Best custom dictionary? (English)

[BaronMoe]

I noticed that the dictionary on Kindles are far superior to the Kobo one. I noticed that a lot of dictionaries are floating around the forum, but does anyone know which is the most comprehensive?

[anacreon]

Look at the thread "index of custom dictionaries".
In my opinion, the most comprehensive English dictionary available there is Owl's (some 56 Mo). The best for syntax and thesaurus is Longman's, which I add to my kobo by disguising it as a dictionary for one of the languages I don't use.

[rtiangha]

The most comprehensive commercial English dictionary is the Oxford English Dictionary (2nd edition is 20 volumes and updating it to the 3rd edition started in 1994 and has a deadline to finish in 2037) and that's what Kindle uses (by default, I think?).

In terms of free dictionaries, the GCIDE takes the copyright free Webster's 1912 dictionary as its base and uses crowdsourcing and adds various Wordnet entries to modernize it a bit. A recent copy of the English Wiktionary would have more entries and more modern entries, though (although some would debate the quality of those definitions). So it depends on your definition of comprehensive.

Personally, I use a stylized version of Wordnet 3.1 combined with GCIDE definitions as my primary English lookup dictionary. It covers most of what I read, the Wordnet-specific definitions are concise but descriptive enough (with the Websters 1912/GCIDE entries where available providing more context if I need it), and some of the Wordnet entries list synonyms, antonyms and hypernyms which is something some commercial dictionaries don't even do.

You could also use penelope to combine a bunch of dictionaries to create your own custom one (for example, you could merge an English dictionary of your choice with a thesaurus to mimic what Wordnet offers, etc.). So there are lots of options.

[snarkophilus]

Originally Posted by rtiangha

Personally, I use a stylized version of Wordnet 3.1 combined with GCIDE definitions as my primary English lookup dictionary.

Is this the same dictionary you posted called "GNU CIDE v0.52" in the Index of Custom Dictionaries for Kobo eReader thread or something different again? Curious to give this a go.

[robp25]

try this one-

https://www.mobileread.com/forums/sh...&postcount=567

Please share if you find a better one

[rtiangha]

Originally Posted by snarkophilus

Is this the same dictionary you posted called "GNU CIDE v0.52" in the Index of Custom Dictionaries for Kobo eReader thread or something different again? Curious to give this a go.

Not exactly. I took that GCIDE version I made and combined it with a different version of Wordnet 3.1 that I found on a Chinese forum (here's the file, but I didn't make it) that had some HTML attached that made the entries look nicer (I didn't know that Kobo dictionaries could be styled like that; it looks really nice when paired with a sans font at a smaller font size using patches). I think it's a conversion of a styled mdict version that I've seen kicking around (although the .ifo file makes mention of babylon so maybe that was the original source format; either way the ifo file has the credits but it's written in Vietnamese. As an aside, the mdict format allows you to attach a css file and when you use pyglossary to convert it to stardict format, it converts that css into inline css per definition, which gets preserved when you use penelope to convert that into Kobo format. It works most of the time without any need for editing so now I try to find mdict versions of things rather than stardict and do that two-point conversion, because there's a higher likelihood that the person that made that mdict dictionary put some effort into making it look nice too.).

I also combined it all with Moby Thesaurus II because I really wanted some kind of thesaurus to do look ups with but didn't want to make a new custom dictionary entry just for that (in order to keep the dictionary menu as lean as possible). As penelope presents results in the order you merge definitions in, the order I used is Wordnet 3.1, GCIDE, Moby Thesaurus (that way, I get the concise Wordnet 3.1 definition first, can scroll backwards to get the Thesaurus entry, or scroll forwards to get the more detailed GCIDE entry if I want it). And all done with only free dictionaries too. To me, that's comprehensive enough.

[skybook]

Would you consider sharing your version?

[rtiangha]

Originally Posted by skybook

Would you consider sharing your version?

Better yet, you can build it yourself!

Grab these versions of Wordnet 3.1, GCIDE 0.52 and Moby Thesaurus II and run:

penelope -i stardict-wordnet3.1-custom.zip,stardict-en-gcide-0.52.zip,stardict-Moby_Thesaurus-2.4.2.zip -j stardict -f en -t en --ignore-case --merge-definitions --sort-by-headword --sort-before --sort-after --merge-separator "<br /><br /><hr /><br /><br />" -p kobo -o dicthtml

[snarkophilus]

Originally Posted by rtiangha

Better yet, you can build it yourself!

