We will increase the disposable income of households and encourage consumption, by introducing policies such as child allowance, free high school education, abolishment of highway tolls, and abolishment of provisional tax rates. These measures will change the Japanese economy to one centered on domestic demand, and will make stable economic growth possible.

The development and proliferation of the latest technologies such as IT, biotechnology, and nanotechnology will be supported. Strong government support will be given particularly to measures against global warming in order to further enhance Japan’s advanced technical skills in this area and to nurture the environment-related sector as a future growth industry.

The agricultural, forestry, fishery, medical and nursing care industries are new growth industries. Providing an individual household income support allowance for farmers, improving working conditions for people in the medical and nursing care sectors and other measures will be taken to enhance attractiveness and growth potential of these industries, and nurture them so that they become industries that create large-scale employment.

I personally think this is a very very poor economic growth strategy, and I will soon elaborate thoroughly on my blog

Right now we support many feeds and methods to get social and learning data, but also to write some content. Soon, we will also release: 1. The ability to create items in any language. 2. The ability for developers to save study data


h2. Some facts about the API

• built on RubyOnRails, powered by Mashery.com
• much more REST than many Open API’s out there
• a lot of hidden controllers that do cool stuff like for example converting Freebase MQL to XMLVocabulary
• I love JSON, so almost every call has a JSON with callback response :]
• making the first steps towards Semantic Web compatibility, by using RDF, well formatted XML and Microformats

Code quote

This is a little piece of code I think kicks ass. I love the nested routes!





h2. More background information

Presentation by me at Tokyo 2.0, explaining about why/how we’ll open up the iKnow! Platform

Old blog entry on Dominiek.com with some thoughts on new ways of API development

At a first glance the concept seems quite easy, we just find all the words on the page you want to read and search for them in our item bank, but in reality it’s a little more complex.

First of all, because English is an inflected language we can not simply search for the words as they appear on the page, since the corresponding items in our database might be in a different inflection or might not be inflected at all.

To make sure we can find all inflections of a word we use a process called lemmatization to get the base forms of each word on the page. This base form is called a lemma and is the form you use to look up the word in a dictionary. In English the lemma of a noun is its singular form, for example friend and book, and for verbs it is the infinitive form, for example talk and eat. So if the word on the page is “ate” we will do a search for both “ate” and “eat”.

To put this in context I’ll show the lemma forms for Japanese as well. Nouns in Japanese do not inflect, so they are always in the base form. For Japanese verbs the lemma is the non-past tense (基本形), for example 話す and 食べる, and the same goes for adjectives, for example やさしい and 高い.

The second trick we use to make the bookmarklet smart is when you click on a word to search for items. We do not just do a plain search for that word, but we also pass along the sentence that contains the word. We use this sentence to figure out what part of speech the word you clicked on has. Then we use this information to only show you items that match this part of speech.

So for example, if you clicked on the word “records” in the sentence “I have many records by Kraftwerk.” you will only get items that contain “record” or “records” that are nouns. But if you clicked on “records” in the sentence “A microphone records sound.” you would only get items that are verbs. This help narrow down the item search results by showing you more relevant hits.

However, languages are very hard to parse for computers, so if we mis-identify a part of speech or lemma we will fall back on searches that give more, but less accurate, results so you will be able to identify the correct item yourself.

Of course you won’t notice any of this directly when you use the bookmarklet, as it all happens in the background, but hopefully it should make your English studying even more effective.

Oh yeah!

http://flickr.com/photos/dominiekterheide/2402176059

And access your FlickR pictures:

You will hear more about the new features soon ;-]