Would you benefit from file manager on your Kobo reader?

[moffattm]

Originally Posted by theonna

The method that you have outlined has not been implemented yet.
File manager is not something that developers team need to develop, the reason Amazon and others don't want file manager has nothing to do with ease or difficulty of its development, because it is already there, it has to do with their desire to protect their content. Operating systems that run reading devices all have file management system, it is a choice of management to close this system from user and offer instead artificial extra layer for "convenience". I am sure designing shelf and database construct is far more laborious than using existing file manager. It is a matter of choice, not resources.

I think you're confusing a File System with a File Manager.

A File System is a set of routines in the operating system that perform file management tasks (open a file, close a file, read from a file, write to a file, delete a file, etc).

A File Manager is an interface between the File System and the user. In text based operating systems (MS-DOS, Unix, etc) it's a command line interface where you type in commands, i.e. COPY, DEL, DIR, etc. In a GUI it's all graphical; you use the mouse and keyboard and can drag and drop representations of the file system around (icons, etc) and the File Manager interprets your requests and passes these on to the File System in a manner that the File System understands.

So you're incorrect when you say that the Kobo devices already have a File Manager. They've got a File System, but not a File Manager. So the Kobo devs would indeed have to put time into developing a File Manager, there isn't one hidden away inside the firmware that they can just turn on for you.

Given that there is just as much, if not more, work involved in creating a File Manager as there is in improving Shelves (i.e. allowing nested shelves), or even in implementing TechniSol's interesting idea, my vote is that the Kobo devs move forward along the path that they've already started on rather than go backwards to implement a File Manager.

[TechniSol]

Guma,

You do realize the main limitation of your concept of autobuilt shelves is that you can only have each book on one shelf, or you require storage of multiple copies... Also, how do you indicate structures for which you don't want shelves built?

My approach through tags allows for books to be cross referenced in any number of ways without any physical or logical structures being built unless upon demand. Flexibility and less requirements for maintaining the database while lowering overhead every time the device is power cycled. If the sorts were saved and adjusted as need be for additions/deletions they would need be performed only ONCE and would simply exist as a linked list or index to the database. It would only get faster to find what you want once a multi-level sort had been run at least once and maintaining them would require very little overhead indeed.

[theonna]

Originally Posted by TechniSol

Guma,

You do realize the main limitation of your concept of autobuilt shelves is that you can only have each book on one shelf, or you require storage of multiple copies... Also, how do you indicate structures for which you don't want shelves built?
.

For file manager it is not a limitation. The only reason that someone might need a book in more than one collection is that collections are flat. If you have hierarchical structure, you do not need to have same book in many collections, because it would already be filed under genre, author and series. So this is a limitation of a flat shelving system and it gets transferred to a collection system derived from folder structure.
Generally you'd specify a directory from which you'd like to build shelves, and it builds them from all books from that directory and all subfolders.

[davidfor]

Originally Posted by theonna

For file manager it is not a limitation. The only reason that someone might need a book in more than one collection is that collections are flat. If you have hierarchical structure, you do not need to have same book in many collections, because it would already be filed under genre, author and series. So this is a limitation of a flat shelving system and it gets transferred to a collection system derived from folder structure.

That is far to limiting a classification for me. I largely read SciFi. But a lot of that is really a Thriller. Or maybe I should say a lot of Thrillers are also SciFi. Being able to put the books in both a "Thriller" and a "SciFi" shelf is what I want. And of course you will have the same author all over the place. There are plenty of authors who write in multiple genres. And multi-author series will just mess it up completely.

[theonna]

Originally Posted by davidfor

That is far to limiting a classification for me.

Well, it would be just as challenging editing metadata for shelves, because that kind of classification is subjective and can only be done manually.
On the other hand, here the search would shine.

[eXistenZ]

Originally Posted by theonna

For file manager it is not a limitation. The only reason that someone might need a book in more than one collection is that collections are flat. If you have hierarchical structure, you do not need to have same book in many collections, because it would already be filed under genre, author and series. So this is a limitation of a flat shelving system and it gets transferred to a collection system derived from folder structure.
Generally you'd specify a directory from which you'd like to build shelves, and it builds them from all books from that directory and all subfolders.

