Storing the publish date for books

[tompe]

Originally Posted by kovidgoyal

Users of private collections are typically more interested in reading than in cataloging They need a tool that allows them to get a certain minimum amount of metadata reasonabl correct, with as little effort as possible. Librarians on the other hand, never actually read all their books, they spend their time making sure all the metadata associated with each book is correct.

I would say that all people I now that have book collections are very interested in cataloging it. They are interested in books and in reading and in the history of the genre they are reading and so on and they read in this context.

[kovidgoyal]

Originally Posted by tompe

I would say that all people I now that have book collections are very interested in cataloging it. They are interested in books and in reading and in the history of the genre they are reading and so on and they read in this context.

Read the part of my post in bold.

[tompe]

Originally Posted by kovidgoyal

Read the part of my post in bold.

Yes, I read it and I realize that the possibility are there. I just think you are wrong in your assumption about the difference between libraries and private book collections.

[kovidgoyal]

Originally Posted by tompe

Yes, I read it and I realize that the possibility are there. I just think you are wrong in your assumption about the difference between libraries and private book collections.

Umm let me rephrase that to be the difference between libraries and *most* private collections.

[tompe]

Originally Posted by kovidgoyal

Umm let me rephrase that to be the difference between libraries and *most* private collections.

A collection is not just a set of books. You kind of have to be interested in the meta information for it to be a book collection.

I think that either people are not interested in meta information like publisher and copyright year at all or they are interested in having a lot of meta information. It is the second category of people that usually starts collections.

[kovidgoyal]

Originally Posted by tompe

A collection is not just a set of books. You kind of have to be interested in the meta information for it to be a book collection.

I think that either people are not interested in meta information like publisher and copyright year at all or they are interested in having a lot of meta information. It is the second category of people that usually starts collections.

Yes, but "a lot of meta information" is entirely subjective. Your definition of "a lot" and my definition of "a lot" differ substantially.

[itimpi]

I am primarily using Calibre to organise and manage my eBooks to facilitate reading them. I therefore find metadata fields like series and index to be of extreme importance. Other fields like publication date are of only minor interest from that perspective. I might like a few new meatadata fields to help with organising my reading - but nowhere near as many as some of the standards that have been mentioned suggest

The thing is to strike the right balance between ease of use and completeness. If there is any conflict I am inclined to lean towards the ease-of-use side.

The other main use of Calibre is its facility to convert between the different formats used by my different ebook capable reading devices (PC, iPhone, Sony Reader, WM phone) so that I can read any particular book on whatever device is currently most convenient.

[kovidgoyal]

At any rate, here's a happy compromise, as far as date is concerned. The next release of calibre will allow you to double click the date entries to change them. This adds no complexity to the GUI but allows people to change the dates to whatever they prefer.

[daneshere]

Thanks so much!

[jowen1]

Originally Posted by kovidgoyal

The needs of a library versus a private collection are very different. calibre is meant to make creating private collections as painless and easy to use as possible. As for dublin core, in the coming digital information revolution, you will find that existing systems of classification and indexing (derived as they are from the needs of a time when "books" were produced in miniscule numbers, by a limited and reasonably well defined set of entities) to be wholly inadequate.

Dublin Core is not a system for classification and indexing, but a protocol for describing information objects. It has been developed to be object-type neutral for the very reason that object-type specific standards (such as the International Standard for Bibliographic Description ISBD) are not adequate in a digital environment where the concept of a 'book' has little or no meaning (or at least remains undefined).

The wider point is that a system such as Calibre, however useful on a personal level, has little future in a coming digital revolution where personal data will be seamlessly linked to or harvested by other users or aggregate systems (e.g. a e-book repository). The digital revolution and the concept of a (semantic or otherwise) web of information is ultimately based on descriptive standards. If Calibre does not adopt these, it will remain a great tool, absolutely isolated from the rest of the digital world.

To be fair, I must add that my 'criticism' only applies to Calibre as an e-book manager. I use Calibre and my e-reader mostly for downloading and reading news content. Calibre is excellent for this, much better than anything I have seen so far. This means that the data resides in Calibre for a few days at most. Inter-operable metadata is not an issue for this kind of use.

I even wonder if devices such as the Sony e-reader might be more successful if they were to be marketed as news-readers, together with Calibre(-like) software. But that is off-topic.

[kovidgoyal]

Originally Posted by jowen1

Dublin Core is not a system for classification and indexing, but a protocol for describing information objects.

Describing objects has only one purpose, classifying and indexing them.

The wider point is that a system such as Calibre, however useful on a personal level, has little future in a coming digital revolution where personal data will be seamlessly linked to or harvested by other users or aggregate systems (e.g. a e-book repository). The digital revolution and the concept of a (semantic or otherwise) web of information is ultimately based on descriptive standards. If Calibre does not adopt these, it will remain a great tool, absolutely isolated from the rest of the digital world.

Actually, if a descriptive standard is so cumbersome to use that no one will ever end up using it, then *it* has no future, not software that actually does something useful. And calibre already exports metadata in Dublin Core, just look at the opf files in the library folders. But if you expect calibre, or any other cataloging program to present user interfaces to get users to enter *all* the metadata elements defined in Dublin Core, you're going to be sadly disappointed.

[jowen1]

Originally Posted by kovidgoyal

And calibre already exports metadata in Dublin Core, just look at the opf files in the library folders.

OK -- problem solved!

[itimpi]

I was thinking - at the moment Calibre supports the concept of tags and you can have multiple tags associated with a book. I was wondering if this could be extended to assign a value to a particular tag instance? That would allow the tags to effectively be treated as dynamic fields extensible by the user, with tag name effectively becoming the field name. Leaving the value empty would make the tags work just as they currently do.

A possible extension to this idea might be to support extending the current search syntax to support items like:

tag:publishyear=1980

The main restriction I would see remaining in place is not being able to sort at the GUI level on tag values.

This might avoid ongoing requests for new meta-data fields.

Anyway just throwing this out as a thought - not to worried if it is not thought worth following up on.

[kovidgoyal]

From a coding perspective, doing that is not much easier than just implementing support for user specified fields. Both would require major surgery at the database and GUI levels.

[itimpi]

I thought the database surgery would be minimal - just adding a Value column to the current Books_Tags_Link table, with a default value of Null?

In terms of GUI developments I was also thinking something simple - perhaps double-clicking on a tag already assigned within the dialog where you assign tags to books.