multiple authors in metadata

[tulsa]

Does anyone know how to properly enter multiple authors into the Calibre author field?
When I have books with two authors, the downloaded metadata is never consistent. Sometimes, it uses an "&", other times a slash.

If I separate them with a comma, will Calibre know that two authors belong to that book?

[GRiker]

If there are multiple unrelated authors, Author Sort should be:
last_primary & first_secondary last_secondary, first_primary

Smith & William Jones, John

which displays on the Kindle as
John Smith & William Jones

If there are multiple related authors, Author Sort should be:
last_primary, first_primary & first_secondary

Smith, John & Nancy

which displays on the Kindle as
John & Nancy Smith

Remember to check Preferences|Conversion|MOBI|Use author sort for author.

G

[tulsa]

Originally Posted by GRiker

If there are multiple unrelated authors, Author Sort should be:
last_primary & first_secondary last_secondary, first_primary

Smith & William Jones, John

which displays on the Kindle as
John Smith & William Jones

If there are multiple related authors, Author Sort should be:
last_primary, first_primary & first_secondary

Smith, John & Nancy

which displays on the Kindle as
John & Nancy Smith

Remember to check Preferences|Conversion|MOBI|Use author sort for author.

G

Ok, thanks. Does that work for the Sony 900, as well?
I was actually thinking along the lines of organization within Calibre. I usually only keep about 10 books on my devices at any given time.
But thanks!

[Starson17]

Originally Posted by GRiker

If there are multiple unrelated authors, Author Sort should be:
last_primary & first_secondary last_secondary, first_primary

which displays on the Kindle as
John Smith & William Jones

That's odd. I always enter the author field as:

John Smith & William Jones

then I click the author_sort button (on the metadata single dialog screen), which enters it as:

Smith, John & Jones, William

If there are multiple related authors, Author Sort should be:
last_primary, first_primary & first_secondary

Smith, John & Nancy

which displays on the Kindle as
John & Nancy Smith

In this case, I just enter:

John Smith & Nancy Smith

which, after clicking the author_sort button, becomes:

Smith, John & Smith, Nancy

I've never tried your method. Is there any problem if I search for Nancy Smith and/or John Smith as the author? IOW, is "Smith, John & Nancy" properly parsed as two authors named Smith?

[GRiker]

My technique is what I use to massage multiple authors for proper display on Kindle. I don't know how other readers display author metadata.

I've never tried your method. Is there any problem if I search for Nancy Smith and/or John Smith as the author? IOW, is "Smith, John & Nancy" properly parsed as two authors named Smith?

Yes, that's been my experience on the Kindle.

G

[Starson17]

Originally Posted by GRiker

Yes, that's been my experience on the Kindle.
G

I assume the "Yes" is that it is properly parsed as two authors named Smith on the Kindle. I wonder about inside the SQL database of Calibre. I don't have an e-reader, so that's all that I see. I was thinking about a search like author:"Nancy Smith" and the same for John Smith. Would it pull up all the books by them, including the joint authorship and sole authorship books.

[GRiker]

I should have made it clear that I was referring to Kindle-optimized author metadata.

On the Kindle, searching for an author is a full-text search. I've never had a problem finding a jointly-authored book using either "John Smith" or "Nancy Smith" as the search term. It could be that the Kindle is actually finding the text within the text, as opposed to in the metadata. My focus has been on presenting the authors in an acceptable fashion on the Kindle, and I discovered that this style worked best for me. I do understand that it's not a correct sorting value, but it's what's needed for the Kindle to show it properly.

Within calibre, I treat the Author field differently - I enter the author attribution as I expect it to ultimately be displayed on the device:

John Smith & William Jones
John Smith & Nancy Smith

That allows me to search for "Nancy Smith" in calibre, and for the author(s) to be properly rendered on the Kindle.

I realize that this may not be an optimal solution for folks with multiple devices utilizing different formats. Once I receive my iPad I may have to revisit this ...

G

[tulsa]

Originally Posted by Starson17

I assume the "Yes" is that it is properly parsed as two authors named Smith on the Kindle. I wonder about inside the SQL database of Calibre. I don't have an e-reader, so that's all that I see. I was thinking about a search like author:"Nancy Smith" and the same for John Smith. Would it pull up all the books by them, including the joint authorship and sole authorship books.

This is kinda what was going through my mind.

This would then bring up the issue about the way Calibre physically stores the ebook. Since its stored by author/title/nameofbook, seems it would be difficult for Calibre to decide which author is the primary author.

There ought to be a new field like 1st author, 2nd author, etc. Or at least have the option of adding an author and then having Calibre always use the first author to store the book under.

dunno.. just thinkin' out loud.

[Starson17]

Originally Posted by tulsa

This is kinda what was going through my mind.

This would then bring up the issue about the way Calibre physically stores the ebook. Since its stored by author/title/nameofbook, seems it would be difficult for Calibre to decide which author is the primary author.

There ought to be a new field like 1st author, 2nd author, etc. Or at least have the option of adding an author and then having Calibre always use the first author to store the book under.

dunno.. just thinkin' out loud.

