Lexiophiles’ Top 100 Language Blogs 2010

Lexiophiles has just announced the Top 100 Language Blogs 2010. The list continues to be dominated by blogs about language learning and language teaching, but it also includes several  good translation blogs (About Translation didn’t make the cut, this year). The translation blogs included among the Top 100 are:  

• 6. Atominium the art of translation
• 9. Beyond Words
• 10. Translation Guy
• 21. ATR Blog
• 22. Translation Blog: Trusted Translations
• 36. Translation Tribulations
• 40. Algo más que traducir
• 42. Words to good effect
• 53. Blog de Traducción: Trusted Translations
• 56. JodyByrne.com
• 66. Medical Translation Blog
• 70. Trad Online
• 76. Spanish Translation Blog
• 79. Thoughts On Translation
• 80. (Not Just) Another Translator
• 94. Les recettes du traducteur
• 97. wéb-tränslatiôns

Congratulations to all the blogs selected!

Check them out, if you didn’t know them already: you’ll surely find some new interesting blog.

Voting is under way for the Top 100 Language Blogs of 2010

As in 2008 and 2009, Lexiophiles is choosing the Top 100 Language Blogs, divided in four different categories: Language Learning, Language Teaching, Language Technology and Language Professionals.

For each category 100 blogs have been shortlisted, and About Translation is included in the “Language Professionals” category. You can vote for one blog in each category.

50% of the final score will be based on user votes; voting started on May 12th and ends on May 24th. Winners will be announced on May 28th.

If you like this blog, please do vote for About Translation (you can do so by using the button below or the one on the top right of the page, and then select the radio button for About Translation in the Lexiophiles’ Language Professionals page), but make sure to also check the other fine blogs listed: you'll probably find some interesting blog you didn't know before.

Logical word order

It is sometimes easy to be misled by he word order of the source text, and to translate using a construction that means something different from the original.

From a contract I recently edited:

English: “Please read the following penalty schedule carefully”
Italian: “Leggere le seguenti informazioni sulle penali con attenzione”

Here, the position of “attenzione” is only awkward, rather than misleading. It would be improved by moving the word closer to the beginning of the sentence: “Leggere con attenzione le seguenti informazioni sulle penali”.

However, in other instances the word order might mislead the reader, even if only for a moment:

English: “… [of the] electronic end user agreement…”
Italian: “… dell’accordo di licenza con l’utente finale elettronico…”

Strictly construed, this translation might be interpreted as “…of the license agreement with the electronic end user…”.Since we do not have “electronic end users”, “electronic” in the original can only refer logically to the agreement; meaning that the agreement appears online or in some electronic media, such as a CD or DVD.

The source text should therefore have been translated as “…dell’accordo elettronico di licenza con l’utente finale…”, or maybe “…dell’accordo di licenza elettronico con l’utente finale…”, but certainly not *“… dell’accordo di licenza con l’utente finale elettronico…”.

Pay attention to the logical word order in your translations: when you read with fresh eyes what you wrote you'll sometimes see it means something different from what you intended.

Which free machine translation works best? The results are in

Some time ago I wrote about the study that Chinese translator Ethan Shen was conducting to compare three different free MT engines (for my earlier articles about this study, see Google, Bing and Babelfish and Google, Bing and Babelfish: some preliminary results).

Ethan has now completed phase 1 of his study, and the results are both interesting and - for me, at least - unexpected. Here below you can read a short report on Ethan's study.

From Ethan’s website you can download the full report, if you prefer to have all the details.

---

Real World Comparison of Online Machine Translators

by Ethan Shen
Gabble On Research Project
research@gabble-on.com

Abstract

This paper evaluates the relative quality of three popular online translation tools: Google Translate, Bing (Microsoft) Translator, and Yahoo Babelfish. The results published below are based on a 6 week survey open to the general internet population which allowed survey takers to choose any language, enter any free-form text, and vote on the best of all translation results side-by-side (www.gabble-on.com/research). The final data reveals that while Google Translate is widely preferred when translating long passages, Microsoft Bing Translator and Yahoo Babelfish often produce better translations for phrases below 140 characters. Also, in general Babelfish performs well in East Asian Languages such as Chinese and Korean and Bing Translator performs well in Spanish, German, and Italian.

Results

Most Preferred Engine and Margin of Preference by Language Pair and Text Length

The above table describes the relationship between user preferences and translated text character length for 15 single direction languages pairings. The most preferred engine is given at each intersection (Google, Babelfish, or Bing) along with the magnitude of its lead over its closest competitor in that category (colored percentage). The language pairings excluded from this table represent sets for which preferences were overwhelming (over 100%) or insufficient data was available.

From this data, the following conclusions can be drawn:

1. For long passages of text up to 2000 characters, survey takers generally prefer Google Translate's results across the board.
   
   a. The extent of Google’s lead varies dramatically from language to language. In some languages such as French, the strength of Google Translate’s engine is overwhelming. However, in several others like German, Italian, and Portuguese, Google holds only a very slim lead when compared to its biggest competitors.
   
   b. These observations validate our Hypothesis 1 that no single engine can perform equally well across a spectrum of languages or conditions.
   