No, it's not as easy as you think, and I can prove it with just a few real life examples:

- "Good Omens" written by Terry Pratchett and Neil Gaiman,genre Humor Fantasy.
In a folder based structure Where do you put it? under Terry Pratchett or Neil Gaiman, or both making a duplicate of the file? or do you create a "Terry Pratchett and Neil Gaiman" folder just for one book?
Same problem with genre, you need to chose between Fantasy and Humor, or duplicate the book.

- Michael Moorcock books, he wrote fantasy books, science fiction books and literary fiction.
In a folder structure like:
Genre
Author
Series
you need to break Moorcock's biography under tree different genres, and there is no way to have a complete view of all his books.
Otherwise if you make a folder structure like:
Author
Genre
Series
You need to create a fantasy folder (and a science fiction one) under each Author, loosing the ability to have a total view of all fantasy books in your library.

- WildCard Series written by George R. R. Martin, R. Zelazny and many others, each book a different writer.
I cannot imagine a folder based way to take the entire series of books together.


A tag based shelf system, is much more flexible to me.

[tshering]

To my mind it is pretty evident that a database based system has many advantages over a file manager. But if there would be a simple file manager, I could imagine some situations I would make use of it. Say I have a lot of reference works or geographical maps, or anything I would not read from beginning to end, and I would not care that the reader opens them at the same point I left them the last time and so on. I would put the whole thing into a hidden directory in order to prevent the reader from scanning them. Nothing would be written the database, nor a set of cover image files produced. If I would want to open the book or image I would simply navigate to it an select it.

[Terisa de morgan]

Originally Posted by theonna

The only reason that someone might need a book in more than one collection is that collections are flat. If you have hierarchical structure, you do not need to have same book in many collections, because it would already be filed under genre, author and series.

I don't think so. I have shelves/collections for:

• Genres -> Tags
• Challenges: Posible Options for future reads
• Init of series: Divided too by genres
• Series -> Books I want to keep out of genres.

If I want to have the same organization, I need a book in three different folders. And no, this doesn't require a big effort of maintenance, only a bit of programming in calibre. And when I want to change a book from a shelve to other, or I want to remove a book from a shelve, I only have to mark some info in calibre GUI, and the changes will be done the next time I connect the ereader to calibre without having to move books or deleting duplicates copies. Not a normal case? It's the same, it's mine, so I need something that a file browser doesn't give me without duplicating books.

And anthologies, with novellas in different series (something quite normal), how would you include then in all the series if you like.

[Terisa de morgan]

Originally Posted by theonna

Well, it would be just as challenging editing metadata for shelves, because that kind of classification is subjective and can only be done manually.
On the other hand, here the search would shine.

Challenging? Not really, I add all the tags/columns/info I like and calibre works later. Any clasification is subjective.

[DNSB]

Originally Posted by theonna

For file manager it is not a limitation. The only reason that someone might need a book in more than one collection is that collections are flat. If you have hierarchical structure, you do not need to have same book in many collections, because it would already be filed under genre, author and series. So this is a limitation of a flat shelving system and it gets transferred to a collection system derived from folder structure.
Generally you'd specify a directory from which you'd like to build shelves, and it builds them from all books from that directory and all subfolders.

Even with filing by author, you would need multiple copies for quite a few books. I have by a quick count, about 400 books on my Aura that either have more than one author, are edited by more than one person or are anthologies where I like to be able to search by all the authors. Do I file Elvenblood under Norton, Andre or Lackey, Mercedes? And then I have those books that are iffy as to whether they are fantasy or science fiction or mystery. Lord Darcy for an example. Do I file Good Omens, the Nice and Accurate Prophecies of Agnes Nutter, Witch under Pratchett, Terry or Gaiman, Neil? Under humour or fantasy genre? " I DON'T CARE WHAT IT SAYS, said the tall biker in the helmet, I NEVER LAID A FINGER ON HIM. "

One example for my wife, she has a series with 5 books. Each are written by two or three authors with a total of 8 authors involved with the 5 books. Search with the series pops up all 5. A file manager? Going to fun trying to devise a file structure that makes it easy to locate all 5 books without having multiple copies.

