Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Google Dictionary (google.com)
106 points by chanux on Dec 4, 2009 | hide | past | favorite | 69 comments


Good dictionaries are hand-compiled by experienced lexicographers. Not Google's strength, to put it mildly.

What Google would be extremely good at is creating corpora for all these languages. An example of an excellent Spanish corpus is the following: http://www.corpusdelespanol.org/x.asp


Good encyclopedias are compiled by experienced... oh wait.

No idea if this specific effort is any good, but Google have done a lot of interesting things with language already.


Good encyclopedias are written by large numbers of individual contributors with expertise in specialized fields, who barely get paid for their efforts, most of the money from encyclopedias going to pay for the organization, compilation, editing, etc. It turns out that the latter can now be much more effectively decentralized than ever before, and contributors no longer need to be explicitly recruited. (which is certainly not to say that Wikipedia isn’t dramatically different from other encyclopedias)

Dictionaries are still compiled by experienced lexicographers.

* * *

Incidentally, does anyone know of a good online thesaurus, now that thesaurus.com is no longer using decent source material (after being hit with a lawsuit for essentially stealing it). All the ones online seem to use Roget’s II, which is vastly inferior to the original Roget’s.


It depends on how we're [re]defining "dictionary." A dictionary has the potential to be a set of pairs of words, and any of the following:

1. how the word is used,

2. what the consensus is on how the word is used,

3. what the consensus is on how we 'should' use the word,

4. what the consensus of people who are, by consensus, considered "experienced lexicographers", is, on how the word is used,

5. and what the consensus of people who are, by consensus, considered "experienced lexicographers," is, on how the word 'should' be used.

Google is trying the first approach, defining words by trying to learn them itself, like a person who naturally "absorbs" a word by hearing it used in the right context enough. The second is how people usually learn words, by asking others what they mean when they hear them. The third is prescriptivism, but ruled by the majority: words such as "ur" 'should' be valid, and so on, in the majority's opinion. The forth is what dictionaries try to contain, and the fifth is what books of "advice" like Strunk and White contain. All are different, though some overlap (as, e.g., 3 may be influenced by 5.)


I don't see how Google does that. I see Google collecting online all the online dictionaries they can find, and uniting them under a single interface, but I haven't found any dictionaries where they have contributed content.

I think they should do that, though.


You mean like Wiktionary?


I have heard about Google using lot of local language expertise for things like validating the ranks for local language search results. So, this might also have been vetted by such a group.


A gringo collecting words for an "Espanol" dictionary. Hilarious.


A lexicographer should be a native speaker. For a corpus-compiler, there's no reason for such a requirement.


It's great that they used definitions from the Collins COBUILD Advanced Learner's English Dictionary (http://www.antimoon.com/how/cobuild-review.htm). Collins was the pioneer of corpus-based full-sentence definitions back in the 90s. Nowadays, there are better choices for an English learner (e.g. the Longman Dictionary of Contemporary English), but COBUILD is still miles ahead of open projects like Wiktionary.

Overall, good news for English learners in the short term, but not so much in the long run (who's going to make dictionaries if you can no longer sell them for money?).


On the off chance, does anyone have recommendations for similar resources in German, French, and/or Spanish?


For translations to/from German from/to (English, Spanish, etc), dict.leo.org is pretty decent. Not sure that's what you're looking for.


Thanks for the recommendation. I've used LEO. However, I was looking for something more comprehensive, (perhaps more authoritative), and also in a more readable format.



I've always typed "define: <word>", but this seems to be superior.


It would be nice if there were a way to get straight to it from a regular search though. It is still more convenient for me (in Firefox) to hit Command+k and type "define:".

Also, it'd be nice if Google could just read my mind instead of me having to laboriously type in search queries.


1. open the link

2. right-click on the search box, select "add a keyword for this search"

3. for "keyword: " type "def" , for "name: " type "Google Dictionary Search"

Now in your address bar, type "def uniquity"

While you're at it, replace all your search engines with this method and remove your search box. optional: install tiny-menu, move everything from your navigation-toolbar to your top toolbar, hide navigation-bar.


That sounds awesome, but it doesn't work for me. I did all that and yet FF just does an "I'm feeling lucky" for whatever I type in.

