Hacker News new | past | comments | ask | show | jobs | submit login
Google to offer for-pay Translate API (googlecode.blogspot.com)
145 points by rryan on June 3, 2011 | hide | past | web | favorite | 40 comments

If they had just announced they would be charging for it people would also be upset. By announcing they're shutting it down, then backpedaling a little by offering to charge for it developers can't be as upset. Genius!

More to that, now they are the good guys who read all the comments, cared about their users, listened to their opinions and are changing their decisions based on those comments. Agreed, Ingenious.

did you even consider that this was genuine?

Maybe they didn't think anyone would pay for it.

I was actually expecting the whole "shutting down" announcement to be a Lean Startup-esque experiment. If it is, I am really glad, because I'd rather that companies (startups or not) focus on things that people really want.

I can't help but imagine all the good Google could have done if it had done this go Wave (either by pivoting it into a better product, or by leaping the whole team into greener fields sooner).

It's also a very good way to figure out how many people would actually be willing to pay for it. Not the best way, obviously, due to the number of irritated developers, but it might provide more accurate results than a survey, and certainly gets people to speak up.

Thanks for listening! (I'm assuming some relevant Google folk will be passing by here.)

A personal qualification: I may be unusual, but my personal concern (as opposed to professional needs) is use by tools such as the Firefox Babelfish extension.

At times, I spend significant time in communities speaking any number of languages. Being able to drag-to-translate various items, pretty much instantaneously (and without forcing the entire page through a single translation), makes participation fluid. It also props up my sometimes dormant, if not non-existent, understanding of those languages.

I hope there is room within the new model to support this, at reasonable cost, whether through being simultaneously signed in to a Google account (hopefully with adequate security filtering for the submitted fragments), or other means.

I can't help thinking there's also an educational aspect to similar scenarios. Always good for a little PR!

P.S. As long as I'm asking, how about HTTPS support?

I am happy to see Google selling clever things like this. I feel that in the long term it's their future, more than advertising.

I have exactly the same sentiment. One of my big fears about Google is that, similar to what happened with Java and Sun, they will ride on their easy money to make cool things and then one day when / if ad revenue dies off (or hell, even just stops growing) it will all go away or get sold to hostile competitors.

In this light, seeing Google actively driving commercial sense into their services and cancelling those that will never fly is, while disappointing, also reassuring.

This is great news for MyGengo (http://mygengo.com/). They have a free machine translation API which can be upgraded to (paid) human translation by toggling a flag.

Granted, they support a lot less languages, but for the supported ones, MyGengo sounds like a much better option, unless Google manages to improve their translation and beat MyGengo pricing.

The quality of the Google Translate is awesome, considering it's automatic, but still falls short of the human doing the job. Also, it'll be interesting to see what Duolingo (http://duolingo.com/) can do, once it launches.

>Granted, they support a lot less languages, but for the supported ones, MyGengo sounds like a much better option, unless Google manages to improve their translation and beat MyGengo pricing.

As far as I know, MyGengo doesn't do in house manchine translations. Am I wrong?

I'm not sure whether they do it in-house or outsource, but they do provide it. From http://mygengo.com/services/api/#benefits page:

You can also get free machine translation through our API - so no need to switch between services. Just pick and choose which of your content you want done by humans, and which by machines. It’s easy and free to start with machine translation, and then upgrade for popular content.

Hey guys

We currently source our API's machine translation from Google, but we're going to be connecting to several other providers in the next few months. Depending on Google's pricing model we may or may not include them in the roster.

We think it's best to connect to our API, because you can 'upgrade' to human translation without switching out code (it's just a single parameter in an API call).

If you want to try us out, we're currently running a $25-credits-free campaign to switch to our API. http://mygengo.com/talk/blog/translation-apis-google-shuttin...

(CEO, myGengo)

I'm glad to see the shit storm did not blow over unnoticed. The translate API is a pretty useful tool and I would happily pay for it. While nothing beats a hand-written translation by a language pro, the automated version is accurate enough for many situations I come across.

I have a feeling their federal sales office out in D.C. would do very well if they could package this up as a network appliance.

This made me consider the problem relying on shitstorms for feedback. The problem might be unrelated (for example, this might have been engineered in advance as a way of minimizing the negative reception of the decision).

That is, in sufficiently large consumer bases, is it safe to rely on shitstorms for negative feedback on business decisions?

Some groups have outspoken representatives that can get the word out there so that it can snowball into something that can be noticed.