Having worked on the code, I know a bit about this. Calibre considers each author to be an item of metadata called "author" in another chunk of metadata (a list) called "authors." Each author is independently linked to the book. Searches for an author will find all of that author's books regardless of whether he's a single or joint author and regardless of whether he's first or last in a list.

Calibre stores a multi-author book in its black box secret and thou-shalt-not-touch storage system under the name of the first author in the list, but that doesn't mean that author is any more primary than the others. It's just how the book is stored, and since we're supposed to ignore that, it makes no difference.

[HarryT]

Originally Posted by Starson17

I assume the "Yes" is that it is properly parsed as two authors named Smith on the Kindle. I wonder about inside the SQL database of Calibre. I don't have an e-reader, so that's all that I see. I was thinking about a search like author:"Nancy Smith" and the same for John Smith. Would it pull up all the books by them, including the joint authorship and sole authorship books.

My experience is that you'd need to enter the author, in Calibre, as "Smith, John & Smith, Nancy". That will definitely store the two authors separately and will subsequently match searches by either name.

If you enter an author of that form, and then edit the metadata for a book and look at the "author" drop-down, you'll find that it lists "Smith, John" and "Smith, Nancy" as separate authors. If, on the other hand, you enter "Smith, John & Nancy", the drop-down will list "Smith, John", but then simply "Nancy".

[chaley]

Originally Posted by Starson17

Having worked on the code, I know a bit about this. Calibre considers each author to be an item of metadata called "author" in another chunk of metadata (a list) called "authors." Each author is independently linked to the book. Searches for an author will find all of that author's books regardless of whether he's a single or joint author and regardless of whether he's first or last in a list.

To add a bit more...

As Starson17 said, author names are stored as individual records. Author lists are a set of words separated by certain magic 'words' such as 'and', 'with', and '&'. Thus you should enter each author in the list as a complete name and separate the authors with one of the magic words. If you do not follow this convention, then some authors will be missing or incorrect. That may be OK (it certainly is for GRiker), or it may not be.

When you ask for your books to be sorted by author, they are not in fact sorted by author but are instead sorted by author_sort. Author_sort is not stored the same way as author. Instead of being an individual record per author, author_sort is kept as a single string per book. When you add a book or press the 'compute author_sort' button in the edit metadata dialog, by default calibre looks at each author in order, attempts to guess the last name, generates a guessed 'lastname, firstname' string, then adds this string to the author_sort field for the book. There are two interesting consequences:

1. calibre can get it wrong, and will get it wrong in the case of names like 'fred smith esq'. This example will become, 'esq, fred smith'.
2. you can change author_sort however you wish without having any effect on the real author name. This permits you to correct any guessing errors, and also lets you choose arbitrary strings for sorting.

Finally, a few people (I am one of them) choose to maintain author names in library order ('ln, fn') instead of spoken order ('fn ln'). If you choose this route, you will need to set a 'tweak' in Calibre so that it won't carefully undo your work and generate nonsense author_sort values. See this post for details on this.

Have fun.

[Starson17]

Originally Posted by chaley

When you ask for your books to be sorted by author, they are not in fact sorted by author but are instead sorted by author_sort. Author_sort is not stored the same way as author. Instead of being an individual record per author, author_sort is kept as a single string per book.

And I suppose this makes sense, as sorting is done per book, while searching is done per author.

Finally, a few people (I am one of them) choose to maintain author names in library order ('ln, fn') instead of spoken order ('fn ln').

One of the reasons there are only "a few" is that it's really hard to switch to your system. You really need to write the script that does a global copy of the author_sort field into the author field. I know, it's tricky when dealing with Esq., De, La, Jr., and multiple authors, but without an automatic tool, switching means facing a long manual editing/renaming session.

[chaley]

Originally Posted by Starson17

One of the reasons there are only "a few" is that it's really hard to switch to your system.

Ahhh, but I was lucky. I did it from the beginning, fighting against the author_sort processing until I didn't need to anymore.

A script (command line program) to copy author_sort to author wouldn't be hard to build, but I suspect that the number of users would be almost indistinguishable from zero.

[Starson17]

Originally Posted by chaley

Ahhh, but I was lucky. I did it from the beginning, fighting against the author_sort processing until I didn't need to anymore.

Yes, that's currently the easiest way - from the beginning.

A script (command line program) to copy author_sort to author wouldn't be hard to build, but I suspect that the number of users would be almost indistinguishable from zero.

Perhaps so, and since you don't need it and you are the one who would have to write it, why should you spend your time writing something that you won't use ........ Hmmmmmm ....... something about that story rings a bell

[HarryT]

Originally Posted by chaley

As Starson17 said, author names are stored as individual records. Author lists are a set of words separated by certain magic 'words' such as 'and', 'with', and '&'. Thus you should enter each author in the list as a complete name and separate the authors with one of the magic words. If you do not follow this convention, then some authors will be missing or incorrect. That may be OK (it certainly is for GRiker), or it may not be.

Thank you - that confirms what I'd determined by experiment, as I said in post #10 above.