Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
The Internet

Open-Source Language Translator Opens For Beta 155

Posted by Hemos from the fear-the-fish dept.
mind21_98 writes "A new machine-translator designed for language translation has offically opened for public testing. GPLTrans is a translator similiar to Babelfish. Pre-alpha testing has shown that it is the most accurate of the major Web-based machine translators. More information can be found here. "
This discussion has been archived. No new comments can be posted.

Open-Source Language Translator Opens For Beta More

Open-Source Language Translator Opens For Beta

Comments Filter:
  • Or is my net connection typically slow?

  • Cool! (Score:2)

    by pb ( 1020 )
    Hopefully we'll see some better translators, because the current ones suck.

    And maybe we'll be able to add on some custom vocabulary, that would be really nice for computer journals (or chemistry, medicine, whatever...)

    ...at least the article wasn't in German, or something. :)
    ---
    pb Reply or e-mail rather than vaguely moderate [152.7.41.11].
  • ... is how babelfish might translate "First Post"

  • Re:/.'ed? (Score:1)

    by pb ( 1020 )
    It's pretty dead.

    I can ping it, but port 80 is pretty non-responsive...
    ---
    pb Reply or e-mail rather than vaguely moderate [152.7.41.11].
  • We need a web translator that accurately translates swear words, or that will at least handle "Will you please fondle my buttocks?" correctly. My nipples explode with delight!
  • It's really good to see that there is work going on to progress these type of programs. What part of the problem with babelfish [altavista.com] is that it doesn't quite get the job done. Several of my classmates have tried to cheat when writting a paper in a different language. Someone in Germany said this to me once in response to my translation...

    "I know what you say, but I don't know what you say. You funny American!"
  • "My hovercraft is full of eels" in foreign languages? Now I can find out! :)
  • I was actually just thinking about a practical way to interface to some translation software to write a real-time IRC bot to translate conversations as they happen. The only free translation software I knew of to do this was Babelfish [altavista.com], and writing an interface to that would be slow as hell for a real-time app, but this thing might be the answer. :)

  • okay its gpl'd...they're using linux...i'd think that 'skript kiddie' should definitely be a supported language!

  • Machine translators (Score:3)

    by theSheep ( 112925 ) on Sunday November 28, 1999 @07:46PM (#1498185)
    While machine translation is very practical, it can also provide entertainment. I remember a story about scientists testing an English-Russian-English translator by translating phrases to Russian and back. Input: "The spirit is willing, but the flesh is weak." Output: "The vodka is good, but the meat is rotten."

  • I need to try this at work (Score:3)

    by MonkeyPaw ( 8286 ) on Sunday November 28, 1999 @07:46PM (#1498186) Homepage
    I'll give this a test at the office,. because half the time I don't understand half of what the customers are saying.

    Perhaps I can use it to translate my words to the customer,. so when I say "Ok,. click on My Computer" they don't hear "restart the computer and click on the first icon you see while hitting the esc key and pulling on the power cord".

  • AI&Babelfish (Score:3)

    by T.Hobbes ( 101603 ) on Sunday November 28, 1999 @07:48PM (#1498187)
    I'm not sure if it has been done yet, but it would be quite helpful if an AI could 'evolve' along with the language (because, as we all know, language changes all the time) based on monitoring of user-editing of the post-process text. For example, if at time 'a' it was programmed to translate 'Cool' to 'Froid' in french, it would (after monitoring the changes made by users) learn to translate 'Cool' to the french equivilent of 'hip'. or something. 'cause, dammit, i can't wait until the AIs take over ;)
  • At one point I was going to write a program to decrypt 5kr1pt k1dd13 5p3ak. But,. figured,. no. I don't really care what they have to say,. heh heh
  • Don't you think IRC is one of the most difficult translation jobs there is? I mean, with all the abbreviations, misspellings and stuff. And few people use complete sentences at all. You would need an immense amount of knowledge gathered from following the conversation (and several at once!) to be able to get anything useful.

    Sorry, but I don't believe it's possible, even if a perfect translator for normal speech existed.

  • Make it a standard desktop component! (Score:4)

    by Anonymous Coward on Sunday November 28, 1999 @08:01PM (#1498191)
    It would be nice if someone were to make a CORBA translation service and add this to one or more of the linux desktops. Then it could be used for email, documentation, irc, coding, etc, not just for the occasional web page. It would also be good if the data at gpltrans was snapshotted regularly and pushed around, ideally so that everyone would have their own copy.

  • It's the Stamp Collector syndrome (Score:5)

    by SurfsUp ( 11523 ) on Sunday November 28, 1999 @08:05PM (#1498192)
    It's common to here the pundits opine that "open source may be good at improving 30-year-old operating systems, but the open-source model just doesn't work when it comes to large scale applications." Various reasons are given, for example: "open source programmers only do what is fun and interesting, and applications aren't interesting". But here we see yet another large-scale application falling to the barbarian hordes.

    Those pundits are wrong: there is no genre of software that the open-source model will never absorb. Simply because the open-source model results in better software, for reasons that are well-known. And no, there is no no software application that is so uninteresting that no volunteer anywhere in the world will touch it. On the contrary: the more an application area remains untouched, the more interesting it becomes to open-source programmers, simply because it's virgin territory.

    This is the "stamp collector" syndrome: when you already have a goodly number of stamps in your collection, adding the missing ones becomes an obsession.
  • This was funny. It deserves a little more than a rating of 'flamebait'

  • Right - when I was getting the idea, it wasn't for regular IRC conversations, rather the "lectures" or "meetings" that happen on IRC sometimes which are conducted with a degree of formality. In those, usually people speak in full sentences. The bot (if and when I get around to writing it) would allow someone to "listen" in any language it supported. Obviously, it would be far from perfect; it would be the equivelent of reading a babelfished web page: nowhere near perfect, but you could get the gist of what was going on.

  • Will users be able to add/update/correct translations or modify dictionaries ala the APT bot in #debian on irc.openprojects.net?

