Calibre - An excellent program, but no thanks

[kovidgoyal]

@QuietSilence: The whole point of calibre is that there is no longer any need to maintain metadata in filenames, which is a very dicey and error prone endeavour in the first place, hence your attempt to come up with file naming conventions. That effort by the way is doomed to failure. What happens with books that have metadata that includes characters that filesystem dont support or whose metadata is longer that the filename limit of the filesystem. Rather than encouraging the preservation of this error prone methodology, calibre delibrately tries to divorce you from over-reliance of filenames.

Maintaining metadata in filenames is sooo yesterday

[kacir]

Originally Posted by tompe

To optimize for command line use you have to think about things like navigating in shells and file name expansion. For example a director or file name starting wit a capital letter is very inconvenient

When I started to use FreeBSD, the first thing I used to do was to install bash, vim and mc.
Later on I tried to use the tcsh shell - the default shell for root. When you are a beginner and want to use examples in The Handbook, tcsh shell is quite essential. I very soon learned that tcsh is a very powerful shell. When I discovered package called tcshrc I fell in love with tcsh.

At the moment I use Mint Linux - an Ubuntu clone. By the way, being able to use Calibre Out-Of-Box is one of reasons. Even in Ubuntu I still use tcsh as my default shell. When you install plain vanilla tcshrc package ( http://sourceforge.net/projects/tcshrc/ ) you get a *very* intelligent autocompletition that works regardless of the case, you get much better history (you type part of the command and press the Up arrow and the history is searched for commands beginning with the typed part - exactly the way the comandline history wirks in Vim)

Do give tcsh in combination with tcshrc package a try.

Or search the net for some clever bashrc file. I am pretty sure that somebody has written a set of fiendishly clever set of bash configuration files that can do <Tab> key completion that works regardless of the letter case.

[SolRaven]

Originally Posted by stisev

Hello everyone,
I've been lurking for a long, long time, but finally decided to register and provide my 2 cents regarding calibre...
...That is why I have decided to "boycott" Calibre

You don't have to use Calibre. Just because a vast many of us enjoy it doesn't mean you must as well. If you don't enjoy it, don't use it. There is no need for a formal boycott of a program that is free for any to enjoy. However, if you do feel the need to provide your $.02, or more, I'm sure with a nice donation you could politely ask the creator if he, or partners, would make some adjustments in the program. That would be the affable thing to do instead of criticizing something that the creator has been so gracious to actualize and freely make available to us all.

[tompe]

Originally Posted by kacir

When you install plain vanilla tcshrc package ( http://sourceforge.net/projects/tcshrc/ ) you get a *very* intelligent autocompletition that works regardless of the case,

I use tcsh. The disadvantage with add on packages is that you have to install it on all computers you use and for all users. And when you help some person solving a problem he will not have the add on installed and youy will get irritated. I usually prefer to use the standard version of programs.

you get much better history (you type part of the command and press the Up arrow and the history is searched for commands beginning with the typed part - exactly the way the comandline history wirks in Vim)

This is in standard tcsh. Use "Esc-p" or maybe "Alt-p" for this functionality. This is the functionality I cannot live without and the reason I still use tcsh.

[Jonezyalz]

I think Calibre is a great program but I have to agree that its really annoying that you can't choose the output file structure. Whilst it is an excellent idea to structure it by Author, the problems arise for me when I have a lot of textbook ebooks and have no idea who the authors are as I can identify them better by category. I just use a different folder then copy them across into folder I wish to use, annoying but such is life I Cant be lazy all the time lol!

[macr0t0r]

Originally Posted by Jonezyalz

I think Calibre is a great program but I have to agree that its really annoying that you can't choose the output file structure. Whilst it is an excellent idea to structure it by Author, the problems arise for me when I have a lot of textbook ebooks and have no idea who the authors are as I can identify them better by category. I just use a different folder then copy them across into folder I wish to use, annoying but such is life I Cant be lazy all the time lol!

But, why not use the Calibre Search Bar, then? Isn't that the whole point of tagging books?

Admittedly, since I see little use for this feature, I suppose I shouldn't be here. But, the whole "sort by folder" thing is something I abandoned years ago. It just doesn't work. Too many titles or subjects have special characters that are not permitted by the OS. Should a book be stored by author, genre, or series? If I want both, should I just keep multiple copies of the book in different locations, but what if I want to update it?

Once I put a book into Calibre, I use Calibre to find the book. I can sort and search by any attribute, and I can add multiple tags to each book. It's just easier.

If the intent is to make Calibre work with other programs, then I'm interested to know if those other programs support a customized directory structure.

