Showing posts with label machine translation. Show all posts
Showing posts with label machine translation. Show all posts

Sunday, April 04, 2010

Being Fast

Ken of the 1-800 Translate company is blogging about Speed.. not the kind that makes you unwell, but the kind that makes you productive.

"You’ve got to do it right, but you’ve got to do it fast too. How fast? It all depends. Most translators in most languages can do somewhere between 2500 and 3000 words per day. That works out to about 300 words per hour. That’s unique words, now, so if the materials to be translated have a lot of repetition (say 50%), then the translator is effectively translating 5000 words instead of 2500 (...) Terminology management, machine pre-translation, and lots of other stuff can push those numbers much higher (...) Edit and proof take more time, say around 1000 words per hour, each, depending on the quality of the original and the talents and diligence of the editors. Traffic becomes a concern, too, in that the document has to be scheduled and assigned to the appropriate linguist by the appropriate project manager or language manager. So all the expectations we have for words per hour get tossed out the window if the one-and-only language manager is backed up or simply having a bad day. "

When all else fails, Ken throws in more translators, but complains that we are like cats, that is, there is more than one way to skin us. Thanks, Ken. I agree that as a bunch of professionals, we find it difficult to agree on anything, starting from translation theory and ending with which style manual is best to use. Don't worry, though - if the automation market keeps growing at the speed it does currently, our own productivity will kill all cats in the linguistic neighbourhood. That achieved, your only worry will be software crashes.


********
For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/

Weaving Trust

Here is an excerpt from a recent interview with Mark Tappling, the CEO of Language Weaver, a venture-backed firm which develops language translation software:

"What we've discovered on the commercial business side, is that there is a massive need for translation of digital content on the Internet. But, because of the large volume of digital content out there, most of it simply is not being considered for translation, because there's just not enough money, people, or time. That said, we made an entry into the commercial market in the digital content space. In order for that to be credible, we found that there was a lack of trust--most people do not think translation software works well enough to be commercially viable. Our strategy was twofold--one, was to go out and get brand name customers, and to translate their content to such a high quality, with no human oversight, that they'd be willing to post that on the internet. We've been able to achieve that (...) The second thing we needed to do, was to engage our users trust. In order to do that, they needed to know when a translation was good or bad. So, we established a patented algorithm--a trust score--which automatically ranks a translation from 1 to 5, five being great, one being poor, so that a customer can determine the trust of that. We calibrated that to our customers own, native speakers, who scored them, so that everyone has trust in the score. That has taken the risk out of the conversation, because we only ask customers to pay when the translations are good."

Is that bit about "no human oversight" giving anyone a chill? If their MT is such high quality, how come Google is struggling? Something is telling me that their average score is around 3/5 and that they sort of provide a tiered system: if it is mediocre, we will get it reviewed by humans, unless of course you are happy with being mediocre in language X (which most of their clients would be, since any language other than English is an enigma and for a pragmatic English-speakers, enigmas are best left at mediocre levels). So close enough is good enough, and if their algorithm produced "3" is 30% better than the "5"s of other MTs, they aren't doing bad.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/

Tuesday, March 16, 2010

Get a Piece of That Cake, Quickly

Is your company doing business with the Arab world?

The Google CEO, on a visit to Riyadh, predicted that the Kingdom would have a bright future because of its large young population. He said the Arabic content on the Worldwide Web was expanding by the day. "The Internet Arabic content will increase further due to the presence of a large Arab population estimated at 300 million," he said, adding that Saudi Arabia would have a big role to play in expanding Arabic content on the Internet. According to statistics published in 2009, 48 million Arabs use the Internet.

Google International has two offices in Egypt and Dubai but no presence in any other Arab country. It has introduced Arabic language in its search engine and provides translation services from different languages to Arabic. But be careful - Arabs are very touchy about their language. After all, the language is sacred because the Koran came down in it. So Google-type mess of machine translation is a very sure way of making unhappy and offended clients.

If you are doing business with the Arabic-speaking world, it is time to localise at least the most important parts of your website NOW. For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at www.arabic.com.au .

Thursday, March 04, 2010

Google Chrome Translation Software




For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Wednesday, March 03, 2010

KIT, lip-reading, and a multilingual mobile phone

From AFP 02/03/10

It has happened to almost everyone. You are sitting on a train or a bus and someone right next to you is annoyingly shouting into his or her mobile phone.