    It seems to me the growth would be incredible if users could modify the dictionary (or atleast add suggestions that could later be added by someone with the appropriate power.

  • I had an ircii (ok well it was BitchX, but probably would have worked on ircii) that worked with babelfish to translate stuff. I will try to dig it up and post a url (or you could e-mail me) It was called gtrans.bx and I dont know if anyone out there kept it up to date.. anyway it wasn't even that bad in realtime since it made a seperate connection for each translation it did.. There was some latency, sure, but it could do 10 or 20 translations at once. I suppose it could have worked on a queue system with http 1.1 to further expedite things but i really didnt get that deep with it. Anyway, it worked like this:

    /mylang
    Sets your default language (put in your startup)

    /de, /en, /es, /pt, /fr, /it
    Translates your typing into the language of choice and funnels it to the current dialog. With all of the translation commands if $mylang is set to a non-english, translation to english is done before translating to another language due to babelfish.

    /mde, /men, /mes, /mpt, /mfr, /mit
    Sends a message to in the specified language

    /flag
    Sets autotranslation of a person or a channel. This was really the coolest command. If some spanish-speaking person came in, you can just /flag juan es and everything you /msg juan was translated to spanish and everything juan said was translated from spanish. Also if you addressed juan in the channel.. eg. said "juan:" it would print in both spanish and your language.

    /trans
    Self explanitory. Output for your eyes only.

    Additionally, there were some new functions that people could use to implement their own fun foreign language commands..

    I have heard there is a babelfish library out there that provides a standard way to interface a program with babelfish. Plus, only one thing has to be updated for all of your babelfish-ized programs to work. With a client like BitchX this would be very easy to simply load and use!

    This GPLTrans thing sounds very exciting and i'd very much like to see about building a new (better) irc script on top of it!

    ~GoRK

    PS. Since the site is slashdotted, could anyone who knows please tell us a little more about it? Can we do the translation on our own hardware or is it central-server based? Can it directly translate between languages where english is neither the source nor the target? Does it provide a standard (e.g. .so loadable lib) interface for other programs to call?

    I would very much like some day to see all of my basic network communications apps (mail client, newsreader, web browser, instant messenger, irc client, etc) have the ability to machine-translate both incoming and outgoing stuff. Everyone seems to be so bent on how "good" the translation is. If a machine can translate something so that I have a basic grasp of what is going on, then the translation has been a monumental success! I would like to machine translation people focus on getting the technology more widespread before they go trying to make their software translate everything perfectly!!

    ~GoRK (again)
  • Yeah, but the people aren't much better. I've got a book at home that claims; When Pepsi got their slogan "Come Alive with Pepsi" translated into Mandarin Chinese it translated as "Pepsi brings your ancestors back from the grave" - which I actually think is a pretty good slogan...

    The problem is that we're all expecting a "Universal Translator" ah la Star Trek.
  • "First Post"->English->Italian->English->German->English- >Portuguese->English->French->English:

    "first wave of the pallet of the beginning"

  • I wonder if the open-source model for something like this could extend to the program's users as well. The idea would be that, as people used the program, it could learn from their input. Thus, every time someone inputs a new word into their local copy, this information could be replicated at some central repository and made available to other users. In fact, you could even ask the user to categorize, define, and give usage examples for each new thing.

    For that matter, you could even have the users refine the system's grammar.

    How hard would that be to implement? Is it totally far-fetched?

  • Better Context Analysis (Score:5)

    by Pingster ( 14864 ) on Sunday November 28, 1999 @08:33PM (#1498200) Homepage

    What most of these language translation programs need is a better understanding of context. I was surprised to find that Altavista's Babelfish utility has very poor analysis of context (possibly none at all). For example, when translating from English to French, "run" always translates to "exécute". For a sentence like
    The computer ran the program.
    you get
    L'ordinateur a exécuté le programme.
    ("The computer executed the program.")
    which is reasonable, but if you translate
    I ran home.
    you get
    J'ai exécuté à la maison.
    ("I executed at the house.")
    which doesn't make any sense. More incredibly, "store" always translates to "mémoire". You would think that, if they were going to force every word to be interpreted in one sense, they would choose the most common meaning. But this choice leads to insanity where
    Tom ran to the store.
    translates to
    Tom a exécuté à la mémoire.
    ("Tom executed to the memory.")

    With knowledge of context, a more advanced system could notice situations in which it was more reasonable for "run" to have a particular meaning. In the last example, "run" is followed by a prepositional phrase indicating a direction, which would imply that the meaning involving physical movement is appropriate, and so on.

    Even more revealing is the fact that the confusion of meaning happens differently for different languages. If you translate

    Tom ran to the store.
    into Spanish, you get the hilarious result:
    Tom se ejecutó al almacén.
    ("Tom executed himself to the warehouse.")
    For translation software that has multiple language targets, i would have expected it to first resolve the meaning of the English sentence into an internal semantic representation before using it to emit Spanish or French. The above would be evidence that the Systran software has no such representation -- or at least that their representation is too weak to indicate the difference between "store" as in "memory" and "store" as in a warehouse.


    -- ?!ng

  • In the context you speak of, in france, they say "cool." Not much tranlsation work there, but it'd be pretty hard for any translator to figure out which context you're talking about. For exmaple, saying "Liquid Nitrogen is cool" is either an understatement or it was from someone who enjoys pouring it over soft solids and shatting then with a hammer, yet it's a perfectly valid statement in either context. But, if you say something more obvious like "Molten Lead is cool" it's pretty easy to assume which version of cool you mean.

    I wonder how current translators solve this problem, or if they even bother. That is, where one word means different things in different contexts, but in another language, there are two different words for it, when the context can be so ambiguous that both contexts can be the same statement.

    Man's unique agony as a species consists in his perpetual conflict between the desire to stand out and the need to blend in.

