Using a scanned dictionary with Kindle 3

[Edmundo]

I have a scanned dictionary that I like to consult. It's a huge PDF file with one image per page. On a powerful desktop machine with a big screen it's not bad to use. With a bit of practice I can find a word in it (with evince) probably a bit faster than I could with the paper version. On a cheap laptop it's usable, though rather slower than using the paper version. On a Kindle 3 (3.0.2) I can view the pages but it would probably take about 15 minutes to look something up as it takes so long to move from page to page with the slowness caused by:

1. The user interface: typing a page number is awfully slow with the "Sym" key mechanism.

2. The hardware and the software: it takes a while to render a page.

3. The screen.

So my first question is: Is there some feature of the Kindle 3's PDF reader that I don't know about that might make it easer to jump to random pages in a big PDF, other than Menu -> Go to... -> page?

I'm guessing that I will probably have to adapt my file somehow to make it easier to use on the Kindle. If at the same time I can make it easier to use on a slow laptop that would be a bonus.

Firstly, I note that the Kindle 3's PDF reader seems to ignore any bookmarks or table of contents in the PDF file. I'm not an expert on everything that's available according to the PDF standard so I might be wrong about that, but I've not yet found a PDF file with an index or table of contents that the Kindle tells me about. Can anyone confirm this?

So, what I tried first was adding some invisible text to each page of my PDF. I was hoping I could jump to a particular page by using Menu -> Search This Document. If I stick to using a PDF I'll have a file that I could use on lots of different platforms.

It works, but it's horribly slow. It seems that the Kindle doesn't compute an index of the text in a PDF file but scans through the entire file, which mostly consists of giant images. That's a shame, and something that Amazon should think about improving.

Is there a way of putting all the text in one place in the PDF file so that the Kindle's PDF reader might search it faster? Obviously I need the text to be linked to the appropriate page and the reader needs to be able to randomly access that page in the file, which is more than 200 MB.

If none of the above is possible, I might have to give up using a PDF and make a MOBI file instead. I expect that's possible, though I can see one potential problem: apparently images in a MOBI file have to be GIFs or JPEGs. My PDF uses G4 compression, which is typically much more efficient for a scan. However, within limits, I don't really care if the file is huge provided there is efficient random access to it.

What might be the best way of creating a MOBI file corresponding to a scanned dictionary?

I have the list of head words and would be prepared to create a list mapping those to page numbers if it let me use the file on my Kindle.

[tomsem]

Kindle does not recognize PDF bookmarks, but you can create 'Kindle' bookmarks for a PDF. You still have to walk through the menus to 'View My Notes & Marks' and then navigate to the bookmark in question. I often create bookmarks at the start of each chapter. If nothing else it shows relative positions of chapters in a PDF.

Search is really slow, because PDF's are not indexed. That also means that you can't search for text inside PDFs from Home screen.

Also Kindle does not support PDF links.

BTW you don't need to use the Sym key to enter numbers. The top row of ten letter keys (Q-P) corresponds to the numbers 1-9, 0. Press 'Alt' and the corresponding letter (simultaneously or in sequence) and it will generate a number.

Converting to MOBI would allow for quicker searches, since Kindle generates an index for this format.

[Edmundo]

Thanks for telling me about "Alt". I must read the documentation at some point!

I've started experimenting with converting to MOBI, using Calibre first of all. I had to modify Calibre to prevent it from turning all the GIFs into JPEGs, and I don't seem to be able to get Calibre to include an index in a MOBI file, though the same input did result in a working index in an EPUB file - working with Calibre's reader, that is.

Does anyone know about generating MOBI files with indexes for the Kindle?

I'd like to be able to search by just typing a word, without having to select "Search" from a menu first, and I don't want the words I'm searching for to appear in the text; I just want them to take me to the GIF image of the appropriate page. (Or quarter of a page. I'm going to have to split up the original pages to make the text legible without a magnifying glass.)

By the way, at some point I'd like to know how to make a dictionary for the Kindle that can be used for quickly looking up words encountered in other documents, like the English dictionaries that came with my Kindle. Does anyone know about that?

[kevinp]

Try using Mobipocket Creator. It will convert PDFs and you can select "Personal Dictionary/Glossary" under "Create New Publication". I've never converted a PDF with it, but I have created a dictionary that works with the Mobipocket Reader.

[Edmundo]

I could try Mobipocket Creator, but isn't that a Windows-only GUI program? So I'd have to obtain access to a Windows machine. In any case I'd eventually want something that is scriptable and runs on the OS of my choice, something that I can put under version control so I can get consistent results and don't have to do things manually each time.

[Edmundo]

I realised that I don't really need to create an index in the MOBI file as the Kindle does a search on the full text that is fast and easy to access, so I made a very simple HTML file and converted it to MOBI using Calibre, which I only had to modify very slightly to stop it converting all the GIFs into JPEGs. It's rough, but it's usable.