Hah, thanks! I was working on figuring out the penelope options myself.

I'll try to finish without looking at your command line.

[rtiangha]

Originally Posted by snarkophilus

Hah, thanks! I was working on figuring out the penelope options myself.

I'll try to finish without looking at your command line.

I think it should work, but I'm going off of memory (not sure if --sort-before is really needed if you have --sort-after, but it still works so whatever). I can't remember if I ran --ignore-synonyms or not. I'm fairly certain I didn't run --flatten-synonyms on the various dictionaries first, but perhaps I did (it's been a while)?? Hmm...

I just realized that I uploaded the wrong version of Moby Thesaurus; it's the stock version and not the one I ended up using in my final version. I basically did some very minor edits to bold the number of entries and italicize the headword. I've updated the file (link remains the same) so if you think the stock thesaurus looks a little boring, try re-downloading the file and try again (It's not that much better, but at least it looks different).

And the GCIDE version in my build directory is slightly newer too, so in case I did run --flatten-synonyms on that dictionary before merging, I also uploaded that version too (same link); I'm not sure if it matters though (from what I can tell, flattening synonyms first may be needed because penelope doesn't migrate stardict tagged synonyms (i.e. <sym>) to Kobo dictionary synonym format (i.e. <var>) at all, so you may miss out on some definitions when searching if they're tagged as synonyms to a different definition as Kobo won't search them out and will only return strict headword definitions. That said, I can't remember if that GCIDE dictionary actually implements tagged synonym functionality).

I also think it all looks better with a sans-serif font rather than the stock Georgia font (I used the "Dictionary text font-family/font-size/line-height" libnickel patch to replace Georgia with generic sans-serif (which currently defaults to Avenir Next) and shrunk the font to 23px and decreased the line spacing to 1.30em, which looks fine on my Glo HD but your miliage may vary depending on your screen size and resolution), but it's up to you.

[Owl_]

Originally Posted by BaronMoe

I noticed that the dictionary on Kindles are far superior to the Kobo one. I noticed that a lot of dictionaries are floating around the forum, but does anyone know which is the most comprehensive?

If you mean with the biggest number of words in the database, you can try mine

https://www.mobileread.com/forums/sh...&postcount=507

enjoy the reading

[skybook]

Originally Posted by rtiangha

Better yet, you can build it yourself!

Grab these versions of Wordnet 3.1, GCIDE 0.52 and Moby Thesaurus II and run:

penelope -i stardict-wordnet3.1-custom.zip,stardict-en-gcide-0.52.zip,stardict-Moby_Thesaurus-2.4.2.zip -j stardict -f en -t en --ignore-case --merge-definitions --sort-by-headword --sort-before --sort-after --merge-separator "<br /><br /><hr /><br /><br />" -p kobo -o dicthtml

Oh no, I just downloaded the 3 resources you posted, and tried converting them using the command you gave, but the resulting dicthtml file doesn't install properly :/

I guess I'll just continue to use the WordNet 2.0 file I found (I really like the formatting, and so far seems fairly comprehensive ) along with Owl's large dictionary (thank you Owl!)

EDIT: until geek1011 finishes dictutil, ofc

[mbovenka]

Originally Posted by rtiangha

The most comprehensive commercial English dictionary is the Oxford English Dictionary (2nd edition is 20 volumes and updating it to the 3rd edition started in 1994 and has a deadline to finish in 2037) and that's what Kindle uses (by default, I think?).

The Kindle doesn't use the full OED, it has (IIRC) the New Oxford American Dictionary.

[rtiangha]

Originally Posted by skybook

Oh no, I just downloaded the 3 resources you posted, and tried converting them using the command you gave, but the resulting dicthtml file doesn't install properly :/

I guess I'll just continue to use the WordNet 2.0 file I found (I really like the formatting, and so far seems fairly comprehensive ) along with Owl's large dictionary (thank you Owl!)

EDIT: until geek1011 finishes dictutil, ofc

File.

[skybook]

Originally Posted by rtiangha

File.

Thank you for the file! Yeah, I must have done something wrong, because my version of this file is only 50 458 624 bytes, vs yours which is 95 195 136 bytes :/