  • Ooo, ooo. What about "Out of sight, out of mind" translating to "Blind Idiot"? Think about it. (I think it was an English-to-Japanese real-time voice translator that managed that effort.)
  • Finally, a project that has been needing to come around. A translator that's fast AND accurate. Best of all, it lets you correct phrases! Babelfish better stick around though.. i always get a kick out of doing things like translating

    'I like to soak my feet in gallons of whipped vanilla pudding'

    and having it finally come out as

    'I appreciate to impregnate my feet in the gallons of the pudding that I have exposed to the flash of the vaniglia.'

  • Translation methods (Score:4)

    by Y ( 13582 ) on Sunday November 28, 1999 @08:42PM (#1498204)
    Although the site has been slashdotted, it would be interesting to see what sort of algorithms it uses to perform the translations. Mmm, open source.

    I would be inclined to say that if it is based on grammar rules, the project won't make much headway - machine translation has been butting its head against this brick wall for forty years. The problem with hard-and-fast grammar rules, e.g.,

    S = NP VP
    NP = Det (Adj)* N
    VP = V (Adv)

    is that they don't account for rapid linguistic change, and people have this nasty habit of twisting grammar to express themselves in new and creative ways. :) In addition to this, it's very difficult to write simple, lucid grammar rules that also count for the myriad exceptions found in language.

    I imagine GPLTrans would probably be using some sort of probability frame of phrases and words occurring together, but one can't be sure without looking at the source. I think the best way to do translation software would be to convert the text into syntax, then into a more abstract semantic form, and from the semantic form, translate back into the target language's syntax, and then into the target language's text. Of course, the trick is to figure out just exactly how to do this. :) The parsing itself is a hefty (and not terribly exciting) task. I attempted to make a term project of a fairly basic English parser and ended up changing the project.

    My 2 cents/Pfennig/lire/pesos,
    Y
  • Woohoo!

    Now I can write back to the Mexicana Chica who works here and explain CLEARLY and CONCISELY, that whilst very attractive and nice, I can't respond to her advances because I am already "with woman".

    The last time I tried, Babelfish somehow made me inform her that I'd love to " kiss her making angry other woman".

    Muy Bien!

    Mong.

    * Paul Madley ...Student, Artist, Techie - Geek *
  • It is taken for granted that the Open Source process would take out all the bugs in this, if enough people look at the code and contribute.

    GPLTrans can be quite good, but imagine it's not (I still can't access). Let's suppose that its translation strategy is not very sophisticated and this system ends up being only marginally better than the others. Now, if somebody comes up with a great idea to improve the design of a machine translation system and wants it to be free, what is (s)he supposed to do?
    1. post it here and hope for the best ?
    2. report it as a bug fix ?
    3. do the coding and contribute a patch ?
    4. fork ?
    5. start from scratch ?
    6. try the first five options, in that order ?
    Does the outcome depend on the people running the original project?

    If they are closed to design improvements contributed by others, is their project truly Open?
  • ... how good is it at translating the GPL? Urgleburgle
  • Back in the 80s, a company produced software which they advertised with the tagline: "Finally, a machine that understands you like your mother."

    The great irony, of course, was that no machine natural language system in the world - even today - can deal with the sentence "Finally, a machine that understands you as well as your mother." (think about the possible shades of meaning)

  • Do you mean Phillip K. Dick's novel "Galactic Pot-Healer" [barnesandnoble.com]? (Stupid title, I know). In it, bored office workers sending a book title or folk saying through multiple translator machines, and challenging their friends to guess the original title.

    • Some of the examples:
    • The Cliche is Inexperinced - The Corn is Green
    • The Chesspiece made Insolvent - The Pawnbroker

    It's just called "The Game" in the book.

  • Those pundits are wrong: there is no genre of software that the open-source model will never absorb.
    I have a few F16s in my backyard, and I want their control software to be OpenSource(tm)d, 'coz I won't trust Lockheed Martin. They probably use cookies to track my flight patterns! Let's start coding! NetBSD people will port it to every piece of hardware in exsistence, from F117s to RC helicopters.

    Moderate this down, citizen.
    --

  • umm hmmm

    so linux is BETTER than windows?

    I see it as different. More stable, less ram hungry, much harder to use, many missing features.

    these things are improving, sure, but its taking a damn long time too.

    remember not everyone is techie, and the goodness of software cant be evaluated by techies alone.

    My guess is that if open-source translator is written better or not depends not so much on wether its open source, but on how talented the main contributors and/or designer(s) are.

  • there was something in a scientific american or some similar magazine about the intermediate representation idea. Someone has thought that out and actually developed such a representation, dont know if any implementations using it exist or not. definitely the right way to go.
  • Actually, I heard "Out of sight, out of mind" translated as "invisible idiot".

  • Babelfish yields some really funny stuff when English creeps into other lauguages. For instance, the English word "teenager" has crept into German; Babelfish translates it as "tea rodent". Reading this in a movie review, a room full my friends nearly died laughing.

  • My favorite example of ambiguity is from a book called "Natural Language Processing" (i think). Anyway the sentence is:

    Rice flies like sand.

    This could mean that the noun "Rice flies" enjoy sand or the the noun "Rice" flies in the same manner as sand.

  • Someone please mirror this software. No surprise their site is down if a slashdot-sized audience is trying a new translation program on one server :-) I can't get at it.

    I hope the word databases and algorithm are easily separable from the implementation. I'm sure they can't have bound it too tightly to PHP and MySQL - the presentation layer should be determined by the user, and use of other databases should be possible.

    Bruce

  • Either Bill Gates or one of his henchman is once quoted as saying something to the effect of "yeah, open source is great and all, but there are certain things that simply REQUIRE corporate backing, such as automatically translating an email message into another language." While obviously isn't the exact same thing, its pretty darn close. Anybody remember the mention of HTTP-DAV in the Halloween documents... the saga continues. If anybody can find the URL of the quote, please post it... I'm sure I saw it on Linux Today, but I can't find it readily in the search.
  • There has been some work (apparently) done in the area of an intermediate representation language for machine translation of human languages, by a team of people associated with the United Nations University [unu.edu] in Tokyo. They claim to have had a conference on the topic on November 18th, but there's no indication of progress or announcements since then on their Web page.
    --
    Paul Gillingwater
  • Maybe I can stop send letters to my french relatives that say: "I am ambiguously gay" instead of "I love my brothers" etc...