­Researchers at Germany's Karlsruhe Institute of Technology (KIT) have developed a method for mobile phones to convert silent mouth movements into speech. The technology is based on the principle of electromyography, that is the acquisition and recording of electrical potentials generated by muscle activity. This muscle activity is measured in the face and converted into speech.

The user can speak into the phone soundlessly, but is still understood by the conversation partner on the other end of the line. As a result, it is possible to communicate in silent environments, at the cinema or theater, without disturbing others. Another field of use is the transmission of confidential information.

"We currently use electrodes which are glued to the skin. In the future, such electrodes might for example by incorporated into cellphones," said Michael Wand, from the KIT.

The technology opens up a host of applications, from helping people who have lost their voice due to illness or accident to telling a trusted friend your PIN number over the phone without anyone eavesdropping -- assuming no lip-readers are around.

The technology can also turn you into an instant polyglot. Because the electrical pulses are universal, they can be immediately transformed into the language of the user's choice.

"Native speakers can silently utter a sentence in their language, and the receivers hear the translated sentence in their language. It appears as if the native speaker produced speech in a foreign language," said Wand.

The translation technology works for languages like English, French and Gernan, but for languages like Chinese, where different tones can hold many different meanings, poses a problem, he added.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Japanese Human Translators Here to Stay

..says Ry Beville, himself a translator from Japanese to English..

The reason, says Beville, is that translation programs like those provided by Google and Babelfish can handle business letters and websites to a reasonable degree; they at least convey the gist of a text, but they still have a long way to go before they become truly viable. The fact that their translations need to be heavily edited isn’t the problem. It’s all the meaning that is lost, especially between Japanese and English. Linguistic science tells us these languages are worlds apart. For a native-English speaker, Japanese is a level-five language, meaning the hardest to master (a distinction it shares with Arabic). Creating a program that can negotiate such vast differences seems like a quixotic dream perhaps exceeded in difficulty only by the pursuit of robust artificial intelligence.

Beville is no Luddite himself, and freely acknowledges the debt of online dictionaries in his career. His proposed solutions are quite interesting in themselves: an open-source translation program that works by concensus, its entries by no means absolute, but rather refined through increasing participation in the project. A program could always map out the structure of a given sentence (remember diagramming sentences?) and use those discrete values to create a correspondence in the target language. The second option would be to begin compiling publicly available translations of identical or similar structures and finding a kind of average or typical rendering of those structures.

But that won't always work, because even though generally accepted grammar structures are finite, possible combinations of words and their meanings within those structures approach infinity. Mapping structures, compiling known translations, refining the system — it all seems so Sisyphean, even for a computer or enormous open-source project.

The full article, worth reading, can be found here.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

ITI Warning Against Machine Translation

From International Trade

The Institute of Translation & Interpreting (ITI), with 3,000 members worldwide, is one of the UK’s primary sources of information on language services for government, industry, the media and the general public. ITI promotes the highest standards in the profession and its members play a key role in promoting cross border trade in products and services. In this article ITI general secretary Alan Wheatley highlights the growth of machine translation services offered by global phenomena such as Google and Facebook and warns against relying on machine translations for business.

If you’re reading this you probably rely heavily on the Internet for your business. Most do, and not just for emails and finding information, but for strengthening brand awareness, and engaging with stakeholders. The interactive nature of Web 2.0 platforms, social media sites such as Facebook and blogs such as micro blogging site Twitter, mean that it’s easier and faster than ever for businesses to communicate on a global scale.

As an international trader, you’ll appreciate the necessity and benefit of having your company literature – and hopefully by this stage your website too – translated into the languages of your target markets, but what do you do when you begin to embrace social media on a wider scale and set up a company Facebook page, a blog, or a Twitter account?

Google Inc is one corporation who believes that having all your online communications translated into the relevant language/s is important and claims to be the largest free language translation service online. Its Google Translation website lists a total of 52 languages available for translating websites, documents, searches and individual words – and all by computers using mathematical equations, at a click of a mouse, without any human translation. In theory, this means that if you have a corporate blog you can simply paste the URL into the Google Translate facility and your blog will appear instantly in a language of your choice.

Others think it’s a good idea too and have their own translation tools. Facebook Inc – with more than 300 million users most of whom are outside the US – for example, has approached it slightly differently and has crowdsourced its users – again not necessarily professional translators, although some reportedly are – to translate the website into 65 different languages. It goes further. Facebook thinks its pro bono crowdsourced translation approach is so good it wants to patent it. It’s also offering users of Facebook Connect the opportunity to tap into the Facebook community to enlist help translating their sites into any language.

