Comments after my week of ebook importing
I have now imported my entire ebook collection of over 7000 ebooks and scanned periodicals. A tremendous amount of work remains. Due to the vagaries of the import processes, Ebook titles need to be restored and/or cover art extracted.
My overall assessment of Calibre is positive enough that I have invested about 40 hours (so far) into moving my collection into the Calibre structure and my old folder-based library now exists only on DVD.
Calibre is good, constantly improving, but not yet great. If not for the "constantly improving" aspect, I do not believe I would have made the move to Calibre.
Tips for those considering a move to Calibre:
#1 DO NOT ATTEMPT TO IMPORT YOUR ENTIRE LIBRARY ALL AT ONCE
#2 DO NOT ATTEMPT TO IMPORT YOUR ENTIRE LIBRARY ALL AT ONCE
Move books in smaller chuncks, check your import, and don't be afraid to delete the imported books and try again with different settings. Learn how the filename filter works. Depending on the size of your library, import one author at a time, or one genre or ?? Learn the ins-and-outs of Calibre as you go.
And in case I forgot to mention: DO NOT ATTEMPT TO IMPORT YOUR ENTIRE LIBRARY ALL AT ONCE
On the issue of Calibre's folder structure.
Kovid is adamant that the database folder structure should not be a user concern as long as Calibre manages the task of database maintenance. I have come to trust Calibre to do so, but I do see some serious caveats.
First take into consideration the wealth of information that may have been encoded into your previous folder and file-name organization scheme. Here is my old system:
Level 1 Folder Name <Language>
Level 2 Folder Name <Fiction/Non-Fiction>
Level 3 Folder Name <Genre>
Level 4 Folder Name <Series or Author name>
Level 5 Folder Name <Title (if multiple files)>
File Name <Author - Series - Series Index - Title>
Now, when I move my ebooks to Calibre, based on import settings, some of the above information must be manually re-entered, especially if I choose to have Calibre set cover art. Now I MUST use Calibre to do my searches, while any OS could do it in the past.
This could be a disaster the Calibre database becomes corrupted. IMHO Metadata storage is not yet mature enough for Calibre to automatically rebuild its database.
The much bigger issue is trust. I have to TRUST my data to Calibre. I do, but I can understand why others would choose otherwise.
What needs to change?
The above is not a call to modify the Calibre folder hierarchy. The better answer is to ensure users feel secure in trusting their valuable data to Calibre. To me, this means to take this really good software package and make it great. To fix what I perceive as shortcomings in metadata redundancy, to implement an automatic database backup/recovery system, to integrate the ability to use external tools into Calibre.
Kovid has done an amazing job, but all of the above it too much to expect, especially in a short time.
What can I (or you) do?
I cannot make any promises, but I truly hope to become an active participant in helping better Calibre. I don't know that I'll manage to produce anything worthwhile, but I am going to try. The first hurdle is learning Python (and Qt). Then on to Calibre itself.