  • Just curious: Is that a reference to something?
  • But, if you say something more obvious like "Molten Lead is cool" it's pretty easy to assume which version of cool you mean.

    Even then, you couldn't be sure, because it could easily just be sarcasm.
  • Forgive me if this sounds off-topic. It's nice to see another new problem set covered by Free Software. In thinking about what can't be covered by free software, the application I focus on is TurboTax. It's the laborious product of accountants and auditors building an expert system, not really the work of programmers. It needs to be accurate enough to persuade IRS not to audit in a tremendous number of situations. It can't ever be optimal, but it shouldn't be too much worse.

    I don't think it's tenable under the Open Source paridigm. I'm sure there are other, similar examples. So, there's room for proprietary software, coexisting with free software and running on a free infrastructure. I'd just rather keep the proprietary stuff in the leaf nodes of the software "tree", where nothing else depends on it.

    Bruce

  • That's an old story about what is said to have happened with one of the first machine translation system back in the 60'ies. Don't know if it's true, or just another urban legend.
  • I don't know what's the case with Babelfish etc. but I know that at least one finnish ->english translator site has used it's logs to improve it's translations. Of course the changes have been made manually, but I see it as a good thing to see it translating something totally wrong, and after some it translates the same sentence correctly.
    --
  • You've just enumerated most of the options that are always open for any open source project. Obviously the best thing is to get involved, with code if possible, with the existing project and hope that the coordinator(s) are smart enough to recognize your contribution as valuable. If not, then you can fork or start from scratch, although at some later date the original project might choose to incorporate your changes anyway. This is precisely what happened with libc and glibc.

    Does anyone have an URL they can send that explains these issues in more detail? The question is just too broad to answer in a /. post.
  • I've studied compiler design, and I've wondered about how human languages compare to programming languages. I would think the biggest hurdle is interpreting ambiguous phrases like, 'fruit flies like a banana'. And all the implied words seem like typecasting, but are also ambigous. '(you/I/they) Come here, dammit'. But I wonder if the entire thing is more than just a really complex language description (in BNF or something) with a big database and a few enumerated phrases.
  • I heard it as the same idea, different phrase:

    "Out of sight, out of mind" [English-Russian-English] = "Invisible; insane"
  • Mon aeroglisseur est rempli d'anguilles...

    My bad spelling...
  • Now all of us German-impaired Slashdotters can
    read the c't articles.
  • Bruce, while I typically agree with your ideology, I take issue with the message you attempt to convey in your above comment:

    In America, at least, everything depends on taxes. Thus, what you wish for is impossible.
  • I think the distinction that you're looking for here is the difference between applications that are fairly open-ended with respect to the feature set versus those where the specification is as complex as the program. In the TurboTax example writing the spec from the tax code is 90% of the work, and the spec cannot be written except by experts. 90% of the work would have to be done beforehand by experts, who presumably would want to be paid. Heck, the more popular TurboTax is, the *less* work for the accountants who helped write it. Compare this to apache, where the specified behavior is fairly loose and subject to modification in many different directions at once. You might only need one person who really knows HTTP to spec that part, the rest is determined by what people want to do.

    It's probably safe to say that most systems that require more domain knowledge than programming knowledge will remain difficult to open source. Can anyone come up with an example of such a system that is an open source success?

    Walt
  • Ah yes, but do you really consider turbo tax a "software application"? Its value is not in its ability to do computation based on questions it asks you. Instead it is more of a service; you're paying for their expertise in preparing the expert system correctly for this years tax laws.

    I'd argue that we already have created this software as opensource: the web browser or other UI toolkits.

    Service will always sell.
  • In conclusion, a machine which includes/understands you love your mother.

    english to french to english on babel. Not as bad as it could have been... :)

  • When I checked it during the week-end, it looked like GPLTrans computed the identity function in all directions. I mean, when you fed it a text x in English and told it to do English->French, it'd output the same text, without any translation.

    And now their server looks like it's down...
    • Babelfish yields some really funny stuff when English creeps into other lauguages. For instance, the English word "teenager" has crept into German; Babelfish translates it as "tea rodent". Reading this in a movie review, a room full my friends nearly died laughing.

    Well even if this word has crept into the german language(we do use it), this mistake has a different origin. If you take the word "teenager" apart you have "Tee" = Tea and "Nager" (shortform of "Nagetier") = rodent.
  • There are lots of companies that make tax software, it wouldn't surprise me if one of them decided to release the core program free (maybe even open source) and just sell "form modules" for various tax situations.
  • Perhaps it would be simpler to change the US tax code. :-)

    When I worked in the US, I couldn't believe that employees would need to pay an accountant to fill their taxes. I mean, I know of no other country like that... In all the Europeans countries I know, you fill in some numbers in a form, you sign and that's it!
  • All that means is that they've perfected English->English, French->French, Spanish->Spanish, etc., etc. :)

    I wish I could look at the source, if anyone has it, post a link or something.

    Someone moderate this up, along with the (real) first post unfairly marked as redundant, and then spank the moderators for me.
    ---
    pb Reply or e-mail rather than vaguely moderate [152.7.41.11].
  • The reverse would be a lot harder. How do you deal with :

    "I clicked on the thing but it didn't work so I clicked on the other thing and it gave me some message"

    Would it be smart enough to translate "I didn't do anything" to "I didn't do anything except replace half the software on the system in a lame attempt at fixing it."