Even Twitter has jumped on the crowdsourced translation bandwagon to have its service available in French, German, Italian and Spanish, in addition to English and Japanese. Twitter plans to enlist its users to offer translations, reportedly alongside some in-house translators.

Where’s the harm in any of this, you might ask? Perhaps none, if all you want to do is exchange a few personal emails with a colleague half way across the world whose language you don’t speak. But if you think you can use Google Translate, for example, to publish your website or blog in other languages, replacing human translation with machine translation, you’re putting your brand and reputation at serious risk.

How can a machine translation ever guarantee accuracy and quality? It’s quite simple – it can’t. Although Facebook and Twitter offerings do have some human intervention in that humans supply possible translations and others vote and the winning vote is applied, there’s still no guarantee that it’s the best translation in that particular context. Translation is a complex task requiring a lot of skill and knowledge. Would you trust an unqualified teacher to teach your children, or an unqualified lawyer to offer you legal advice? If it were easy to speak each other’s language we would all be able to do it by now and the profession of translation, which is almost as old as language itself, wouldn’t exist.

There is much more to being a translator than meets the eye. A good translator has an exceptional command of the native or ‘source’ language and can provide outstanding copywriting, proofreading and editorial services. They usually operate in highly specialist fields and possess the relevant technical vocabulary. They guarantee your communications are fit for purpose. Professional translators also identify unarticulated needs, thanks to their acute appreciation of cultural differences and approaches to business and industry practices.

It’s true that professional translators themselves use technology to speed up their work. However, specialist computational linguists have been working for over 50 years to achieve a high degree of quality and the translators themselves programme their software to suit their particular area of focus. A professional translator would never rely purely on a machine translation because computers will never understand how a language works and therefore neglect the essential aspects of good communication – accuracy, clarity, style, nuance and cultural sensitivity.

Errors in translation and interpreting can waste enormous amounts of time and money, resulting in incalculable costs in terms of misunderstanding and loss of prestige. Professional translators understand this and know business leaders care about the effect their online – and off line – communications will have on the business and will ensure the company communicates effectively to enhance the perception of your brand and reputation.

Sadly, it can be very easy to get it wrong by cutting corners and ITI members are often called in to rectify strategic translation errors. Rather than hiring a professional from the outset, some companies believe they can save money and achieve the desired result by using a machine translation or enlisting someone in-house who just happens to speak the language/s in question.

Poorly translated communications can mean your brand will be poorly perceived resulting in limited or no success, and by not using a professional translator you will not only waste money, but you will also miss opportunities, fail to attract appropriate media attention and damage your reputation. It makes sense to get it right from the outset and call in a professional translator.


To identify professional translators – and interpreters – there are some obvious checkpoints. You can start by confirming qualifications, references and memberships of professional bodies. A professional translator or interpreter will always be happy to provide these. ITI members, for example, demonstrate their commitment to the profession by joining the organisation and adhering to a strict code of conduct. This is essential in an unregulated profession. You can find the award winning ‘Translation Getting it Right’ booklet available in the Advice to Business section of www.iti.org.uk plus a Directory of Members.

You’ll know when you’re working with a professional because he or she will offer consultancy, add value, translate into their native language unless they are multilingual, ensure the end translation is fit for purpose, save you money, help safeguard brand reputation, offer specialist knowledge, request feedback, solve problems, think ahead and advise. No computer will ever be able to do all that!

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Wednesday, January 20, 2010

Engrish Resources for Blooper Collectors

The term “Engrish” widely refers to the incorrect usage of the English language – usually in written form – by people in East Asian countries, including Japan, Korea and China. “Chinglish” is the term used to describe incorrect usage of the English language in China or by Chinese people.

Things associated with the west, including English, are wildly popular in East Asian countries because they are seen as exotic. So the Chinese, for example, slap English sayings on anything from t-shirts to pencil cases to bubble gum wrappers. Sometimes the translations are done by professionals and checked for quality- but often they are done using raw machine translation, a dictionary or a person who knows very little English, producing the “Chinglish” phrases that we all love so much.