   
2. The greatest relative strength of statistical translation focused engine (Google Translate) has not clustered around the European Union working languages as expected. German, Italian, and Portuguese, all EU working languages are the most hotly contested from a performance perspective.
   
   a. One possible explanation is that large additional bodies of parallel English-French text are available from the government of Canada for which are official documents are translated into both. To a lesser extent this could explain the strength of Google Translate in Spanish as many Latin American country offer English Translations of official documents.
   
   b. This data partially refutes Hypothesis 2.
   
   
3. Traditional Rules Based Translation Engines (Babelfish) performed generally well in East Asian languages such and Chinese and Korean.
   
   a. One possible reason for this outperformance is likely that the language specific grammar and word usages rules are more effective that association based transliteration in these situations.
   
   b. These finding are in line with Hypothesis 3, but the size of the data set is not large enough to confirm in a statistical significant manner.
   
   
4. Across almost every language Bing Translator and Yahoo Babelfish gain ground or surpass Google Translate as the text length gets shorter.
   
   a. In Chinese, the gradual erosion of Google relative performance as total text length shrinks from 2000 characters to 50 characters is stark and representative of the comparative strength Rules Based or Hybrid Translation Engines as phrases get shorter and more straight forward.
   
   b. It appears that at 150 characters or less, the fiercest competition between performance of different translation models become the most heated. Some similar effects were seen at 200 characters, but to a less significant extent.
   
   c. Though data is not shown, a similar effect is seen for passages that are only one sentence compared to passages with multiple sentences
   
   d. This data strongly validates Hypothesis 4.
   
   
5. The most interesting observation is that translation quality is not a two way street. The engine that is best for translating in one direction is not necessarily the best tool to translate back the other way.
   
   a. The two most obvious cases of this are French and German. Though Google Translation dominates when translating both these languages to English. It faces heavy competition when translating back from English to the foreign language.

These results are taken from a longer full research write-up.
To read the hypothesis, experiment design, extended results, practical applications and references, the full report is provided here: http://www.gabble-on.com/files/phase1_full_research_report.pdf.

Idiomizer: a wiki for idiom translation

I recently received an announcement about Idiomizer, a new wiki site that collects idioms in many languages and their translations into other languages.

According to the announcement,

IDIOMIZER is a translation
reference for idiomatic exchange across languages, absolutely vital
because different cultures use different phrases to impart the same
meaning. One of its many powerful features is the ability to view
multiple languages simultaneously. 

IDIOMIZER seeks the input of translators, linguists
and language lovers worldwide in making the site a useful and enjoyable
tool and we encourage you to register and add to our idioms.

Registration is required, but free. Once you have registered, you can search through the idioms, add new idioms, add translations and definitions, and generally contribute to the site.

Perhaps this is not “absolutely vital”, but it should prove a useful addition to many a translator's toolbox.

The launch of Translation Workspace ("...you can start working for free")

What first comes to mind when someone tells you that “you can start working for free”?

My first thought was “start working but not get paid for it”. I know, Lionbridge didn't mean it that way in Monday's promo offer for Translation Workspace. They meant that you can start using their new translation platform without paying for it (but only until the end of June).

Given, however, that you will have to pay for Translation Workspace to continue translating for Lionbridge (doing work that you used to be able to do with the clunky, but free, Logoport), the first interpretation contains at least a grain of truth.

Lionbridge claims that you will be able to use the new system for other translation projects, no longer for Lionbridge jobs only. Since most professional translators already own one or more translation memory programs, having to pay for an unwanted extra tool is an unappetizing prospect.

I doubt that many translation companies will switch from other CAT tools to Translation Workspace: apart from technical considerations (why move to a tool that uses the MS Word interface, when most other CAT programs are moving away from it?), Jost Zetzshe mentions another issue in his latest Tool Kit:

I foresee a huge problem once Lionbridge starts talking to other LSPs, who are of course direct competitors. I imagine a response something like this: “They want me to give them a month-by-month rundown of how much I translate?”.

Since the monthly payments depend on the number of words handled, Lionbridge would know how many words each translation company runs through the new platform.

The same for freelancers, of course. Would you tell your customers what percentage of your turnover they represent? Think how such information could be used against you: if Lionbridge knew that most of your work is with them, they would be in a better position to play hardball when negotiating rates or demanding discounts.

For more about Translation Workspace, see my previous post: Lionbridge’s Translation Workspace: my thoughts.

