something like Calibre that doesn't change filenames and folders

[Havoc]

I know and understand the reasons why Calibre works like it does and for some things it would work fine for me. But not for everything. And keeping everything twice through exporting isn't really an alternative.

Most of my books are scans of old books in pdf. Epub isn't an option as they are full of diagrams and photo's. And the epub versions are just ocr's full of scrambled words. And I'd like to keep the filename so I can know if I already downloaded it. So I also cannot change the folders because that would make it impossible to compare without first downloading with is what I'd like to avoid.

Another part are manuals. These are sorted by manufacturer and device and I want to be able to keep them that way so I can access them without having to open another application first. The same goes for a rather large collection of datasheets of electronic components and electric standards.

Like in the manuals, there might be a repair manual that contains info about some generic part. Would be nice to be able to tag it so that a search could turn up that repair manual for some unrelated device that contains the same part. And the datasheets are refered to frm the pcb design software. So searching outside of it would be easier with some tags but re-doing the whole component database isn't an option.

But having a title that actually says something (instead of some abstract filename) and some tags would make it a lot easier. I don't mind if it would add a bunch of metadata files around in the folders.

Maybe some simple database would be a good solution but are there pre-defined versions around for this? Or at least something easy to configure. Databases are out of my comfort zone.

[canpolat]

Hi Havoc,

I'm working on something like you describe. Not being able to use my own folder structure was not an option for me either, so I started hacking something that would scan certain folders and then list the contained books in a simple manner. Since this is a side project, I cannot provide any due dates (and the first platform will be Linux), so don't hold your breath. But I will definitely announce it here when I have something to show.

For the tagging, I thought of actually embedding them in the files themselves instead of just keeping in the database (so that even if you just send the file to someone, that someone will be also able to just add it to their database and see the same tags you do).

Since I'm in the early phases of the project, I would like to hear what other requirements would be beneficial to you (and others, of course). Please feel free to suggest features I cannot promise that I will implement them, but a good idea is rather difficult to resist to

Please also state the platform you are using (Linux, Windows, Mac).

Edit: You may also want to check MikeX7's Book Pouch. It's also under development...

[Havoc]

Hello, I knew there would be others in the same position.

Honestly I think that nothing more than a specialised front-end for a database is needed. Just some random ideas:
- not limited to ebooks but able to integrate also locations of real paper books. This is what makes me think that a database is better. You cannot put the tags into paper books. Also if you want to expand to other "archivable items" like electronic drawings I'm not sure you could put tags into someting like .dxf files without lots of trouble.
- most of these programs are too obsessed with authors. I only use authors when refering to "literature". What I mean by that is "I feel like I'd like to read something of Woodhouse, maybe one of the Jeeves stories". Then I look through the authors and so further. But most of my library consists of technical reference books. So it goes more like "where do I find something about Westinghouse steam turbines?". And then I do not care about authors. This is done by tags, but never is there a hierarchy in the tags like engineering>steam>theory>turbines>types. Why not based on the Dewey Decimal System?
- integrated versioning system like svn so you can put your work in progress into it
- should also be able to work with plans, electronic bills, music scores. Why not even internet sites?

Linux is fine

[canpolat]

Interesting ideas...

- I haven't thought about adding physical books to the library. I think that can be added with minimal effort, but I need to think about it. I still prefer the files to keep the information about themselves, but in the case of physical books (and other documents) that is a challenge (let me see if I can find a way for my program to print a metadata card and put that in the book -without user interaction :P).

- Implementing the Dewey Decimal System is above me. I don't have the expertise, and I think that would be an overkill for most of the personal bookshelves. So, hierarchical tagging is not something I would go for (not from day one). But a well implemented seach function may be helpful finding books about "Westinghouse steam turbines".

- I think I didn't get your point about the versioning system. Do you want to keep several versions of a book in your folder (e.g., mybook_v1, mybook_v2) or do you want the database to keep previous version of the same book? I don't see a problem with the former. If you want the latter I will ask you to elaborate (btw: I use git on a daily basis).

- I guess if I happen to implement an option to cover physical books, it will also cover other document formats that are not easy/possible to embed metadata into.

Linux is fine

[Havoc]

Physical books could be represented by a xml file storing the info for instance. If Calibre would just store the location I gave the book in its .xml metainfo file then it would be done. Or just make symlinks to the location and store those into its filetree. If it would then use the OS and filesystem to prevent access to the original files it would be perfect. But I can't program my way out of a paper bag so delving into the code is no option for me.

The versioning system is indeed something like git, subversion (dare I say it: second source). I'm not a programmer but doing electronic design. So regulary you have datasheets that update regulary. Now only the last one is interesting, but it should be findable "on the same shelf, the same position". Wouldn't be the fist time that stuff is left out new versions or that you need to go back to find why something was done a certain way.

For reading books this is the equivalent of editions. Like what happened with THHGTTG that got changed with every edition.

I agree that Dewey might be a step to far. But you asked for ideas...

[canpolat]

Originally Posted by Havoc

Physical books could be represented by a xml file storing the info for instance.

I was thinking the same... And it could be applied to other "unsupported" file formats.

Originally Posted by Havoc

The versioning system is indeed something like git, subversion (dare I say it: second source). I'm not a programmer but doing electronic design. So regulary you have datasheets that update regulary. Now only the last one is interesting, but it should be findable "on the same shelf, the same position". Wouldn't be the fist time that stuff is left out new versions or that you need to go back to find why something was done a certain way.

For reading books this is the equivalent of editions. Like what happened with THHGTTG that got changed with every edition.

I need to think about this. I don't yet have a good idea as to how this can be solved.

Originally Posted by Havoc

I agree that Dewey might be a step to far. But you asked for ideas...

Fair enough... I just wanted to discuss them to possibly find reasonable solutions (like providing a good search instead of implementing hierarchical tagging)

[Havoc]

Tought a bit more about hierarchical tagging. It doesn't need to be hierarchical, but it can be presented as such. Suppose you have a lot of tags. Now when starting to search you enter a in a bow your first tag. Then a second box comes which only lists those tags that are used together with the first tag you enter. It would very fast narrow down to a couple of books.

[canpolat]

Originally Posted by Havoc

Tought a bit more about hierarchical tagging. It doesn't need to be hierarchical, but it can be presented as such. Suppose you have a lot of tags. Now when starting to search you enter a in a bow your first tag. Then a second box comes which only lists those tags that are used together with the first tag you enter. It would very fast narrow down to a couple of books.

I believe, something similar to this can be implemented. Another good idea...

[james320]

i am very happy today




_____________________
anosh