Now you can watch a slide show on Flickr over 1,300 photos of Chinglish, or visit Engrish.com to read about Engrish and see hundreds of hilarious photos of objects with Engrish texts written on them.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Free TM software reviews

OmegaT:

A CAT tool which some of its features include user-customisable segmentation using regular expressions, translation memory, fuzzy matching, match propagation, glossary matching, context searching in translation memories and keyword search in reference materials.

OmegaT uses international standard TMX (Translation Memory eXchange) as its native translation memory format which most CAT tools are able to import and export. TMX is an open XML standard which allows easier exchange of translation memory among translators with different application tools.

Although it lacks of features in comparison with other CAT tools, OmegaT is fast in segmenting a source file, compiling target files (known as ‘clean up’), and its fuzzy matching. In its earlier version, OmegaT segments the source file by paragraph, rather than sentences. Now, you can select paragraph or sentence-level segmentation. By using the function called Editing Behavior under Options, you can have the selection of displaying in source text or best fuzzy match. This is a CAT tool that is worth trying out.

OmegaT includes the following features: – Fuzzy matching – Match propagation – Simultaneous processing of multiple-file projects – Simultaneous use of multiple translation memories – External glossaries – Document file formats include:

- XHTML and HTML Microsoft Office 2007 XML OpenOffice.org/StarOffice XLIFF (Okapi) MediaWiki (Wikipedia) Plain text

- Unicode (UTF-8) support: can be used with non-Latin alphabets – Support for right-to-left languages – Compatible with other translation memory applications (TMX)

Some third-party software that can be used with OmegaT are OpenOffice.org, Rainbow, bitext2tmx, and Samuel Murray’s scripts and procedures.

Download

Anaphraseus :

ACAT tool for creating, managing and using bilingual Translation Memories. Some of its main features Include text segmentation, terminology Recognition, plain-text TM (Unicode UTF-16), fuzzy search in TM, UTF-16 TMX export/import, user glossary and OpenOffice. org extension.

Anaphraseus works in OpenOffice.org as an extension and work seamlessly with TM created in Wordfast. You can install it using Extension Manager under Tools in OpenOffice.org. During the translation process, Anaphraseus divides the translated text into segments or sentences. When a segment is selected, Anaphraseus displays the closest match found in the TM.

As an OpenOffice.org extension, it allows you to use all the features of the word processor “on the fly”. When it comes to performance, it is slower than Wordfast, especially when handling large files. Anaphraseus works with translation memory in TMX format, which allows you to work with most CAT tools in the market.

Anaphraseus includes the following features:

- Text segmentation – Terminology recognition – Plain-text TM (Unicode UTF-16) – Fuzzy search in Translation Memory – Unicode UTF-16 TMX export/import – User glossary

Since Anaphraseus is still in beta test, it currently works only with unformatted text, skipping all formulas, pictures and similar objects. If you need a CAT tool that won’t cost you anything and works with OpenOffice.org, Anaphraseus is worth a try.

Download

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

It is still Greek to me..

..or rather, it is still Gibberish. Greek, provided it is correct, is fine.

The Oregon Department of Administrative Services announced Friday that the state’s official Web sites have added Google Translate, giving users the ability to view those sites in 25 languages in addition to English.

A product of Google, the feature is available on Oregon.gov.

“By adding this feature to our Web sites, we can make important government information available to more people who need it,” said Wally Rogers, Oregon’s e-government manager. “We are able to offer Google’s page translation service for over 40,000 pages without spending a dollar. Here in Oregon we have many residents who speak languages other than English, and we have increased our ability to reach out to them through Oregon.gov.”

Visitors to Oregon.gov will now find a “drop-down” box in the upper right corner of every state Web site for which the address ends in Oregon.gov. The label on the box is “Select Language.” The user simply activates the box by clicking on it and selecting any one of the 25 available languages. The page then quickly appears in the selected language.

The following languages are now available on the state Web sites, in addition to English: Arabic, Chinese (simplified and traditional), Danish, Dutch, Filipino, Finnish, French, German, Greek, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Persian, Portuguese, Russian, Spanish, Swahili, Swedish, Thai, Ukrainian and Vietnamese.

Yikes! I would hate to live in Oregon and have to deal in legal issues through a Google translation..

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Too Many Words, Too Short a Deadline - a solution?