Just much easier to be able to search by author, title, series, etc.

Regards,
David

[PeterT]

Nonsense. We just need to use symbolic links on the file system to handle that.

OOPS... It's a FAT32 format file system. Guess we need to request that the file system move to ext2 / ext3, but still be accessible to the PC running Windows that it's attached to.

Originally Posted by theonna

Well, it would be just as challenging editing metadata for shelves, because that kind of classification is subjective and can only be done manually.

On the Kobo device as it exists today, true. On a computer running calibre, it is not quite true.

Say you want to create shelves based upon the author. You can define shelves based upon that by modifying Calibre's configuration. Everything is handled automatically. Books with multiple authors will even appear on multiple shelves. If you don't like that approach, you can take a semi-automated approach by defining a shelf (e.g. Favourite Authors), setting the search criteria (e.g. a list of authors you like), then dragging and dropping the results onto the shelf. By saving the search criteria, you can easily update the shelves when you update your library. You can define arbitrary shelves based upon arbitrary criteria, even if it means a book is on several shelves (or no shelf at all).

Directories work well for something that has a clear hierarchical structure. Sometimes that works for books, sometimes it doesn't. Tags can be used whether there is a hierarchical relationship or not.

Originally Posted by PeterT

Nonsense. We just need to use symbolic links on the file system to handle that.

OOPS... It's a FAT32 format file system. Guess we need to request that the file system move to ext2 / ext3, but still be accessible to the PC running Windows that it's attached to.

Second OOPS ... delete the original file and all of those symbolic links are broken. Better make it hard links. Now you have to make sure users are clear on the difference between symbolic links and hard links. Mass confusion.

Third OOPS ... deleting a book now requires tracking down every copy of it. It's a little extra work if all of the file names are the same, and a lot of extra work if the filenames are different.

[theonna]

Originally Posted by Terisa de morgan

Challenging? Not really, I add all the tags/columns/info I like and calibre works later. Any clasification is subjective.

Exactly. You need Calibre. You also need to go and manually add tags/columns/info.
DNSB, PeterT, eXistenZ, davidfor all offered great examples where their use of their library would require a system to include a book in multiple sets, without having to have more than one copy of a book. It is all perfectly correct, and books having correct metainformation, ability to have more than one tag for each field, should work in searches and database.
As it is now, we do not have ability to search like that on our devices, we also do not have books with correct metainformation. So though I agree it would be a great system if it worked and we had it, but we don't.
We have a device that requires to manage shelves one book at a time, takes long time processing new books, takes long time dealing with many shelves and with managing them. All of it would be unnecessary if you had access to file system, it would not suit all requests for categories and collections, but it is still personal, and you can arrange your books how you like them so you know where you'd find series, authors, you'd know how you arrange books with multiple authors, it is not different from having actual bookshelf at home, you have all your library at your fingertips, without extra wait. It is your library, so you know where you have all your stuff.
You can tell me that the structure of your preferred collection could not be satisfied with a library organized by folders, and that is fine, while programmers work to make it to your liking, you can still have well organized collection of books with a fraction of time if you had a browser.
And after all the great programming is done, foreign language books, which alphabet is not supported yet, would be still unavailable for any search.
Cant you see, file manager is a simple tool, that could work now, while all your requirements aren't met, even without having to have long processing time, without needing Calibre.
It is not perfect but as it is right now, it is still better what we have now.

[guma]

This whole discussion seems a bit strange to me since it is contrasting a 'happy case' scenario (fully tagged libraries, ergonomic tag/library management tools etc.) with a request to - potentially - improve an existing device which does currently not remotely offer the possibility to use - let alone manage - metadata in an ergonomic way.

I mean, this is about KOBO devices, right? Not about what can/could be achieved using Calibre and the like, nor about why, in theory, a meta-data based file organization paradigm is more powerful compared to a file structure based approach.