New feature request - Sortable Dewey Metadata

[garcle]

Originally Posted by itimpi

The big problem with the Dewey system is that is hierarchical. I can see that this is necessary for physical books as they have to be filed somewhere. However that means that one is at the mercy of the person who decided what the hierarchy should be for a particular book. I much prefer the tag system as that means I can add any sort of classifications to a book, and do not have to worry about their relationship to each other. It therefore seems a much more powerful method of categorising books than Dewey provides.

Yes I agree!! .. except for one thing .. *sorting*!
Tags are great for filtering but as currently implemented in Calibre they are useless for sorting
The whole point of having Dewey implemented would be that you can "cluster" books by high level category. So that you can locate books that are related to, but not necessarily precisely, what you were looking for initially.

As for being "at the mercy of", I find that I am currently "at the mercy of" the persons that assigned tags to a book in BookThing or Isbndb. The only difference between that and Dewey is that Dewey has been assigned by people who are experts at classification. So currently I have to laboriously go thru the imported tags and remove tags that are totally useless and often just plain wrong. At least, by importing Dewey, we could start from a known reference point and add tags that are meaningful to ourselves from there.

[garcle]

Originally Posted by Manichean

I have to add my two cents to this discussion, because one point hasn't been made yet:
I'm absolutely no expert on library sorting systems, but a quick (single-digit minutes) research on Wikipedia seems to suggest to me that the Dewey system is well and good for english- language libraries, but may be too specific for other languages. If any sorting system should be integrated into Calibre, I believe it should be one that supports at least the major languages well- off the top of my head I would say english, spanish, french, german at the very least, maybe russian and portugese as well. That should cover at least a large amount of the userbase, I believe.

There are only two main classification systems used by libraries world wide. one is UDC (which is based on DCC) and the other is the Chinese system. Most libraries that hold books in multiple languages, separate books first by language, and then by classification.

[Lady Fitzgerald]

Originally Posted by garcle

Yes I agree!! .. except for one thing .. *sorting*!
Tags are great for filtering but as currently implemented in Calibre they are useless for sorting
The whole point of having Dewey implemented would be that you can "cluster" books by high level category. So that you can locate books that are related to, but not necessarily precisely, what you were looking for initially.

As for being "at the mercy of", I find that I am currently "at the mercy of" the persons that assigned tags to a book in BookThing or Isbndb. The only difference between that and Dewey is that Dewey has been assigned by people who are experts at classification. So currently I have to laboriously go thru the imported tags and remove tags that are totally useless and often just plain wrong. At least, by importing Dewey, we could start from a known reference point and add tags that are meaningful to ourselves from there.

Curious. I have no trouble sorting with tags. Have you tried using the tag browser?

[garcle]

Originally Posted by Lady Fitzgerald

Curious. I have no trouble sorting with tags. Have you tried using the tag browser?

Yes, but the tag browser is for filtering, not "sort-ordering". We may be having a semantic disconnect on this basic term. When I use sorting I mean in the "sequencing" sense of the word (for example sort into alphabetical or numeric order), not in the selecting sense (for example sort into red and blue). The term sorting is broadly used in database technology in the sequencing sense not in the selection sense.

To put some context to this my non fiction library has about 8250 titles in it. On average, each Title has 9 tags. Currently, the tag field in the book list is ordered alphabetically. So of the 9 tags per book, the books will be clustered by whichever tag is first alphabetically (if you click the tag column heading). The problem is, is that the first tag alphabetically, may be my 9th-least-important tag hierarchically. So i get books clumped together by, possibly, the least important tag in the list of tags for those books.