This was proposed by Oxford Translation Ltd in "Content for Reprint" (we are talking about 600 pages of a technical report needed in 48 hrs):

"If the translator cannot produce the best possible full translation within the time available, then an "optimal" solution should be agreed upon. Such translations could fall into one of the following categories: - "a translation that provides what is needed for the stated purpose" - "as good as possible in view of the customer's requirements" (...) Due to the obvious time limitations, a summary could be proposed as an alternative. This could have the additional advantage in that key bullet points raised for discussion in the international meeting would be more easily accessible in a summary. Of course, production of the summary would itself involve the translator reading all 600 pages of the report and that alone could probably take more than the two days available. So from a practical point of view, it might be that the person who is asking for the translation could provide a summary in his or her own language - after all that one person is most likely the one person who is fully aware of the salient points in the report (...) One of the ways in which an experienced translator can produce a long translation in a short space of time is by dictating the translation into a digital recorder and then emailing the translation to one or more typists who in turn type out the translation and send it back to the originating translator for correction before delivery to the customer. Using this technique, a translator can produce upwards of 10,000 words per day. Almost everyone can speak faster than they can type. [Sure, but can the typist type as fast as you talk? And do you "sight translate" at the same speed as you talk to your friends about the latest fishing trip??]...A similar technique uses one of the computer speech recognition programs, which have seen a great deal of improvement over the last few years. Using one of these systems, the translator can dictate the translation into a small microphone while at the same time his or her computer types the translation directly onto the computer screen. When completed, it can then be corrected, proof-read and sent to the customer. Of course a translation agency is always pleased to accept a long translation (as long as the deadline for completing the translation leaves enough time for the translation itself and proof-reading, etc.). [Unreal. My Dragon Speaks Badly was ditched after producing French text when I was talking in English to it. I don't speak French, and I don't have a French accent! As for proof-reading 600 pages, that needs a bit more than just 48 hrs, right?]If the original text is a technical document, translation memory systems can be used and this ensures that repeated expressions are translated automatically by the memory system. The reader is then reassured by repetition of the same phrase for the same action and vocabulary is standardized by the system, avoiding any ambiguity to the reader. The same word is always used for the same object. Lower costs can result from this, which can be passed on to the translation agency's customer. [Provided, of course, you are already in possession of such a TM and glossaries].

No further comments. I am a craftsman. I do not encourage clients to believe in myths. It would make much more sense to actually educate the client as to the need to include possible language needs into the business process from the very start.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Wednesday, January 13, 2010

Got website? Globalising?

According to Common Sense Advisory, more than 50 percent of Web users who purchased online buy only at Web sites where the information is presented in their language. (Going from Simple Translation to Successful Transactions on Global Websites, Common Sense Advisory, March 2007).

Here are the top tips to consider for translating Web content from Mark Tapling of the Language Weaver:

1. Pay attention to the content type

Every content type has an audience, a value that it serves for the business, and a value that it serves for the audience. Because of this, different content has different requirements for translation.

  • Some types of content – like advertising material – are highly influential and need to be perfect, thereby requiring human translation.
  • Other types of content - like documentation - need to be near perfect but don’t require the same nuance as other types of content; this type of content can be served up with translation software and post-edited by a human reviewer. (Maybe!)
  • Some content - like knowledge bases and FAQs - simply needs to convey facts (but needs to do so quickly); automated translation software from Language Weaver can be effectively leveraged for this type of content. (Oooops! Can't judge the leverage.)


When looking to start or add additional translation, take a look at the content in your organization and consider other translation options to cost-effectively translate the information while still meeting the needs of the business and the audience.



2. Match the translation cycle to the business need

Some types of content only need to be updated and translated every time a new product or service is introduced, such as product documentation, marketing material and product oriented Web content. Other types of content should be translated as soon as they are written, such as bug fixes, FAQs and knowledge base articles, in order to save on ongoing support costs for known issues.

3. Don’t forget about search

When updating Web content, think about how users will find that information if they speak another language. Search is how most people will find information, so keep in mind that content needs to be translated to make it searchable in a visitor’s native language. Some sites pre-publish all translated content to make it visible to search engines, others don’t. Companies that don’t translate content up front are subjecting their site to an unreliable user experience delivered by a free translation plug-in. In theory, this can work, but remember that free translation sites aren’t trained to understand your brand voice and terminology so there is risk built in when putting translation in the hands of free translation.

4. Maintain control of your brand

Whether you use human translation, post-edit output from translation software or use automated translation, brand and term management is something that is crucial for a growing business. Your company has likely spent a lot of money on its brand, so you don’t want it to be at risk when expanding into new regions. For example, what should your product name be in Chinese or Japanese? How should you describe a particular experience with your product or service in another language?