  • Re:Better Context Analysis (Score:3)

    by moore ( 3400 ) on Sunday November 28, 1999 @10:08PM (#1498242) Homepage
    The problime is that most if not all of
    these systomes know nothing about meaning at all.
    All that do is try to match one set of strings to
    a difrent set of strings.
    GPL Trans works by the substuation methoud.

    >from: Mooneer Salem
    >
    > It is a system where words in a phrase that
    > can be substituted are
    > marked by %phrase%
    > For example:
    >
    > English: My name is %phrase1%.
    > Spanish: Me llamo %phrase1%.
    >

    This genreal systome can be extended in to a
    phrase sturcture grammer with pares of rules for
    each language. ex:
    english: S -> NP1 V NP2
    irish: S -> V NP1 NP2

    these rules would modal sentences like:
    english: the cat chased the dog.
    irish: chased the cat the dog.

    All this is oversimplifyed but you get the poin.
    The real problime is that you need to be trained
    as a linguist to understand what the structer of
    many seantences are and even linguestes aruge a
    LOT. The phrase structal aprouch is probly what
    altavista a such do. All thoe I rilly like the
    idea to GPL Trans I do not thik there aproch will
    get them to far; but it will be fun to see what
    thay can do.

  • This stuff is hard (Score:3)

    by moore ( 3400 ) on Sunday November 28, 1999 @10:11PM (#1498244) Homepage
    I posted this a reply to a comment but then thought maby it should be its own thread.

    The problime is that most if not all of
    these systomes know nothing about meaning at all.
    All that do is try to match one set of strings to
    a difrent set of strings.
    GPL Trans works by the substuation methoud.

    >from: Mooneer Salem
    >
    > It is a system where words in a phrase that
    > can be substituted are
    > marked by %phrase%
    > For example:
    >
    > English: My name is %phrase1%.
    > Spanish: Me llamo %phrase1%.
    >

    This genreal systome can be extended in to a
    phrase sturcture grammer with pares of rules for
    each language. ex:
    english: S -> NP1 V NP2
    irish: S -> V NP1 NP2

    these rules would modal sentences like:
    english: the cat chased the dog.
    irish: chased the cat the dog.

    All this is oversimplifyed but you get the poin.
    The real problime is that you need to be trained
    as a linguist to understand what the structer of
    many seantences are and even linguestes aruge a
    LOT. The phrase structal aprouch is probly what
    altavista a such do. All thoe I rilly like the
    idea to GPL Trans I do not thik there aproch will
    get them to far; but it will be fun to see what
    thay can do.
  • If you really want the final frontier, think about kid's software. One of the best-selling software packages last year was a Barbie dress-up program. It's really hard to imagine Gnu 'Rugrats at the beach'.

    And you couldn't even get started if you wanted to. Trademarks are so tightly entwined with the software in that field, that it's just about impossible to Open Source anything.

    So, yes, there's plenty of room for proprietary software in the leaf nodes. It's funny, folks talk about the "desktop" as if the home market and the business workplace were similar markets. They're very different in many ways, but luckily much of the traditional home apps are moving to the web, where we can use them on decent operating systems.

    As far as TurboTax goes, an open sourced Tax program would be a great thing, since stability and lack of error is one of the major goals. I don't think it will happen though. Accountants don't rush home after work to work on personal accounting projects in the way many programmers do.
  • Not just sarcasm, but simple fallacy. Context (and non-sensicality) has to play a major role in refining things like that.
  • In order to come creature who lives with voltages he on the Pepsi-voltages!



    Woo-ee, babelfish is smoking crack tonight. It's starting to sound like a religious prophet. The Bible, by Babelfish, anyone?
    ---
    pb Reply or e-mail rather than vaguely moderate [152.7.41.11].
  • You're right, this is a big problem, and one which Cyc [cyc.com] will hopefully solve. Don't expect *that* to be Open Source anytime soon, it requires a huge amount of tedious work to make something like Cyc, so they're pretty careful about holding onto it...
    ---
    pb Reply or e-mail rather than vaguely moderate [152.7.41.11].

  • Re:Machine translators (Score:4)

    by Arjen ( 52387 ) <poutsmaNO@SPAMyahoo.com> on Sunday November 28, 1999 @10:33PM (#1498251)
    This is an urban legend. According to MACHINE TRANSLATION: An Introductory Guide [essex.ac.uk]:

    The `spirit is willing' story is amusing, and it really is a pity that it is not true. However, like most MT `howlers' it is a fabrication. In fact, for the most part, they were in circulation long before any MT system could have produced them (variants of the `spirit is willing' example can be found in the American press as early as 1956, but sadly, there does not seem to have been an MT system in America which could translate from English into Russian until much more recently --- for sound strategic reasons, work in the USA had concentrated on the translation of Russian into English, not the other way round). Of course, there are real MT howlers. Two of the nicest are the translation of French avocat (`advocate', `lawyer' or `barrister') as avocado, and the translation of Les soldats sont dans le café as The soldiers are in the coffee. However, they are not as easy to find as the reader might think, and they certainly do not show that MT is useless.

    BTW, since this book is no longer available in the stores, the whole contents is placed online [essex.ac.uk]. I recommend reading this book to anyone who is interested into the subject of MT. It really is a nice introduction into the subject.

  • There are several examples like that. I don't know the origin, but I think it was a newspaper article quite a few years ago (I have it around here.. somewhere).
    For example, when Nova (the car) was brought to Spain, it didn't sell very well since Nova (no va) translates into "doesn't go". Ford Pinto didn't fare much better; who would drive a car named "small male appendage"? Nike cought on fire (literally!) when an angry mob informed them of that "air" on one of their products was strikingly similar the arabic "Allah". Branif translated it's airline slogan, "Fly in leather" into Spanish as "Fly naked", and the most horrible error was probably some random baby food manufacturer who began selling their product in South Africa. What they didn't think of was that most products in South Africa are labeled with a picture of the food inside the container (due to illiteracy). Their product was of course labeled with a baby, since that was whom the product was supposed for. Imagine the horror -- tinned babies!?
  • I had a senior-level research project that I did on the subject of comparing a variety of language parsing systems. It was supposed to be a straight comparison of augmented transition node networks (ATNs) as compared to something else (I can't remember what).

