Determining original format of book

[cybmole]

Originally Posted by kovidgoyal

For that matter it has to deal with ORIGINAL_FMT also when doing a FMT to FMT conversion.

& cope with other insane user actions , e.g. wheat should the program do/show if user deletes the first added format....
show "gone AWOL", next best choice/earliest available format....

or, if user overwrites the first added format with an ill-judged conversion from another format.....
is it still the "first added" even after being overwritten ?

this is all better done manually, methinks

and arguably the custom colunm should be source ( i.e. book store), not format. If you mess up and need to go download a fresh copy, then knowing where you bought it is more useful than knowing what format it was in ?
in most cases , name of book store implies format anyway. Amazon do not sell epub, Kobo do not sell mobi....

this is all better done manually, methinks

There certainly are a lot of issues to be addressed if such functionality were to be made generally available, but in my own case, doing it manually is inferior to doing it automatically, and I know this because of the pain I went through after switching from B&N to Amazon.

A big reason why I've never bothered trying to make this a more general solution (perhaps via a plugin) is precisely because it fits a certain workflow, i.e. mine, and would be hard to generalize. However, within that workflow, this has been an extrordinarily useful feature. If I'm reading a book that I purchased months ago and notice that there was some problem during conversion when I first imported (let's say spacing was too-aggressively removed between paragraphs, so it included scene breaks, too), I can just go to the book, hit “Convert”, make a small change, and go. Calibre does the right thing, too and uses the ORIGINAL_ format for conversion if one exists.

Basically, this tweak makes things a lot easier and less error prone for me, but that's because it's tailored for how I work.

[LadyKate]

Originally Posted by eschwartz

If a book was bought from Amazon, it will have an exth 113 field, which contains the ASIN (product identifier). When calibre imports the AZW* file, it will add this ASIN as the identifier "mobi-asin".

If a book has the mobi-asin identifier, it is a pretty good bet that the book was originally from Amazon.

Of course, I also add a tag for any book that isn't from Amazon, to indicate where it came from. (Tags are useless, so I use them to keep track of other information, like original source, if the book is an omnibus/anthology, and whether or not any metadata or editing work needs to be done.)

Out of curiosity, why do you say tags are useless? On those files that I have modified the tags to my own specifications I find tags are very useful for finding a book in the genre I want to read next.

[eschwartz]

I do use a few tags, but in order to get meaningful tags you need to add them yourself. And usually that is too much work.

[cybmole]

Depends on how often you add a new book. I average maybe one a week so the extra workload of tagging it is manageable