Identify the brand voice you want to maintain, key product names and references, and company terminology up front so that you know how they should be represented in other languages. Working this out ahead of time also helps any translation project go smoothly and lets you maintain more control of your brand across languages.

5. Use the "NOTRANSLATE" tag

A site owner can control content by ensuring proper tagging of the information. While there are several page-level tags available to exercise this control, the most exciting one was recently added by Google. At the page level, site owners can now add a tag called “NOTRANSLATE.” This tag is interpreted by the crawler as a “do not translate” instruction. By setting the “NOTRANSLATE” tag, you ensure that Google’s automatic translation is not offered as an automatic option to anyone performing a search – especially for those pages or parts of your site where you want to make sure YOU, the site and brand owner, decide what the translation should look like without handing over control of the translation to Google’s algorithms. (My favourite tag of the year, that one!)

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

10 Tips On How to Come Across Polite in a Business Email

Here we go, thanks to One Hour Translations

1.Use the recipient’s correct title.

2. Use the correct greeting.

3. Introduce your company and yourself.

4. Write short, simple sentences. (This could backfire, though. Not all cultures respect simplicity as clarity)

5. Avoid idioms and slang language.

6. Do not refuse requests directly, and avoid criticism in emails or letters.

7. End your letters and emails properly.

8. Where possible, send the letter or email in the recipient’s native language. This shows that you respect and care for their language and culture. Using the recipient’s native language will usually be accepted very positively and will usually reduce cultural gaps or potential misunderstandings.

9. Use only native-speaking translators. The recipient will know immediately if the translator is not a native speaker of his or her target language. Avoid using machine translation (like Google Translate or Yahoo’s Babel Fish): automatic translation is still very far from producing an acceptable result. In many cases, it can completely distort the original meaning. (Sorry, had to copy this in its entirety)

10. If possible, use an independent proofreader who is also a native speaker of the target language: two pairs of eyes are always better than one. The proofreader can review the style, fix any typos and ensure the translation is perfect.

8-10 will cost you money. But it will cost you more if you offend your client.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Monday, January 11, 2010

Dow Jones has a new translation gadget

From VIP Live Wire

Dow Jones Factiva has launched the beta version of a new translation tool. It uses technology from Language Weaver, and as with other automated translation systems, the results are mixed. However, this free service gives a general indication of meaning and will thus open a significant volume of international content to business researchers.

VIP readers will already be aware of the various free online tools for translating text and web pages, including Yahoo's Babelfish and Google Translate. They are useful for getting the sense of source material but still no match for the human translator.

Factiva's tool is a welcome addition, being an integral part of the system, thus saving the researcher time. It allows articles to be machine-translated from seven languages (Spanish, French, Italian, German, Russian, Chinese Traditional, and Chinese Simple) into English as well as English into the same languages. Translation between these languages (i.e., French to Spanish) is not available as yet.

Not all content is available for translation. The Factiva announcement states that 5,800 sources from approximately 360 information providers are currently available including Dow Jones, Interfax, N.C.N. Limited of Xinhua News Agency, Russica-Izvestia Information, South American Business Information and Newsquest Digital Media.

Clicking the Translate Article icon above available articles shows the languages options in a pop-up box. A new window opens to display the translation. At the moment there is no indication from the headings list whether an article can be translated, so users need to open the full article to find out. Changing this would be a useful modification.



Translation programmes are based on different approaches and systems – Babelfish uses SYSTRAN, and Google and Language Weaver have both developed statistically based technology. Looking at some German language examples (into and from English), Factiva's Language Weaver translation gave the gist, but had some difficulties. Users with a high need for such tools may want to compare Babelfish, Google Translate and Language Weaver to see which best fits their needs for international business news.

One should always turn to a professional human translator for mission critical information in a foreign language, however, for those of us who regularly research international markets, free online translation has been an important tool for quickly accessing content. Despite the mixed results, the results can provide key clues for further research. Even Factiva has this warning in place:

"The content of this translation is generated by way of automated translation software and is no substitute for a professional human translator. Only the text from the Dow Jones Factiva service is translated; additional web sources and links to the Internet are not translated as a part of this service. Dow Jones does not take responsibility for the accuracy, reliability or completeness of this translation and the results of this translation are “as is.” Dow Jones makes no representations or warranties of any kind with respect to the translation on this site or its contents and disclaims any liability, loss or damages suffered based on the user’s reliance of any kind or for any purpose with respect to the content of this translation."

