Odyssey Bookeen looking for Dictionary Beta testers

[DuckieTigger]

Originally Posted by bob123

Better yet: http://176.31.234.27/dic/

thank you, downloading now. it is a shame they could not send an email out to the ones that did sign up letting them know this way too.

Edit: Wow, how come the It-It is only 2 mb, but the EN-EN is 32?

[DuckieTigger]

Not happy so far with the dictionary support. If a book does not have a language set, it will display the dictionary in the language your reader is set at. If you have more than one english dictionary (i downloaded all of course to see what happens) it will always choose the en-en dictionary, unless you have french selected in your language setting then it will only show the EN-FR. automatic selection based on book language (or lack thereof) will not work. automatic selection might be just fine, but there should be a option to choose a different one later. And forcing behavior by guessing how the algorithm is to get a certain dictionary is bad.

bookeens answer: oh just make sure you set the language in calibre right. boo hoo - that still wont let you select en-fr instead of en-en for instance if both are installed - only by changing the language of the reader itself. plus i havent found a way to remove a dictionary again to test out a different one. simply deleting the en-en dictionary removes look up of english books w/ english language - still won't let you go to EN-FR instead.


in a nutshell: automatic might be good so you have something to start with, but please let us customize it.

[bob123]

Originally Posted by DuckieTigger

Edit: Wow, how come the It-It is only 2 mb, but the EN-EN is 32?

Interesting question.

[DuckieTigger]

EN-EN dictionary, funny findings (looking up different words):
clicked on Kurds, the answer is:
Kurd,Kurd prop.n. A member of a people who inhabit a mountainous region of Western Asia, sometimes referred to as Kurdistan, spread over an area [...]

now Switzerland -> nothing.

Sudan -> Does not have this in the dictionary either, but it does give me options (since it was obviously misspelled ???) of (each I can click on an look the definition up): Sudamen, Sudamina, Sudarium, Sudary, Sudation, Sudatoria, Sudatories, Sudatorium, Sudatory (2x noun and adj.).

now the best thing so far:

Iraq -> does not exist, BUT it thinks Iraq and IRA have something in common:

1st IRA == individual retirement account
2nd IRA == The Irish Republican Army

Now, does that mean that once you quit the IRA, you go retire to Iraq for a 2nd chance terrorist home ????

[AlPe]

Originally Posted by DuckieTigger

Edit: Wow, how come the It-It is only 2 mb, but the EN-EN is 32?

Because the IT-IT is generated from the Italian version of the Wiktionary, while the EN-EN comes from the GCIDE and it contains a lot more stuff than the it.wiktionary.

[EowynCarter]

Originally Posted by bob123

Better yet: http://176.31.234.27/dic/

Oh, if you have to check the blog

[AlPe]

Originally Posted by DuckieTigger

EN-EN dictionary, funny findings (looking up different words):
clicked on Kurds, the answer is:
Kurd,Kurd prop.n. A member of a people who inhabit a mountainous region of Western Asia, sometimes referred to as Kurdistan, spread over an area [...]

now Switzerland -> nothing.

Sudan -> Does not have this in the dictionary either, but it does give me options (since it was obviously misspelled ???) of (each I can click on an look the definition up): Sudamen, Sudamina, Sudarium, Sudary, Sudation, Sudatoria, Sudatories, Sudatorium, Sudatory (2x noun and adj.).

now the best thing so far:

Iraq -> does not exist, BUT it thinks Iraq and IRA have something in common:

1st IRA == individual retirement account
2nd IRA == The Irish Republican Army

Now, does that mean that once you quit the IRA, you go retire to Iraq for a 2nd chance terrorist home ????

As I said in the previous post, the EN-EN dictionary is essentially the GCIDE, which is based on the freeware Webster 1913. Some country names and other "modern" terms might be missing. The Odyssey simply tries to return a list of words "similar" to the one being searched for.

[AlPe]

Originally Posted by DuckieTigger

Not happy so far with the dictionary support. If a book does not have a language set, it will display the dictionary in the language your reader is set at. If you have more than one english dictionary (i downloaded all of course to see what happens) it will always choose the en-en dictionary, unless you have french selected in your language setting then it will only show the EN-FR. automatic selection based on book language (or lack thereof) will not work. automatic selection might be just fine, but there should be a option to choose a different one later. And forcing behavior by guessing how the algorithm is to get a certain dictionary is bad.

As I said before, the Odyssey selects the language by looking at the ebook's language metadatum. That's why (as the current state of the thing) you cannot have EN-EN and, say, EN-FR at the same time. Otherwise the mechanism you wrote takes over: if there is bilingual dictionary (e.g. EN-FR) ending in your device's language (e.g. FR) then it will have priority over the monolingual (e.g. EN-EN).

