make a quick run of the books and fix the layout where necessary using Sigil. With many bought books nowadays, that comes down to removing whitespace between paragraphs and maybe updating the cover to a better one.

6. Last, I remove the books from the Workplace library, and import the books from disk again (the ones I just ran through Sigil, obviously). If all is well, I move them to my main Calibre library.

So, the books in the save/import are the fixed/optimized versions of the original files, and they are the basis for my Calibre library. They go into a backup folder. Calibre or Sigil will never touch them again, except if I have a very good reason for that.

This seems a lot of work, but it really isn't, if the original books don't need a lot of fixing with Sigil. I mostly acquire books in series, so Calibre's plugins and bulk metadata change takes care of most of the work. Running the books throuhg Sigil takes me maybe 30 seconds per book.

The end result:
1. Original file without DRM (never used again, normally)
2. To my liking optimized/Fixed original file (never used again, normally)
3. That file, and all conversions, in the Calibre database

And I have backups of all three.

===

After the books are in my main Calibre library, I couldn't care less what Calibre does to them. If anything goes wrong with one or more of those books, I can just re-import them from my base files, including all metadata that is most important to me.

This way, I have all advantages: everything Calibre provides *and* a completely clean Calibre-independent base library. The only three things in those books that refers to Calibre are the meta-data tags for series, index, and dc:contributor which is set to "Calibre <version>". If another program that I may use in the future does not read the series and index tags, then they'll be ignored.