Creating Dictionary in Epub

[Krystl]

I searched the site here, but did not find a direct answer to my question:

Is it possible to create a "dictionary" format in epub (for sony reader) with a look-up function (using i.e. wordsearch). It does not have to connect to documents or be used from inside a different book--it would "stand alone" as a reference book. I have several personal, homemade "in house-dictionaries" with unique language needed for projects and would like to have them in my reader for more convenient transportation.

I have read that mobi has a "dictionary" format that can be used. Is there something similar for epub or a work-around?

I have considered just saving them into "A6 format" pdfs (is that the correct size? from what I read it is) hyperlinking the letters (as chapters), but the idea of a usable search would just be too perfect.
Is there a better option? Has anyone done this? (the dual column is what throws me--formatting is important here) Can anyone point me in the correct direction?

Ideas? Thank you for any information available.

[Jellby]

If it doesn't have to work as a dictionary, and you only want a list of words, then of course it's possible (and it doesn't have to be in two columns, in fact, it's better if it isn't). But if you want it to work as a dictionary, it's not possible because no such thing is defined in ePub, and each reader device or program uses a different format and method.

[Krystl]

Thank you for the info, Jellby.

Actually, I have it in a "table view" to allow easier language differences.

but what I understand is that the "search" function or something similar as a e-dictionary is not possible.

hmmm... well... guess I get to play a bit on now it would work best.

[Doitsu]

Originally Posted by Krystl

Is it possible to create a "dictionary" format in epub (for sony reader) with a look-up function (using i.e. wordsearch).

AFAIK, this would require rooting the firmware and installing custom apps.

What you can do is create an ePub dictionary with a comprehensive TOC. For an example, have a look at the free MS Reader Dictionaries. (Since they're not DRMed, you can deflate them with ConvertLIT.)

The main TOC of these dictionaries contains links for each letter of the alphabet. When you click a letter a sub TOC page is displayed that narrows the letter range further down. For an example see the screenshots of the Italian-English MS Reader dictionary.

[Krystl]

Thank you Doitsu. that is sort of what I was looking for and the alphabet TOC was what I figured I would need.

I am going to play with it for a bit and see if I can really make this work for me.

Are tables (basic list tables) horrible for epub?

i.e. If I were to import/resave the word doc table with the alphabet hyperlink TOC (I can get that far without any problem) to html (it can be saved that way from the program--simple) can I then convert it fairly cleanly via calibre? Or will the table formatting screw everything up?

Would that work or do I have to figure out sigil or some other more complex conversion tool?

Thank you for your help and karma all around.

[nrapallo]

Originally Posted by Krystl

Thank you Doitsu. that is sort of what I was looking for and the alphabet TOC was what I figured I would need.

A few years back, I created a huge epub dictionary, Webster's Dictionary 1913 and settled on a Summary TOC (just the letters ABC..) supplemented by a more detailed TOC inspired by those MS Reader Dictionaries and my own experimentation. I still continually tweak it and will update it once the epub spec incorporates better dictionary guidelines. I have my own ideas, but without ebook reader software search engine/lookup it could never be used as you would want to use it!

It was even noticed by the Sony ebook users and sticky'ed in their forum here. But that's about all the notice it got.

I am going to play with it for a bit and see if I can really make this work for me.

Are tables (basic list tables) horrible for epub?

i.e. If I were to import/resave the word doc table with the alphabet hyperlink TOC (I can get that far without any problem) to html (it can be saved that way from the program--simple) can I then convert it fairly cleanly via calibre? Or will the table formatting screw everything up?

Yes, tables would be a huge hindrance as ADE renders only the first part of long tables that spill over the page break. I would avoid it and only use lists or definitions. In fact the EPUB 3 spec was to include something to help better define the use/creation of epub dictionaries, but it was postponed until a later date.

Would that work or do I have to figure out sigil or some other more complex conversion tool?

Thank you for your help and karma all around.

Building the dictionary in html is all you really need, as afterward calibre and Sigil can convert same into an epub quite trivially.

[Doitsu]

Originally Posted by Krystl

Are tables (basic list tables) horrible for epub?

Not necessarily. It depends on how you formatted them. Why don't you simply copy a couple of rows, paste them into the default document in Sigil in Book View (Wysiwig) mode and then save it as an ePub to test it with your Sony.

Originally Posted by Krystl

i.e. If I were to import/resave the word doc table with the alphabet hyperlink TOC (I can get that far without any problem) to html (it can be saved that way from the program--simple) can I then convert it fairly cleanly via calibre? Or will the table formatting screw everything up?

Simply save your Word doc as filtered .html file, add the file to a new Sigil project and then save it as an ePub. This should give you a pretty good idea what the end result will look like. (The final file will require some fine-tuning. E.g splitting the main document into several smaller chapters, style changes, generating an .ncx file etc.)

However, if I had your reader, I'd go a different route. I'd root it and install Colordict and FBReader and/or CoolReader. Then I'd convert my dictionary to the StarDict format, which is supported by Colordict and use it with my books in FBReader or CoolReader.

BTW, creating a StarDict dicionary is rather easy, since there are several scripts available for converting a tab delimited text file to the StarDict input format and the actual StarDict dictionary generator is also available for free.
You also wouldn't have to create a TOC, because StarDict automatically generates a search index.

If you decide to go that route, you should also make sure that StarDict dictionaries for the language combination that you're interested in don't already exist, because there's no point in reinventing the wheel.

BTW, what language combination are you interested in?

[ProDigit]

not to bust your pleasure, but if you plan on having each word hyperlinked, you might end up with a 25MB file, where of 20MB is hyperlinks, and background stuff, while only 5MB is an actual dictionary.

The majority of the time you'd be hyperlinking.

[Doitsu]

Originally Posted by ProDigit

not to bust your pleasure, but if you plan on having each word hyperlinked, you might end up with a 25MB file, where of 20MB is hyperlinks, and background stuff, while only 5MB is an actual dictionary.

The majority of the time you'd be hyperlinking.

Had you read my posts more carefully and actually looked at the screen capture of the MS Reader dictionary TOC you'd have noticed that the TOC does NOT contain a link for every single word, but word ranges, which is quite manageable.

BTW, Kindle readers don't have these problems because it's very easy to create huge .mobi lookup dictionaries, which are natively supported by the Kindle.