Thanks for the reply,
CWatkinsNash.
While I'm waiting for someone to come up with an automated way to do this, I have been re-reading a fascinating article by Clay Shirky about ontology versus crowd-sourced tags, entitled "
Ontology is Overrated: Categories, Links, and Tags". I'm not sure that Shirky's conclusion, that crowd-sourced tags are better than a fixed ontology, is applicable to my own personal ebook library. Here Shirky is concentrating more on collections that are, essentially, publicly owned or publicly accessible. My ebook collection is my own, and is not accessible to anyone but myself. In these days of DRM and anti-piracy legislation, I may well publicize my catalog, but I will not share my content. ;-)
Therefore crowd-sourced tags are great as a suggestion only, and I like your idea above,
CWatkinsNash, of verifying tags against the content of the book, but I think I need to retain ownership of the metadata for my comparatively small collection.
In practical terms, what this means is probably creating duplicate tags, but prefixing them with something, such as "gsafd.Humour" and "gsafd.Science Fiction", and using these tags exclusively for my own library, eventually deleting any tag that doesn't conform to my own standard. Once they are all processed, I can go back and remove the prefix from all of them if I choose to. Again, it would be great if there was an automated way to achieve this. I'm facing many hours of unpaid work to catalog my ebook library to a self-imposed standard. Welcome to the wonderful world of Asperger's. Haha.
Next problem: GSAFD tags only exist for fictional works, which is what it is designed for. What happens when I come across a non-fiction book? Should I revert to using Dewey's categories? Here again though, as Shirky points out, Dewey is only designed to categorize a book into a single category...
Quote:
|
It isn't the ideas in a book that have to be in one place -- a book can be about several things at once. It is the book itself, the physical fact of the bound object, that has to be one place, and if it's one place, it can't also be in another place. And this in turn means that a book has to be declared to be about some main thing. A book which is equally about two things breaks the 'be in one place' requirement, so each book needs to be declared to about one thing more than others, regardless of its actual contents.
|
So, I could just use multiple categories sourced from the Dewey Decimal System, as an authority, and compare those to the crowd-sourced tags. I could then add a GSAFD tag for "Non-Fiction", and my tag list for a particular book could end up looking like this...
gsafd.|Non-Fiction
nf.|Endangered Species
nf.|Kakapo
nf.|Nature
nf.|New Zealand
nf.|Science
My prefixes could be anything, as long as I standardize them. In the example above I have used "nf" as a prefix, meaning "Non-Fiction". Once I have processed all my books, I can remove these prefixes if I want to anyway, or they may be useful to retain in the event that I add more ebooks to my library later, and have to categorize new books using crowd-sourced metadata again.
Sorry if this all seems very wordy. I'm getting my thoughts out in here and asking for input at the same time.