However, what if the average emotional response to the decision had been the same, but no shitstorm was raised? Does it happen often? In what conditions?

Hopefully they allow some kind of free tier for minimal use.

I agree. I really like playing with translation software as the web does not speak one language. There are, however, loads of free alternative but they are a bit harder to plug into.

Interesting, would love to be a fly on the wall at internal discussions at MyGengo

I thought MyGengo was in the business of real translations, not machine translations. How is this relevant?

Full Disclosure: I work with myGengo.

We offer machine translations for free, but also offer human translations for (IMO) pretty good rates, and they can all be gotten from the API. I assume the parent comment (of the parent comment of this comment... yeah) was speculating on how we were planning to deal with Google ever deprecating their machine translation. :)

tl;dr I don't believe we were ever too worried, as we're certainly no slouches in the tech department ourselves. Initial shock was there though, definitely a "wtf" moment.

Yeah not only that. Now that it might be a paid product, they might start focusing more on it? and lead to overall improvements, with paying customers keeping them up at night.

Yeah but i would assume that they would still have a fairly large, overlapping customer base?

Also see comment above, about machine translation support at mygengo.

It's absolutely unbelievable that for a solved problem we don't have reliable offerings and need to beg Google/Microsoft for an api access or buy expensive enterprise software. How difficult is to take one of the thousands open source translation libraries that every university's linguistics department offers and turn it to some semi-correct service

If it's that easy, perhaps you should do it in your spare time next weekend.

Well I didn't say it's easy (at least to me, I have zero education in machine translation) but some people spend their entire lives working on it(using our tax money) and just don't feel the need to make something public facing out of their projects except dumping the zipped source code on sourceforge.

P.S. And by the way I tried not to use the google api (well before the announcement) and wrote a basic word by word translation which obviously gives dismal results (https://github.com/aparij/bl)

Great news for YouLing, my mouse-over translator for Mac. I just made it a couple of weeks ago, and was very disappointed when I heard about the API closing! http://goo.gl/MsV4D

a new revenue stream, sweet. To bad Facebook and Bing are winning the web away.

Bing has made huge strides in their search and mapping system. I love it. However, for what ever reason, I still do my searches via Google.

Microsoft Ad center was also a walk in the park compared to Google Adsense. I am a novice though.

As long as Bing continues to have to advertise, I doubt they're winning.

Its true, facebook is making a new type of internet. Its hardly about search anymore, now people search facebook for content. Social networks are a fad like anything else but they are winning now.

Wait what? People search on Facebook for content? Maybe for content of what their friends are up to. I've never considered a social networking site to be a search engine, and considering Facebook's abysmal attempts at targeted marketing (their ads SUCK at relevance) no matter how many time I tell it that a certain ad is of no relevance. I just don't see how Facebook is "winning the web war". Maybe I have different friends than most, but pretty much everyone I know uses a search engine to find things. Even my mom, who isn't what I would call computer savy.

That is certainly a valid point. Mostly what happens with people on facebook is that they stay on facebook. They play games, they chat, they look at pictures they stay on facebook. When you goto a movie the ads don't have websites, they have facebook pages. Facebook has created a new type of internet, a more closed siloed internet (Dave Winer reference). People used to use the internet, now they use sites.

The new layer is data locked into private site that farm content.

people (non-hackers) are searching for content (and people) on facebook. pages alone host a considerable amount of fresh content from businesses, institutions and organizations.

you do have to consider many of those searches as looking for stuff already in their social graph (friends, pages, etc). i know i use their search to navigate all the time.

They are winning the web war since their messaging has basically completely displaced AIM, MSN and GChat in the US in the 16-26 year olds. Its also put a decent dent into email for asynch communication, and I only expect that to continue.

I don't see facebook overtaking email in professional situations for a long time.

I Agree. If a business contact wants to get to me through facebook I cringe. Although I can see the current generation using it in a more professional sense.

Honestly, I think it's pretty likely that Facebook will tend towards with people having "Professional" profiles offsetting LinkedIn and they will offset email.

Obviously at this point it is largely just offsetting personal email, but I don't think it's reasonable to dismiss that as insignificant.

Oh no! What will Google do without all the money they make from GChat?

I don't really understand your point; any amount of money that they make from search or gmail or docs is just because someone is on that page and sees ads. The exact same situation applies with GChat assuming you use it through the gmail interface rather than through an alternate client (which most people do).

Registration is open for Startup School 2019. Classes start July 22nd.

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