The Global Discussion
Several weeks ago I sat down with Nathan Lipson, a journalist from Israel, and talked to him about the market meltdown, venture capital, and what is going on in the US economy. I blogged about it the next day and that started a great discussion that currently numbers a great discussion.
Nathan posted his article sometime in the past day or two and it’s available here. Unfortunately I can’t read Hebrew and so I’ve asked a few friends of mine who do to let me know how the article came out.
There are many web page translation services on the web and I tried out a few that claimed to do hebrew to english but gave up after a few failures. I am sure that there’s a tool that would work, but I didn’t find it and like many things on the web, I gave up after a few tries.
But this experience brings up a broader issue. The readers of this blog are global. In the past month, this blog’s web pages have been visited by almost 120,000 people from 165 countries. The top six countries are all english speaking (US, UK, Canada, India, Australia, Ireland) but the rest of the top ten are France, Germany, Netherlands, and Singapore. The top fifty countries all registered 100 visitors or more and include places like Vietnam, Estonia, and Pakistan.
The posts and comments and discussion on this blog are entirely in english, which makes sense, but it does limit the discussion. It would be great if there was a "translate" button on the front page of the blog and in the comment system and people could participate in this "social media" regardless of language.
And because of the architecture of the web, those translations don’t have to be provided by Six Apart and/or Disqus.
I am wondering if there is an api version of the translation web services that I visited this morning seeking a hebrew to english translation? There should be. And it could be peer produced and open sourced like wikipedia is. We could collectively translate the web for each other. Maybe someone is doing this. If so, I’d like to learn about it. If not, let’s make it happen.
suggest you look into alfabetic (sp). Also, I think Google offers an API (widget?) for their language translation tool …Doesn’t it always suck when technology is ALMOST there, but not quite?www.twitter.com/A_F
If you still need a translation of this article I’ll be happy to help.
As an FYI, if case you haven’t seen this, one interesting similar site for crowdsourcing of video language translation (subtitling) is http://www.dotsub.com
I had a feeling that others would be doing this
Google Translate can handle Hebrew. Here’s a translated version of the article -http://translate.google.com…About as good as computer translation gets nowadays…
That’s painful readingThe technology has to get better
Google does offer an API for their translation service…although I’m not sure how good it is for Hebrew.There was a demo of adding full-translation to any page through their API at google-IO this year.
Alfabetic presented a solution like this at Techcrunch 50 this year. They create hosted, translated versions of popular blogs and hope to make money by selling ads on the international versions of those sites and split revenue with the original publisher. I believe they use real people to translate, but that the service isn’t designed to scale. Rather, they’re focused on translating just the most popular blogs. Techcrunch is one of their beta customers…http://www.alfabetic.net/In terms of self-serve API’s, Google and Yahoo! both offer machine-based translation API’s, but the translation quality isn’t very good. Amazon’s Mechanical Turk is a good middle ground. You can submit content to be translated through an API, but real people do the work.
Another vote for Mechanical Turk – it’s very well suited to this – work done in small amounts, by humans, and an API to control how it gets parcelled out and results collected. Works for me…
I work in Natural Language Processing and am sorry to say that the state of the art isn’t good enough to do what you want automatically. The result is too garbled to be informative, and too stilted to make appealing daily reading. Appealing in the sense intended by the writer at least — check out the Google translation:http://translate.google.com…Is there a decline worth of startups – recruitment new nostrils?Absolutely!
To reiterate what others have said, Google Translate offers a gadget you can drop into your blog and it will translate for 35 languages, including Hebrew and English, as best as the current technology allows. Seehttp://translate.google.com/translate_tools
ThanksIt’s not very good in hebrew but maybe better in other languages
Hi Fred! I understand exactly what you mean here, perhaps because I read and write in two languages. I think that both people and technology will bridge that gap – more people in non-English-speaking countries blog in English (a few Israelis too, e.g. Orli Yakuel), and in the meantime technology makes it easier (and more accessible) to translate whole sentences and even passages. But this gap will never be fully bridged – too few people are really multi-lingual and technology cannot capture the intricacies of natural language.
That last line may well be true. But I wish it were not so
You point to a big gap regarding the blogosphere. Having faced this for TechCrunch i can tell you there is no good solution but human. The guys at alphabetic (israeli based) may have found a solution which is partly automatic and partly human.Google translate in hebrew, like in any other language is far from accurate and brings up many nonesense. So if you have a good friend in Israel just ask him to have it translated for you or put it in a wiki and i am sure you ll get it translated pretty quickly
Thanks Ouriel. You clearly are living this on a daily basis
Surely with all of the brilliant, innovative technology, education, and linguistics companies out there, more viable solutions can be near. Perhaps there is a start up waiting to happen, that embraces this challenge. Someone has to pick up where Rosetta Stone got Fat and happy?I have seen the difference full communication has made, first hand. The sounds and site of words, exchanging our ideas should no longer prevent exactly that.Dr. Israelian learned a second language [4th really] and became a world class doctor. Her sister did not (formally), and became a seamstress with little voice.
There are a lot of problems concerning translations at e.g. Google translate. Among other, it often doesn’t detect typically Hebrew constructions of e.g. a “prefixed” preposition to a noun, and this is what renders artificial intelligence in this sector nearly unusable. I wrote a comment on this phenomenon on my own blog only some days ago (unfortunately for you it’s in… German).This said, if you’d like a usable translation of the article, I’d volunteer as well.
You are not thinking like an American (in a good sense). Few Americans care about the language issues especially from the perspective of non-Americans.I also thought about this. As fas as I know, there is no good solution for what you are looking for. I don’t think any of the existing approaches (I have researched before) has the potential to nail it down.
Yes, such software exists by one of the portfolio companies that nfactorial Software ( http://www.nfactorial-softw…) is incubating. The company name is Image Point Technologies ( http://www.imgpoint.com/ind…
Perhaps test drive this widget http://googlewebmastercentr…
Science Foundation Ireland (SFI) recently funded a large-scale effort (Academia-Industry collaboration) for ‘Next Generation Localisation’: http://www.cngl.ie. We’re just starting out, but allowing “people to participate in this ‘social media’ regardless of language” is right where we’re aiming and a lot of our recent discussions have revolved around peer production and open source (in addition to language technologies). Will keep you posted…