    However, my conclusion was that each method (and there are more than two) had both its strengths and weaknesses, and no one of them was "better" than any other in general.


    I then went on to propose that the best solution would be to have a "blackboard" system, whereby you allow each parsing methodology to do what it does best and you don't try to twist each of them to handle everything, and they each contribute their own part to the mapping and parsing of the input.

    The result being that you can have multiple feedback loops, and the total output should be better than the sum of individual outputs of the various subsystems.


    It wasn't exactly the paper that had originally been envisioned, and my adviser only gave me a "B" for it. I wish I had a copy of it online, so that I could provide an URL to it. Hopefully, I've still got a floppy disk around somewhere that I could pull up that has a copy of it. If I ever manage to get a copy and put it up, I'll let you folks know.


    Anyway, it seems to me that the sort of systems that Systrans and GPLtrans have created would be ideal applications of this methodology -- take what they have now (strict sentence/phrase/word substitution, or whatever), and combine that with a system that could tag and direct the substitution based on contextual clues.

    Implemented properly, you should be able to continue to extend and improve this sort of a system pretty much indefinitely.
  • Actually the difficulties here wouldn't be technical at all. You're correct in that the lion's share of the work is in translating the tax code into directions simple enough a computer would understand, but in the US at least this is already done for you by the IRS... actually they take it a step farther and translate the tax code into directions any idiot with a GED can understand. When I was in college I translated the 1040EZ and 540E (State of California) forms into Pascal in under a week for crying out loud.

    No... the hard part about tax software isn't the code... it's the legalesee... who is Joe SixPack going to sue when GnuTax-1040 causes him to be audited? Can we get an addendum to the GPL that says if you use the results without verifying them then you use them at your own risk (oh... wait the wording on that reminded me... isn't there already a "use at your own risk" clause in the GPL?)

    Another sticky situation is the trust aspect... are people going to trust us to not collect their personal info? Lately I'm not so sure they're going to trust anyone... OSS or not. ('cause even if they *can* read the source it doesn't mean they'll understand it.)

    Being OSS also brings up another point... let's say you and I put out GnuTax and have correctly translated all the tables and formulary... then some 'leet haxor goes and patches it for something (say performance... or "privacy") and breaks the math... who's to blame? (I hate to think this way... but with something like this the blame game is going to be important... just ask Intuit's legal department).

    well... just to be on topic I was going to translate this to french or something... but the poor server is slashdotted....
  • was this one pass from english to something and back to english? or the full degenerative case of repeating until stability?

  • Re:How does it work? (Score:1)

    by Anonymous Coward
    Look here [ogi.edu]. Cool stuff. You just have to associate more info with each word (its semantic type(s), in addition to its syntactic type(s)).
  • I see it as different. More stable, less ram hungry, much harder to use, many missing features.
    "less ram hungry"???
    If you are comparing windows and Linux..then you are probably comparing the desktopedness of linux to windows (running X, wm, a net browser, etc), and Linux is very RAM hungry...netscape/mozilla both devour RAM, so does X itself.
    Now, if you are comparing Linux the server to windows the server, then ya sure Linux doesn't use half as much ram..(assuming that you aren't running X)

    sorry to pick nits.

    PS try out corel Linux...its should be pretty nice for non-techie people.
  • I was surprised to find that Altavista's Babelfish utility has very poor analysis of context (possibly none at all).

    While contextual knowledge can increase the qualitiy of a translation; the amount of world knowledge [essex.ac.uk] necessary to translate a typical web page is simply astounding. Most users of a translation system simply do not want to wait for hours to translate a simple sentence.

    And, there is the problem of linguistic knowledge. Most web pages are not written in "proper" English, but in some Web-speak-lingo. This requires the system to be very robust.

    The most successful use of MT in corporations today are situations where a very simple grammar and lexicon is used, and very little world knowledge ois required. For instance, the Xerox corporation has its own translation system that translates component manuals. The technical writers that write the original version of the manual are required to use very simple English only, without any ambiguities and with very simple constructions.

    For translation software that has multiple language targets, i would have expected it to first resolve the meaning of the English sentence into an internal semantic representation before using it to emit Spanish or French.

    This "internal semantic representation" is called an Interlingua [essex.ac.uk]. It has been used in various MT systems, with varied amounts of succes.

    The most important advantage of an Interlingua-based MT system is that is does not require a translation engine for each language pair. For instance, if you create a system for English, French, Dutch and German texts, you only need to create four analysis engines:

    1. English -> interlingua
    2. French -> interlingua
    3. German -> interlingua
    4. Dutch -> interlingua
    And four generation engines:
    1. interlingua -> English
    2. interlingua -> French
    3. interlingua -> German
    4. interlingua -> Dutch
    With a non-interlingua system (which is called a Transfer system [essex.ac.uk]), you'd have to create 3^2=9 engines:
    1. English -> French
    2. English -> German
    3. English -> Dutch
    1. French -> English
    2. French -> German
    3. French -> Dutch
    etc..

    Clearly, it is easier to integrate new languages into a interlingua system than into a transfer system.

  • Well, actually if I could really get that chance (and if the "OS" there wouldnt be hardcoded) I would *love* to toy with that; which only proves the point... :-)
  • Hello,

    Ive a masters degree in computational linguistics, and I predict this effort will totally fail. Research on automatic translation is about 40 years now and a lot of money has been spent.

    However there is still no working solutions, as problems are still far too big. Id suggest everybody participating in discussion should read a good book on linguistics.
  • If you'd read the update text at the top of the page, you'd have realised that it says "French, German and Portuguese have been added, but they currently don't do anything"!
  • Just had a quick look at the tarball. It is a 22 MB substitution phrase based DB read by a reasonably simple PHP script.