In any case, I still think the exported link structure will probably be your best compromise, but I hope that at least one of you requesters know a touch of python to get things rolling. Otherwise, I think it's becoming a bit clear that this feature will be a low priority unless one of the other supporters step up and accepts this project.

- Jim

[tompe]

Originally Posted by macr0t0r

Admittedly, since I see little use for this feature, I suppose I shouldn't be here. But, the whole "sort by folder" thing is something I abandoned years ago. It just doesn't work. Too many titles or subjects have special characters that are not permitted by the OS. Should a book be stored by author, genre, or series? If I want both, should I just keep multiple copies of the book in different locations, but what if I want to update it?

Since most people manage to sort their library of paper books and find them I do not see why it should be harder for ebooks. And for a folder sorted structure you can always use find or some other command line tool to find the book if you happen to forget were you put it.

Also a library of book you usually want to have distributed on a lot of computers.

[Xenophon]

Originally Posted by tompe

Since most people manage to sort their library of paper books and find them I do not see why it should be harder for ebooks. And for a folder sorted structure you can always use find or some other command line tool to find the book if you happen to forget were you put it.

Also a library of book you usually want to have distributed on a lot of computers.

You could use DropBox. Or you could stick both Calibre and the library on a USB key. There are many other choices. For my particular uses, I've never needed to access the files directly in the file system rather than through Calibre. Your mileage may vary.

My approach is to treat the file structure maintained by Calibre as an opaque database, and to use Calibre itself for finding books and loading them onto my various readers. When my wife and I need the books on another machine, we simply blast the entire structure to the other machine (optimizing, of course, via rsync or SuperDuper or some other such utility). Works fine.

Meanwhile, the structure used by Calibre plays nicely with TimeMachine which makes it easy to get backups without repeatedly backing up a single ever-growing database (as with the original version that stored all book in the database itself).

Xenophon

[tompe]

Originally Posted by Xenophon

My approach is to treat the file structure maintained by Calibre as an opaque database, and to use Calibre itself for finding books and loading them onto my various readers. When my wife and I need the books on another machine, we simply blast the entire structure to the other machine (optimizing, of course, via rsync or SuperDuper or some other such utility). Works fine.

The problem is how to synchronize all the copies. You need some merge functionality.

Is there a web interface to the calibre database? Using my own web server is one way I make all my books available to my phone or my N810.

[kovidgoyal]

Originally Posted by tompe

Since most people manage to sort their library of paper books and find them I do not see why it should be harder for ebooks.

Because you are likely to have far more e-books than paper books.

[itimpi]

Originally Posted by tompe

Is there a web interface to the calibre database? Using my own web server is one way I make all my books available to my phone or my N810.

Calibre has a built-in Web Server.

Alternatively if you are using Windows you can use the Calibre2Web script (search this forum for the thread on the details).

[tompe]

Originally Posted by kovidgoyal

Because you are likely to have far more e-books than paper books.

The sorting principles does not vary depending on the number of books. So I really do not believe that. When you search for a book you nearly always know what type of book it is and if not you can easily check that with google.

If you search for a book you can never be sure that you do not have the book if for example the author name is not spellt correctly. Checking in the correct sub-folder you can be sure that you do not have a book.

[kovidgoyal]

Originally Posted by tompe

The sorting principles does not vary depending on the number of books. So I really do not believe that. When you search for a book you nearly always know what type of book it is and if not you can easily check that with google.

I must introduce you to a few librarians.

If you search for a book you can never be sure that you do not have the book if for example the author name is not spellt correctly. Checking in the correct sub-folder you can be sure that you do not have a book.

I'm sorry, but I simply disagree that categorization is superior to search for finding books. As for you specific example just sort the calibre book list on author and scroll down it to roughly where you expect the author to be.

[tompe]

Originally Posted by kovidgoyal

I'm sorry, but I simply disagree that categorization is superior to search for finding books. As for you specific example just sort the calibre book list on author and scroll down it to roughly where you expect the author to be.

I think both methods have problems. I just do not agree that search is superior in all circumstances. I usually want to have both search and the possibility to search in a folder structure. And that is exactly what I have in a library or with my paper books at home (structured search in the book shelves and unstructured search in my book register file).

[kovidgoyal]

Originally Posted by tompe

I think both methods have problems. I just do not agree that search is superior in all circumstances. I usually want to have both search and the possibility to search in a folder structure. And that is exactly what I have in a library or with my paper books at home (structured search in the book shelves and unstructured search in my book register file).

You talk as though searching in the calibre folder structure is impossible.