Implementing large, specialized dictionary

[two_wheels]

I am trying to create an ePub book rendition of a large, specialized dictionary. I initially attempted to create it as an instance of a generic ePub book. At 40,000 plus entries, with other supplementary material and images, the book comes in over 1.5 GB and simply isn't functional in any reader that I've tested it in. I then turned my attention to the dictionary format/application, thinking, perhaps, that it would be possible to capture the dictionary as an ePub dictionary, and create a simple, ePub book/document as a portal to the dictionary. The idea was that I could somehow embed lookups to dictionary terms in the portal, which would then allow me to access the dictionary that way. If that is even possible is a problem I've struggled to resolve, however, simply because of the scarcity of demonstrated examples of epub dictionary implementation on the web. Or else I'm just looking in the wrong place.

I would appreciate any guidance or recommendation.

[rkomar]

I don't think that epubs are a good fit with dictionaries. They don't differentiate between any of the contents, making searching more difficult (you'll be searching the definitions as well as the words being defined). Generally, electronic dictionaries are created using some specific structure that a dictionary application understands (e.g. xdxf or sdict markup). You would probably be better off creating such a structured dictionary, and then converting that into whatever format is needed by your device.

I had a similar problem with a 1600 page dictionary I scanned for personal use. I created a PDF file of the scanned images, and used PDFMARKs to create an index for the dictionary. At the top level, I had a bookmark for each letter of the alphabet. Next level down, I grouped eight pages together, labeling them with the first and last word in each grouping. At the final level, I had a bookmark for each page, labeling it with the first and last word on that page. That way, I could find the location for each word with three clicks through the index. Obviously, that's a lot of work (almost 1800 bookmarks in my case), but it is worth it if the dictionary is used extensively afterwards.

If you insist on epub, maybe you can divide the contents in a similar way that would make navigation easier.

[fbrzvnrnd]

mobipocket had support for dictionary. ePub hasn't (yet).
You can find information about future working about dictionary and EPUB3 here:

http://www.idpf.org/epub/dict/

[rkomar]

Interesting. Is there any application that understands or plans to understand such dictionaries?

[fbrzvnrnd]

afaik: none. You can find some 2015 update here:

http://www.idpf.org/epub/dict/roadmap.html

[Doitsu]

@two_wheels: IMHO, the best approach would be to offer StarDict and Kindle dictionaries. This should cover most platforms and for each format at least one free reading app exists. Also many Android epub reading apps have hooks for third-party dictionary apps.

Since StarDict dictionaries can't be DRM-protected, you'll have to sell them on your own website or include a password-protected download link in the epub.

Depending on the bookstore, you also might be able to create an epub3 book with JavaScript that searches an embedded sqlite database.

I toyed with this idea some time ago and created a proof-of-concept epub3 book that allows you to search a 10 word English-French word list.

Unfortunately, it only works with the iOS and Android versions of ADE 4.5 and Azardi. (It doesn't work with Sigil or Calibre Editor.)
However, since I only have rudimentary JavaScript skills there might be a bug in one of the JavaScript files.