 

---

I begun writing this post Monday morning, after receiving Lionbridge’s promo offer for Translation Workspace. In the evening, I took part in the CTA “Thought Swap”, where some other questions were raised about Workspace. I have now added below some of these questions, with my ideas about their answers:

• Is the data safe? Won’t we risk Lionbridge having access to the memories we use for other customers? From what I hear, such concerns are unfounded: the servers in which the data is stored are under third-party control, and Lionbridge will not have access to other users’ private memories and data.
• Is this an attempt by Lionbridge to monopolize the translation market? Even such a large player as Lionbridge has not a big enough market share to monopolize the market. According to Common Sense Advisory “Ranking of Top 30 Language Services Companies”, in 2009 Lionbridge was the second largest translation company in the world by revenue. The market share of the thirty largest translation companies combined,  however, was only about a quarter of the global market. Our industry is still very fragmented, and no player is able to monopolize it. Translation Workbench is certainly aimed at improving Lionbridge’s position. Whether it will succeed, is an open question.
• Are they going to use the data to feed machine translation? No way of knowing for sure, of course. Many translation companies (and other companies as well: see Google) are mining the data they own to build up translation memories and feed them to statistical MT systems. This is going to continue and increase in the future. I believe that translator will have to adapt to this and learn to use MT as a tool (just like we did with TM earlier).
• What about LSPs (or freelancers) who work mainly for Lionbridge: won’t they be forced to adopt Translation Workspace? Companies and people who rely for most of their income from a sole customer are in a much weaker position when trying to resist that customer’s demands.
• Will we have to pay for the words we translate in Translation Workspace for Lionbridge? (That is, will the words translated for Lionbridge be applied against the subscription?) No. According to all Lionbridge’s material on Translation Workspace, work done for Lionbridge will not count against the words purchased with the subscription.
• What about words translated for Lionbridge, but indirectly (for example, an LSP accepts a project from Lionbridge, and then assigns it to freelancers): will these be applied against the words paid for? Probably not. The LSP should be able to assign the work to its freelancers in a way that does not otherwise affect their TW tenancy. It might depend on how the projects are set up, though.

---

 Please note: all of the above is my own opinion and interpretation, based on the information I have. It does not reflect or represent in any way Lionbridge's official position. I am not affiliated with Lionbridge, and have no access to any Lionbridge insider or confidential information.

New book on the business of freelance translation

Judy and Dagmar Jenner have just published “The Entrepreneurial Linguist: The Business-School Approach to Freelance Translation”.

Judy gave a presentation on her “business-school approach” to freelancing some time ago, at a CTA session, and I was impressed how helpful her approach was. The book is available both as a paperback and as a downloadable e-book. I just ordered my copy, and plan to review the book here, as soon as I have a chance to read it.

Congratulations to Judy and Dagmar!

Watercooler is going to change its home

You may have noticed the link to Watercooler is no longer displayed on the right. Andrew Bell is planning to move it from its present location as Ning network to some other service (this because of the changes that are happening at Ning).

In the meantime, the link to Watercooler.com would no longer work. For the time being Watercooler can be accessed at the following address: http://translationandlanguage.ning.com/?xg_source=msg_mes_network

A great resource for translators

I don't remember if I mentioned it before, but if you work in translation and want to stay up to date with what is happening technically in our field, what new tools are coming, what are the best tools for our job, and what is the best way to use the tools you already have, a great resource is Jost Zetzsche's The Tool Kit, a computer newsletter for translation professionals.

Time estimating for dummies

A brief glance through the message board of most translators’ portals will find dozens of messages like this one:

I'm a beginner, can anyone tell me how many pages usually a translator should asked to translate a day?

The messages may differ slightly, asking for words or lines instead of pages, but they are all essentially the same: “I don’t know what I’m doing: how long will it take?”

If you are a real beginner, you may really not know, but asking others is not going to help you either, even when other translators provide an answer: knowing that others translate 2,000 words a day on average, doesn’t mean you’ll also translate at that speed.

As a beginner, you need to learn how to estimate how long you’ll take to do a job:

• If you don’t have already a word count for the job you are going to do, first count how long the text to translate is. If it is a hardcopy or pdf job, a rough estimate (to the nearest 100 words) is good enough.
• Then time yourself carefully, starting from the moment you begin a job and until you deliver it to your customer.
• Include in the time you count all the time you spend on the job (including time spent researching the assignment, translating it, proofing it and entering it in your accounting system).
• Also include in the time count any short breaks you would normally take. It is tempting to stop the watch any time you go to the bathroom or answer the phone, but it would be wrong: if these are normal activities, you need to include them in your time reckoning.
• Do not include in the time count any time spent on other translations you might be doing in addition to the job you are timing, and do not include really major interruptions that would not normally happen.

At the end of the job, a simple division will tell you your hourly speed: if the job was 1,800 words and it took you 6.5 hours to complete, your speed was 1,800 / 6.5 = 277 words per hour. Now from your hourly speed you can calculate your daily speed for that job: 277 * 8 = 2,216 words / day.

Continue to do this until you have a good idea what your speed is under different conditions: different kinds of assignments, using different tools, and so on. Repeat this exercise from time to time, to make sure your statistics are still valid, and repeat it again any time a major change happens in your routine or in the tools you use.

And, please, don’t fool yourself that if you normally can translate 2,200 words a day, you can accept that tempting 8,000-word assignment due tomorrow, if only you can stay awake long enough: you may be able to get away with it, some times. Disaster will strike in other occasions (and customers will remember).