Edit: It only works when I add keywords to the already installed search engines.


Ignore the above comment. I must have been up way too late last night. Thanks for that piece of info.


To add any search to the search bar in Firefox:

- install addon Add to Search Bar

https://addons.mozilla.org/en-US/firefox/addon/3682

Then right click in the search page's search field and choose "Add to search bar ..."


Actually, I already know about that. What I wanted you to tell me (because I was too lazy to find it out for myself) was that I can change the search engine used via the Command+Up/Down keys.

Still, those arrow keys are all the way down there and I can type "define:" pretty fast...


This looks similar to the define: results, but juiced up on steroids. Maybe at some point this will be what you get when you do a define: search?


For some test English words, it doesn't look any better than Wiktionary:

http://en.wiktionary.org/wiki/pants http://en.wiktionary.org/wiki/ostensibly

http://www.google.com/dictionary?aq=f&langpair=en|en&... http://www.google.com/dictionary?aq=f&langpair=en|en&...

As a plus, Wiktionary lets you download their data :). If Google opens this up, then it could get pretty interesting


I'm learning Dutch, and it's even worse. Really basic, commonly used words aren't there.

Good dictionaries are hard to build. At the same time, the web isn't constrained by space the same way a physical book is, which means that a result should have even more information. You shouldn't accept an online dictionary that gives you only one simple usage of a word.

A good dictionary will list the short & sweet definition, alternative definitions, forms, example usages, pronunciation, history, usages in various situations/jargon, idioms, and word history. This dictionary just points you to a wiki.


I'm Dutch, and I use the electronic Van Dale Groot Woordenboek (€99). Google will give you wiktionary for the word "slim" ( http://nl.wikipedia.org/wiki/Slim )

Van Dale gives you this: http://dl.dropbox.com/u/381493/van_dale_slim.jpeg


You should really check out this website, which I love with all my Dutch-learning heart.

http://www.mijnwoordenboek.nl/

"Slim" isn't a very good example for that dictionary, but check out something with a lot of possible EN-NL translations, usages, and jargon.

http://www.mijnwoordenboek.nl/vertalen.php?src=NL&des=EN...

That site refers you to "Babylon 8", another installable dictionary which just came out with a Mac version. I'm not sure if it returns the same results as the site, but if it does I'd say it's a pretty good competitor for the VDGW (which is truly fantastic, itself).


Now we need an API for this, please!


Google's webpages are really easy to scrape. lxml, django & 40 lines later and you have a really nice json formatted data stream


I would be almost useless without API..


Don't be so hard on yourself, I'm sure a lot of nice people like your regular UI just fine.


I think you can do the dictionary from Google Language API?


"hackers plural

1. A computer hacker is someone who tries to break into computer systems, especially in order to get secret information.

2. A computer hacker is someone who uses a computer a lot, especially so much that they have no time to do anything else."

http://www.google.com/dictionary?langpair=en%7Cen&q=hack...

I guess we're number two?


Contrast this with the Oxford English Dictionary's entry for Hacker, noun:

  1.  One who hacks; one who hoes with a hack.
  2.  That which hacks; an implement for hacking, chopping
      wood, or breaking up earth; a chopper, cleaver; a hoe,
      mattock.
  3a. A person with an enthusiasm for programming or using
      computers as an end in itself. colloq. (orig. U.S.).
  3b. A person who uses his skill with computers to try to
      gain unauthorized access to computer files or
      networks. colloq.
One of the quotations listed under 3b. is also notable:

     1983 Byte May 298/1 ‘Hacker’ seems to have originated at
     MIT. The original German/Yiddish expression referred to
     someone so inept as to make furniture with an axe, but
     somehow the meaning has been twisted so that it now
     generally connotes someone obsessed with programming and
     computers but possessing a fair degree of skill and
     competence.


Tried the first word that randomly popped into my head.

panoply /p'ænəpli/ Synonyms: noun: armour, armor

Huh? Since when are "panoply" and "armor" synonyms?


From dictionary.com:

1. a wide-ranging and impressive array or display: the dazzling panoply of the maharaja's procession; the panoply of European history.

2. a complete suit of armor.

3. a protective covering.

4. full ceremonial attire or paraphernalia; special dress and equipment.