    It seems to have basis ability to correctly position Proper-Nouns using wild card characters within phrases.

    No clever grammar rules etc which is probably a good thing. Stick on a 'did this translate properly' button and let users add to the vocabulary is probably a better approach long term approach with enough users that a clever grammatical algorythm.




  • Anyone else worried by the fact that they ask for your POP3 password and sent it to their server?

    I ran the English->Spanish translation on my homepage and, although I don't speak Spanish, it is quite clear that it sucked! Much development work to be done I think. A VERY good idea in principle though.
  • Now this is what I call a powerful demonstration of the quality of open source software:
    English: "I am a small fish who wants to live in your ear."
    German: "Ich bin a small fish who wants to live in your ear."
    Astounding. I couldn't have done it better myself, and it was 6 years since I last took a German class... Wow. Also, I find this part of the Note at the bottom of each page particularily qualitative, too:
    Note: this computer-automated translation is not guranteed. It'll screw up with some text. If it does in fact screw up, first make sure you spelt everything properely.
    My note: I have mucho respect and understanding for alpha releases. It's just that I'm a nitpicking bastard, and this was quite funny. ;^)
  • If anyone wants to mirror this, please email me. I've had to point the domain to an unused IP at the moment, absolutely couldn't handle the load any longer (especially for a freely hosted user).

    Of course, it would of helped had the author (who had hours of advance notice apparently) had emailed with I or my associate that agreed to host his site letting us know he was going to be on slashdot, then arrangements could of been made much earlier. He posted a notice on his site that it was happening, but failed to notify either one of us. (Can you tell I'm not real happy with him right now?).

    So if anyone has the resources to mirror this, contact me and I'll arrange it with the author, or contact him directly and arrange it. Either way works.

    --
    William X. Walsh
    william@dso.net

  • What I have right now is v0.01-a-pre. Wanna test?

    11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.

    Moderate this down, citizen.
    --

  • Unover the true meaning of words...

    Try it with "slash dot":

    nonpersonal of the opening

    There went my karma...

    --

  • It's from an episode of "Monty Pthon's flying
    circus". It was about a hungarian phrasebook
    translating (if I recall it correctly) a
    hungarian phrase with the meaning "How can I
    get to the train station" to "My hovercraft
    is full of eels" (and other such nonsense).
  • I came across another funny Babelfish translation a while ago while reading a German article that mentioned Microsoft. The translation was pretty good until I came across the phrase "talking moon" in the middle of a sentence. It didn't make any sense at all, so I looked at the original and reallised Babelfish had translated "Redmond" as "talking moon."

    Of course, after I saw this, I remembered from my high school German than "to speak" is "reden" and "moon" is "Mond," so I can understand how Babelfish got confused ;)
  • For Pete's sake, get a spell checker, will you? Spelling is *not* supposed to be made up as you go along. I almost needed Babelfish just to read what you wrote!

    The real problime is that you need to be trained as a linguist to understand what the structer of many seantences are and even linguestes aruge a LOT.

    IMO, linguistics is just as woolly as psychology. That's why they argue; because many of the more subtle assertions about grammar that have been published aren't much more than unsubstantiated opinion.

    The human brain uses grammar up to a point, and then dispenses with it. There is no reason to expect that the grammar that has evolved in every language has to be completely regular. So you can formulate a consistent set of grammatical rules to deal with basic usage, but the more complex things get the more often the rules will be broken.

    The difference between linguistics and zoology or botany is that the latter subjects only attempt to catalogue a finite number of real living species. But when grammatical rules are flexible or disposable, the number of potential structures is almost as limitless as the number of potential utterances (which Chomsky put a number to, I seem to remember).

    In this case, beyond a small core of prescriptive grammar everything else is purely descriptive. To catalogue the resulting infinity of possible verbal blunders and call this zoo a formal grammar is pointless.

    Also, even with simple phrases you can have two different interpretations (and two complete but mutually exclusive superimposed structures) whose meaning cannot be resolved without context.

    Because of all this, a phrase structural approach, or any other rule based method is ultimately doomed. However, insofar as the linguistics community utilises Artifical Intelligence concepts (as in natural language processing studies), they are it appears still dominated by those who swear by symbolic logic.

    I'm inclined to believe that the most effective natural language parsers will always be connectionist rather than rule-based. Connection machines (such as neural nets) can encompass rule-based logic but also have the flexibility to make an "educated guess". Thus they are much more capable of parsing ungrammatical language.

    After all, our brains work the very same way when we speak or listen.

    Consciousness is not what it thinks it is
    Thought exists only as an abstraction
  • Alas, the website has been /.'ed, so I can't look at the translator, but there are some serious questions to ask.

    1 - testing: They claim to be the most accurate of the web-based translators. Based on what corpus and measured in what way? This isn't a trivial question, there are no benchmarks for translation programmes.

    2 - parsing. If this program uses American style phrase grammar, it will inevitably break down. Phrase grammar is counterintuitive and for AI purposes pretty unproductive. It is computationally simple - see Charniak's last book for good parsing algorithms - but almost certainly isn't the way humans process language.

    All of the most successful natural language translation systems are, in one way or another, dependency grammar based. Dependency based systems are also generally more portable to other languages.

    3 - morphology. English is very morphology poor. If morphology is only minimally accounted for (as a lot of poorly thought out, English based NLP systems are), I don't see how it can hope to work in Russian, or Turkish or dozens of other major languages with rich morphology. Furthermore, what kinds of morphological rules can it accept? There are languages that use prefix, postfix and infix morphology. The kinds of simple rules that can account for English will not go vert far with other languages.