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Thursday, January 07, 2010

nFluent in gibberish?

These are my comments on the CNN feature about IBM's attempt to create an "instantaneous translation" gadget.

(a) I hate scientific reductionism and the bruhaha of "technology can do all". It can't. Technology cannot bring you back from the dead, cure you of cancer, make someone fall in love with you, make your kid brighter - or suddenly translate fluently into a score of languages.

(b) "We have a web page interface, where you type in a URL and it automatically translates the web page for you (...) "We also have an app that you can put on a web page and when users arrive... they can pull down a menu and change the language. "The ability to translate URLs is something that our customers love a lot, because once you translate the page, you can click on all the links and suddenly you are exploring the foreign language web as an English speaker." As a native English-language speaker, I find web pages written/translated into poor English annoying, disrespectful and not generating trust in the business/person behind it. Same, I assume, applies to any native speaker of languages X, Y or Z. I don't just want to get an idea about what's there. I want to enjoy the cleverness of the writer. I want to be sure I understand what the writer means - the idioms, the metaphor, the innuendo, the sarcasm and the stuff between the lines.

(c)"n.Fluent" began in 2006 as one of 10 innovations sponsored by IBM's chairman Samuel J. Palmisano. The company decided that the language barrier was a key issue, both for global businesses and companies with clients worldwide and so resolved to find ways of addressing the problem." More BS. Language is only one aspect of the many problems. I can learn to speak fluent Japanese in a vacuum, and it will not make me a good communicator with the Japanese. There are things such as cultural issues, perceptions, beliefs, unforgotten histories, economic inequalities, etc. that affect your way of penetrating a foreign market. Language is not actually the biggest of these issues.

(d)"Google are currently working on a tool that will translate not only web pages -- but web searches as well (...) Imagine what it would be like if there was a tool built into the search engine which translated my search query into every language and then searched the entire world's web sites." Awful, that what! I get enough gibberish when I do searches in English, and I have two other languages I am fluent in that I can do the same search in. Using the "Advanced Search" on Google already allows me to find websites with my favourite words in specific languages. At least here I have the choice to filter - badly, since my Arabic pulls down more Farsi websites than Arabic anyway. Imagine what it would be like if you were now getting million more hits because your search phrase is mistranslated in another language: "spirit" becomes "soul" instead of "alcohol". Or "balls" become.. you get the gist, right?

(e)"..the Tele Scouter will mean conversations won't get lost in translation. Unveiled last November the device is a set of headsets and glasses that can automatically translate spoken words and display them on a tiny retinal display." Oh, imagine a specialist telling his dying patient who does not speak English that there is nothing more they can do.. through this Tele Scouter, hanging over one eye, obliterating the patient and relegating him to some exotic item on the hospital bed. Or imagine a lover declaiming his poem in in Arabic to his new British girlfriend.. "Oh my gazelle, your oranges are like a full moon .. oh, my eye, oh, my night.. and sorry I can't look into your beautiful blue eyes because I need to squint into my Tele Scouter." Give me a break, mate!

(f) "Vernacular and jargon can be particularly problematic for translation software, so "n.Fluent" has been designed to learn from its mistakes and pick up specific terms used within IBM. To do this the project has been opened up to all 400,000 staff working for IBM around the world, and uses this "crowd sourcing" to access their expertise to feedback on the project. Over a two-week period in October last year IBM launched a "worldwide translation challenge" to its workforce, which resulted in two million words of text being translated. Incentives in the form of charitable donations and other prizes were offered to staff who took part." And how many of these 400,000 staff - mostly new grads - are linguistically aware of the subtleties of vernacular and jargon in two languages to be able to correct the language? Or is it basically "open to interpretation", where "God, she hit me hard" can be both "she assaulted me" and "I am madly in love with her"? And what happens when this vernacular, a living, breathing, pulsating creature, changes?

Now, the crux of the matter:

(g)"Rapid, accurate translation of such literature published online can deflect calls from call centers, and bring significant savings." When our Dell (or rather Dill) laptop started behaving badly (and it does it religiously), we phoned the Australian Tech Support Centre. The call was deflected to India (nothing unusual). India has a rather poor telecommunication infrastructure, and it was the monsoon season, so we heard crackling and distant moaning (must have been the operator drowning), but at least there was a human being you could vent your spleen on at the other end. Now we will be getting little pop-ups telling you in gibberish English (or another language) that your problem cannot be solved due to a SQL database problem on the website. Contact us for a solution. Oooops, another gibberish pop-up, still a SQL database problem. Enjoy! Don't pull too hard on your hair, it won't regrow.

