In an effort to meet the rapidly evolving needs of webmasters all around the world, Google recently announced on its official Webmaster Central Blog the release of a new widgets to allow for a better language translation of the website contents, as well as new HTML tags that will be parsed by Google translator to disable automatic translating on specific sections of the site.

Automatic translators have been around the Web for a while, the most notable being Yahoo's Babelfish, and provide a useful although not always accurate tool for both international visitors and webmasters who want to allow their visitors to understand the content featured on their sites. Despite long-time efforts to perfect the art of automatic translation, this field, along with the rest of artificial intelligence, hasn't quite reached the stage where it can understand (or even represent properly) the contexts and semantics that are typical of the human mind.

The problem with natural language processing in general is also one in which the greatest steps are still to be taken. It is not by chance that the so-called "Turing test", designed to test the development level of an artificial intelligence, relies completely on natural language processing, given the difficulties that are inherent to the process of understanding (and later translating, in the case of Google's tool) the human language.

One of the most publicized competitions that involve a Turing test is the Loebner Prize for Artificial Intelligence, in which the best "chat bots", or automated chatting robots, compete against each other attempting to simulate a human chatter to pass the Turing test. The competition has been going on for over 15 years now, with scarce results: one of the classic examples of such poor results is that the bots aren't even able to answer apparently trivial questions such as "What color is a blue

car?" unless they have explicitely programmed to do so.

According to many experts, even with all the investments Google has been making in the field of making their translation tool smarter than ever, their results won't be satisfactory until they will be able to successfully "parse the meaning" of sentences, something we are currently still very far from achieving from a technological point of view. Today, Google's translator relies on what many consider "the next best thing", extensive language-to-language words and expressions dictionary-like databases and occasional Web searches to provide some basic understanding of the context where needed.

While the results are far from being grammatically — and, sometimes, even semantically — correct, they often enable international users to get the gist of the content being translated and still represent an essential tool for many webmasters. Google's latest update of its translation tool adds a useful widget, now available from this address, that makes translating content on the fly even easier for visitors.

Knowing the limits of its translation tool, or maybe simply in an effort to adapt to the various needs of the webmasters relying on Google's tools to personalize their sites, the company also introduced two HTML tags that, if embedded into the HTML code of a website, can prevent the selected area from being translated automatically by Google's tool.

The first one is a META tag to be embedded in the HEAD section of the page which prevents entire pages from being translated:

<meta name="google" value="notranslate">

The second, more flexible, can select a specific region of the page:

<span class="notranslate">This text can't be translated</span>

The tags are quite straightforward to use and don't require further explanation. Lastly, Google engineer Josh Estelle suggests making use of the AJAX Language API introduced by the company back in March to personalize the translation option even more.