    I haven't seen this program, and I don't know how seriously these issues have been considered, but they are the kinds of things to keep in mind when looking at machine translation programs.
  • Okay... consider that Windows (starting from 1.0) has been around longer than Linux (starting from 0.whatever), has been developed as a desktop OS for a MUCH longer time, and has consistently been able to stifle other OSes. Considering what it's working against, Linux as a desktop OS is progressing very quickly. Maybe it hasn't surpassed Windows for the average user yet, but it will.
    --
  • Okay. Here it is: the Moore-ish -> English translator! And it's open source!

    aproch -> approach
    problime -> problem
    systome -> system
    difrent -> different
    all thoe -> although
    linguestes -> linguists
    aruge -> argue
    substuation -> substitution
    rilly -> really

    This is obviously not complete, but hey, it's the first version :)

    The interesting thing about Moore's spelling is that he's consistent. More consistent than, (to bring it back on topic) translating from German to English.
    --

  • I just got a look at the source code. I think in a few years they might have a real translation database, but right now they only have a few hundred Spanish words and a few dozen German, French, and Portugese. It's a toy program. Not a bad place to start, but hardly worth the press release.

    Bruce

  • Some respondents have pointed out the difficulty in making translations contextually sensible ... whether 'run' should be translated as 'execute,' rather than 'quick bipedal motion.'

    I don't see an easy way to get out of this -- the needed 'world knowledge' that people have pointed out as necessary for this really is huge.

    But (and this is why I mention slashdot's metamoderation), there is a certain amount of brute-forcing which could serve as a useful basis for creating improved context interpretation. For instance, let's say you visit this translation engine and choose some text for it to translate ("Mein Hund ist in dein Aktentasche," say). At the same time, there might be a few selections of recent translations requested by others, and the resultant translations, which could be shown to you based on the languages you know. (Not telepathically ;) -- based on your own self-declaration, perhaps followed by a quiz to establish competency.)

    The resultant translations could be joined with alternate tranlations / permutations, and each reader could (say), rank-order them, or choose the best one, as far as they can determine by context, etc.

    And hopefully, the program can then be taught (wrong word, but I'm being figurative)that (anthropomorphically), something like "OK, if there are several computer-related terms in the translated text, like megabyte and power-supply, 'run' is likely to mean 'execute.' If 'run' however appears in a context which does not indicate computer use, and / or directly before the paired words 'away from,' it should probably be the bipedal-movement one. And if it's in front of a business-type name, like 'bank,' 'lemonade stand' or 'brothel,' then it is likely to mean 'manage' or 'administer."

    In my (interested but ignorant layman's) understanding of AI translators, this is the kind of discrimination that they try to make, nothing out of the ordinary. But, because words can fit into so many categories, I think this sort of gradual, piecemiel accumulation holds hope of making it work better over the long haul. It would take too many linguists to account for all the wacky ways that words get used.

    Just thoughts,

    timothy


  • I need NetBSD ported to a tank gun (a.k.a. a bazooka) so I can deal with all the obnoxious space-hogging SUVs here in the Silicon Valley. Does anyone want to pool coding resources? I'll set up a project here at http://www.bsdvssuv.org/ [bsdvssuv.org].

    Vovida, OS VoIP
    Beer recipe: free! #Source
    Cold pints: $2 #Product

  • This just goes to show, the Matrix [whatisthematrix.com] has us. It surrounds you. Everything you see, hear, feel, or taste is part of it. ;)

    I'll take a stab at your puzzle: "I toss my cookies down the toilet." Just a guess, highly dependent on humorous context. ;)

    Vovida, OS VoIP
    Beer recipe: free! #Source
    Cold pints: $2 #Product

  • Yeah, I think this is also a good idea. The problem with it is that search engines themselves can only supply answers based on statistics, not judgement. It would be useful to do a search engine search like you say, but the translator engine would have to have a good idea of what size chunks to divide the original text into.

    Anyhow, no conflict here -- I think translation engines are going to have to use a number of strategies on every input text and see which ones make the most sense in the end, then applying the information that for text-chunk X, translation X-prime (or whichever) was the best translation. That way when phrasings similar / identical to ones in text-chunk X appear again, there is at least a reference to check against.

    timothy
  • "But, if you say something more obvious like "Molten Lead is cool" it's pretty easy to assume which version of cool you mean."

    Couldn't one also use antonyms in this case. I.e. a word/phrase can be a replacement, if it is synonymous, and /not/ antonymous in the context. For example, molten describes the noun. Molten is probably also partially synonymous with "hot". Since "hot" is the antonym of "cool", in the temperature sense, then one would not use "froid" to describe it in French, but instead the appropriate term for "cool" ("cool" itself I guess), which would not be antonymous with "hot".
  • You misunderstand me. Of course grammar exists but it is not a complete, consistent logical system like mathematics (is meant to be), it is completely invented, mostly by accident.

    The result is that there are some phrase structures which you want to add to in order to complete the sentence but you can't do it without breaking the rules or generating a sentence of incomprehensible drivel.

    Most people prefer to break the rules than spout drivel, so for complex sentences in the real world, grammar often breaks down.

    BTW, It's obvious that there is an innate potential for grammar in the human brain but I don't agree with Chomsky that we are all born with the same basic grammatical structures hardwired. If you wonder how it is that so many of us end up sharing a similar meta-grammar (to coin a phrase) then you ought to read William H Calvin's book The Cerebral Code [washington.edu] (yes, the whole thing is online, thanks Prof!). He shows at the end precisely how neural structures to support basic grammar could form spontaneously to enable thoughts about who did what to whom, and with what. The same structures are probably used to generate the word order when the thought is spoken.

    You may have noticed that the higher apes (principally chimps and gorillas) used in language experiments have demonstrated the ability to form simple grammatical structures too. There were also reputedly some experiments with an African Grey parrot which demonstrated similar ability (but I've not often heard the work cited and don't know how reliable it is).

    PS. If you like Calvin's book, his latest one Lingua ex Machina [washington.edu] is all about the evolutionary development of language. Like all of his books this one's online too.


    Consciousness is not what it thinks it is
    Thought exists only as an abstraction

Slashdot Top Deals

The herd instinct among economists makes sheep look like independent thinkers.

Close