So, right now, all you can do is either setting a different interface language for your device or keep copying/deleting the relevant dictionaries to/from your Odyssey.

I agree that automatic selection of the dictionary + possibility to manually change it would be a better option.

[EowynCarter]

it will always choose the en-en dictionary, unless you have french selected in your language setting then it will only show the EN-FR

Tanks for the info

[vishcompany]

Originally Posted by AlPe

As I said in the previous post, the EN-EN dictionary is essentially the GCIDE, which is based on the freeware Webster 1913. Some country names and other "modern" terms might be missing. The Odyssey simply tries to return a list of words "similar" to the one being searched for.

I had the Webster 1913 on my pocketbook and yes, it's a dictionary, but yes, it's (almost) 100 years old. I really did not like it.
I hope, bookeen uses the Webster just for testing purposes and will come up with something decent in the end. I still am interested in the Odyssey (i really like, how it handles fonts), but a proper English dictionary (and options to install others) is compulsory. I have no problem paying for a good dictionary. I really loved the PONS EN-GE I had bought for my Gen3 (mobi). It still is the best to me (also beats the Colins, which is on Sony).

Is there a tool, to transform mobi dictionaries to whatever format is required by the Odyssey?

[DuckieTigger]

Originally Posted by vishcompany

Is there a tool, to transform mobi dictionaries to whatever format is required by the Odyssey?

Unfortunately no. Mobi has a standard for dictionaries. Every reader that supports mobi also has to support dictionary lookup for those mobi dictionaries. It is beyond me why epub does not have a standardized dictionary (not sure about epub3, have not checked that out in detail).

But as a good thing, the dictionary lookup works in PDF files as well.

[HarryT]

Originally Posted by DuckieTigger

Unfortunately no. Mobi has a standard for dictionaries. Every reader that supports mobi also has to support dictionary lookup for those mobi dictionaries. It is beyond me why epub does not have a standardized dictionary (not sure about epub3, have not checked that out in detail).

No, no dictionary support in ePub 3.

I completely agree with you that it's a real shortcoming with ePub. With Mobi, you could buy your device, and then get whatever dictionary you wanted for it - and pretty much all the commercially published dictionary are available, from Oxford, Chambers, Collins, Webster, etc.

With ePub, every device maker has to reinvent the wheel, and come up with their own dictionary implementation, and only the big players - the Sonys and B&Ns of this world - can afford to license decent, modern, commercial dictionaries. The smaller players, such as Bookeen and Pocketbook, can only use what's in the public domain.

[AlPe]

Originally Posted by vishcompany

Is there a tool, to transform mobi dictionaries to whatever format is required by the Odyssey?

Actually yes, there is. I "reversed engineered" the format of Odyssey dictionaries with the help of fellows SBF-forum friends. After figuring out how it is done, I wrote my own Bash+Python scripts to convert stardict/xml dictionaries to the crazy format required by the Odyssey. Proof:



You might want to read the following SBF-forum thread, maybe using Google Translate, as it is in Italian:

http://forum.simplicissimus.it/booke...-come-crearli/

[AlPe]

Originally Posted by HarryT

With ePub, every device maker has to reinvent the wheel, and come up with their own dictionary implementation,

False. They just have to adopt some "standardized" format, like stardict or XDXF, instead of trying to play as Apple does with its devices... the "reinventing the wheel" thing is done on purpose, in this case, I suspect.

BTW, I do not think that EPUB specs should rule the dictionary mechanism. The only thing I think they should include is some sort of optional support to language tagging, so that I can have some text in, say Italian, inside an ebook that is written in English. By properly marking these "enclaves", an e-reader can offer better support to things like text-to-speech or dictionary lookups.

Originally Posted by HarryT

and only the big players - the Sonys and B&Ns of this world - can afford to license decent, modern, commercial dictionaries. The smaller players, such as Bookeen and Pocketbook, can only use what's in the public domain.

This is the main reason why Bookeen is using GPL-ed / CC-ed source.

[AlPe]

Originally Posted by vishcompany

I hope, bookeen uses the Webster just for testing purposes and will come up with something decent in the end. I still am interested in the Odyssey (i really like, how it handles fonts), but a proper English dictionary (and options do install others) is compulsory. I have no problem paying for a good dictionary. I really loved the PONS EN-GE I had bought for my Gen3 (mobi). It still is the best to me (also beats the Colins, which is on Sony).

I hope they are going to sign some agreements with some good publishers of dictionaries, so probably you are right: they are going to use these first dictionaries they published on their blog in order to "test" the response of their customers.

BTW, they have to improve their grammar algorithms a lot. Right now, they are working decently for French and partially for English, but - for example - the Italian parser lacks many features (like properly associating plural words to the root, singular word included in the dictionary)