"Whoever wins the fight for market share, be it IBM, Google or others hoping to close the language divide, advanced translation software looks set to make a huge splash and businesses should get ready: it looks like the world may be about to shrink yet again." The world won't shrink. That's against natural laws. But our brains surely will!


For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Tuesday, January 05, 2010

Plurk Twitters in Tongues - for free?


Plurk, a microblogging site, is taking off where Twitter is struggling.

Plurk offers services in 33 different local languages, including Russian, Hindi, Chinese in traditional and simplified character forms, Arabic, Hebrew, Greek and Japanese; with 45 more getting ready for release. While Twitter is behind in this language support race, they have quickly added Italian, French and Spanish within the last month in response.

While Twitter blames the difficulties of localising in Asian languages, Plurk seems to have solved the same issue with their language strategy: they send out an e-mail with a string of English to volunteer translators to localize the content and send it back. Head translators lead teams of users, and they vote for the best language usage when they run into unusual English slang or a new phrase. Many translators come from the open-source community and are willing to work without pay. Their only reward, aside from Plurk in their own language, is a virtual Rosetta Stone badge they can put on their Plurk user profile.

There is a surprising amount of people who have a strong desire to localize for free a site to have it in their own language. Plurk is not the only such website - Facebook is also dependant on crowd-translation.

Beats me why. Maybe I am a greedy capitalist, or just read in too many languages to bother translating for free, or maybe it is a generational gap. The motto seems to be "If you preach information for free, you need to slave for it". Fair enough!

But it doesn't always work well - one irate user wrote: "This is complete nonsense. It isn't localized in 33 languages or anywhere near it. The only localization I can see is for the word "is" in the two languages I checked - and in one of them (Irish) it's syntactically incorrect as the developers have failed to understand the most basic concepts of internationalization. All BS." :-D

Still some work left for us, professional, paid translators. Hurray!

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Monday, July 13, 2009

Apologia Machina

Agence France-Presse is telling us that the French business newspaper La Tribune is launching a multi-lingual version of its website using automatic translation, dispensing with journalists but producing often comic results.

"Ryanair loan to make travel of the passengers upright," "The Chinese car in ambush," "Internet Explorer: mistrust!" and "Assets of the continental right in management of the crisis" were some of the mysterious headlines the same day on the site, which is still in an experimental phase.

But the paper's editors are confident that the project will, once the software is refined and a human hired to tweak the texts, open La Tribune to a potentially huge international audience. The editors acknowledged that the results on La Tribune were still far from ideal, but said that the software was being continually updated and that within a few months it would achieve "almost perfect" news articles.

*****
Now comes the stick-it-in-apologia-for-us-human-translators:

But it is generally accepted that translation software cannot, as Google Translate admits on its site, "approach the fluency of a native speaker or possess the skill of a professional translator".

Why do I call it an apologia? BECAUSE: "Most of the English articles on La Tribune's site were, with a little effort, understandable despite their many linguistic oddities."

We are willing to pay less and put "a little effort" instead to read garbled English which no longer gets defined as "incorrect" but as "odd".

Except that I wouldn't get paid for being "odd", but some idiot would pay hundreds of thousands of dollars to purchase a software that is. Amazing!

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.

Friday, May 08, 2009

Gripping About Translators

From the Masks of Eris (Finland):

"I acknowledge translating stories is difficult, okay? A machine can translate 80% of the content/words and 10% of the mood/associations, but (at the moment) it takes a very skilled, careful and thoughtful human artisan to drag both to the nineties (...) the best possible translator would be a foreign mirror image of the original author (...) You have a chimp translating Shakespeare, you get chimp adventures; you got a Goethe translating Shakespeare and you get Shakes-worthy results. You got a Goethe translating chimp scribblings, I don’t know what the heck you’ll get. Maybe a book better than the original."

Apparently, he would have to come up with something Faustian, he would.. No excuse for having a bad translation, that.. err.. chimps scribblings.

For all your English to Arabic and vice versa translations that will help you expand your business into the Middle East visit Arabic Language Experts at http://www.arabic.com.au/.