The word "panoply" derives from a compound Greek word, "pan" = "all" and "hoply," from the "hoplite" or "armed soldier" of ancient times. Hence, the original idea behind the word is that of being "fully (or heavily) armed."


yes. Words are defined by their usage... even if the original meaning was different (or even opposite) to to current usage.

"Hacker" has come to primarily mean #1 ... only a small subset of technical people and academics still use it like #2. Similar to the way "Begs the question" has changed in the last few years to mean "directly leads to us asking".


"Words are defined by their usage."

Indeed, definitions are like history. It doesn't always work out the way one might like it to.


Mac OS X has an excellent dictionary app, highlight any word, and from the app menu select Services -> Look Up in Dictionary. Even works when you're offline :-)

Seems to me this used to be bound to a keyboard shortcut back in the NEXTSTEP OS, though it's possible to add one in the current Mac OS.


Or even better: hold down Ctrl-Cmd-D and hover over any word.

Works in Cocoa apps that use native text controls -- Safari, Mail, etc. all work.


I tried "the" first, and then "friend".

It looks like it's generated from statistics plus a language model plus the contents of the web (possibly and/or google books).

If that's the goal it looks like they're doing some great work. If being a dictionary is the goal it looks like there are currently better solutions.

However a) give it a few years and b) if they're able to extract that amount of information from their corpus, just imagine how that could be applied to a search engine - I get the feeling that they're showing their hand a little in their progress on the natural language processing front.

If that's so then I can't wait to see them apply it to search.


What you're seeing for English is the Collins Cobuild dictionary.


Clean interface and ability to save favorite words, nice but after a few tests i'm not too confident in the quality of the translation (e.g. some related phrases with eng<->ita are just wrong). A good start,btw.


I wish there was Latin too.


You could always get this exact functionality by searching for "define: word". But still, I like that they have a wrapper interface now.

For times when my word choice matters greatly I still refer to the Oxford English Dictionary, however. As someone else noted, there's a far cry between Google Dictionary and a hand-compiled one.


I recommend taking a look at http://www.wordnik.com. It's a startup launched by an experienced dictionary editor (sorry, Erin, I don't know what else to refer to the profession as:)) and they have an API as well.


That's great! I've usually used answers.com or dictionary.com, but they're slow and cluttered compared to this. Also, answers.com often gives you a Wikipedia entry when you want a simple dictionary definition.


This still seems a bit cluttered to me and as usual it doesn't even look like the visual design was given any though below the page header.


how is this cluttered when there is basically one drop down, one text input, and one button. Go to homepage of answers.com and they got a bunch of crap on it.


I've been using ninjawords[1] for awhile now. Looks like they have pretty similar source material, but ninjawords actually feels much faster.

[1] http://ninjawords.com/



All things considered, it could have been a lot worse: http://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffalo...


The interface is uniform, but the quality is highly variable. There are some excellent electronics dictionaries on the market, but generally not for free on the web.


It would be great if "define: WORD" returns the definitions. It appears to be more better than Wordnetweb's results.


You can actually do that from the Google search box


But, a query of [define:WORD] only shows 'web definitions' for WORD. It doesn't show these new dictionary definitions, and it doesn't even include the handy top-right-header link to dictionary definitions that a normal query does!


i can't help but wonder, if anyone else except from me was using it for at least last few weeks? otherwise, it means i had breaking news at hands without realizing it :(


Someone pointed it out to me at least a month ago as well. I'm not sure why it is getting press now...

Maybe this is just the power of marketing? I imagine usage was fairly low before people started talking about it.


i believe this is because they started including this in some search results and it got attention...

anyway, at least i don't feel bad anymore, thanks :-))


This is pretty cool.. Hope they make a mobile version soon.


I hope an Android client is in the works.


no japanese? curious to know why.


日本語わ、むぞかしですね。。。


日本語は actually.

Pronounced wa, but spelled ha.


needs audio



English <-> Spanish has it too.


I'm liking the interface, they even include IPA transcriptions!

I bet this will get most online and many offline dictionaries out of business.


they even include IPA transcriptions

This is very good, as it will promote learning the International Phonetic Alphabet.

http://en.wikipedia.org/wiki/International_Phonetic_Alphabet




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: