Today, we had our first live day of simultaneous interpreting, with a guest speaker. The manuscript he gave us beforehand would take about 15 minutes to read, and he spoke for about 90, so we got some good practice at interpreting without a script.

The course has been a great learning experience so far. For one thing, it's shown me that I need to work on expressing myself in Japanese.

The techniques we're learning for simultaneous interpreting between Japanese and English have been illuminating. Japanese and English are very different structurally, and usually in order to translate a sentence, you need to look at the end.

Here's a contrived example, with the parts color-coded:

新しいバージョンは、Xがあって、Yがあって、前のバージョンより機能性が高くなりました。

You've got a topic, a bunch of "modifiers," or supporting information, and then the predicate. Normally, I'd translate that sentence something like the following, moving the predicate to the beginning, and turning the supporting information into one or more subsequent sentences:

The new version has improved capabilities over the previous one. It has X as well as Y.

This of course won't do with simultaneous interpretation. Especially considering that "X" and "Y" could be of arbitrary length (it's not uncommon for the topic and predicate to be far, far separated, even in spoken Japanese), if you waited until the end, you'd forget the sentence. Or even if you didn't forget it, while you were processing this sentence the next one would be flying by.

So they're teaching us techniques to kind of listen until you get a meaningful "chunk" of information, and discharge that while you listen to the next chunk. Here's an example from a passage where I can still remember how I interpreted it.

Japanese:

(1)バングラディッシュ　チッタゴン大学(2)の経済学教授ムハマド　ユヌス(3)と学生達が、ある村を調査している(4)最中に、素朴な竹製家具を作っている女性に出会いました。

My English interpretation:

(1)Chittagong University in Bangladesh (2)is where professor of economics Muhammad Yunus works. (3)He and his students were on a field trip to a village. (4)There, they encountered a woman who made simple bamboo furniture.

We always do a lot of preparation before each piece. In this case, the preparation included the Wikipedia pages on Grameen Bank and Muhammad Yunus, and the original English version of the article. Incidentally, the English version started, "During a field trip to a village, Muhammad Yunus…"

Tomorrow, we get to interpret for a mock trial, where a GI has been arrested for theft and assault. Very Okinawan topic!

I'm not really interested in becoming an interpreter, simultaneous or otherwise, but the profession has always fascinated me — simultaneous interpreters are kind of like the thoroughbreds to the slow-but-steady translator plow-horses.

My main goal for the course is to stretch myself in new ways. In particular, I'm hoping to train myself in interpreters' ability to think on their feet. I also am looking forward to going both from and into Japanese. I've lately been trying to write more Japanese, and study more translations from English into Japanese, and I think that this has increased my toolset for translating into English.

The course also promises to teach some people skills: everything from eye contact to correct posture (urp). I'm looking forward to an exhausting but valuable eight days.

Earlier in my career, I tried to stick to the style and terminology of the reference translations as much as possible. I figured that since the client had accepted those translations in the past, that was what it wanted. Even if the translation was very bad, I'd just plug my nose and send that pile off.

That turned out to be a bad idea. If I passed the trial, I would then need to translate 10 or more times the volume of the trial using that same painfully bad style.

After a few years, I started ignoring these "reference" translations, and doing what I thought was the best translation. I might look at references to get a feel for terminology, but I'll also diverge from the terminology in the reference if I think it's wrong.

I figured that even if the client didn't agree with me on translation style, at least this way I could look forward to passing the trial with anticipation instead of dread. I work almost exclusively for Japanese companies, and I've also found that I don't want to work for clients that prefer brain-dead chokuyaku. We're both better off if those types of clients find translators who write English just like in their middle-school English textbooks.

It also turns out to be a lot easier and less stressful to do what you consider to be your best work, rather than second-guessing your translation, wondering if this is what the client wants.

To my surprise, I found that after I took the policy of ignoring reference translations, I started passing a lot more trials. But when I thought about it, it made sense: the client was looking for a new translator because they weren't happy with their current provider. In retrospect, it seems like a pretty bad idea to copy the work of the guys who just got fired.

So, back to the trial from a few weeks ago. The prior translations were very bad, and since I'm knowledgeable about the subject matter (software specifications), I ignored them completely and did what I considered to be a good translation. Fortunately, the client not only liked the trial: they were thrilled with it, and although they were originally planning to place a 10-page job, that has turned into a 30-page job, with a promise of more work to come.

And the best thing is that I can use my own judgment to produce the best translation possible.

For some reason, the accepted translation for this term seems to have become "recycling oriented society" (it's the only one listed in the venerable Green Goddess).

The problem is that you almost never see this term used in genuine English documents. Since it doesn't seem very likely that this concept is unique to (or even originates in) Japan, I'm always on the lookout for a better way to express it.

So far, the best English equivalent I've found is "closed loop economy" ("economy" often serves when the Japanese talks about "society"). Unfortunately, clients often aren't adventurous enough to stray from the anointed dictionary translation, so I tend to keep a little collection of links that support my use "non-dictionary" translations.

Here's a comic where I spotted "closing the loop" being used in much the same way as Japanese uses 循環型社会:

We must "close the loop" and make all products
recyclable, reusable, or compostable

They recently contacted me about taking part of a big job, and agreed to my usual rates, so I accepted.

Big, BIG mistake.

Firstly, they sent me the file, saying it was about 3 pages. Then they immediately sent me three other files, saying they were higher priority, so I put the first file on the back burner and did the high-priority ones.

Then I open up the "3-page" file and found that it had a lot more than three pages. I called them in a panic, because the translation was due the next day.

Me: This file has a lot more text than 3 pages. Is this the right file, or do you just need part of it done?

Agency: Whoops, sorry. I left off a "0″ from the count. It's actually 30 pages.

Me: Holy crap, this job is due tomorrow and it's 10 times what you told me. But actually, the file has way more than even 30 pages.

Agency: Well yeah, actually it's 90 pages, but 60 pages worth are repetitions, so we just count the new stuff.

Me: ::boggle:: … Leaving aside how hard it makes it to schedule my time when you give me a word count 30 times lower than actual, I suppose that this means that you don't plan to pay anything for the repetitions that I translate.

Agency: Of course, since it's just a search and replace…

Me ::double boggle:: … Oh and by the way, I notice that you've done some sort of search and replace from your glossary, inserting English words into the middle of the Japanese text (and incidentally corrupting the file due to something wrong with the tool you used). I suppose that you don't plan to pay for the Japanese words that were replaced?

Agency: No, of course not, since we only pay for translation, and those words are already translated for you.

There's no way I'm going to do 90 pages in a day, no matter how much of it is "just search and replace" and helpfully interspersed with horrible translations from a glossary with sometimes comical results, but I agreed to do one third of the job because I made the commitment. But I am truly flabbergasted that a reputable translation agency could think that this is a normal way to deal with translators. Needless to say, this is the last job I will do for this agency.

My friends asked me if I was going to take the call. I pulled out the phone, and saw it was from a translation agency, so I put it on "manner" mode (vibration mode), and said, "No, I just put it on manner mode. Actually, I wish this translation agency would learn some manners."

Then I looked down at the phone, and saw that I had actually taken the call instead of putting it on manner mode, and the agency had heard everything I'd just said. Oops!

I thought about apologizing, but then I thought that what I had said was exactly right, although perhaps not delivered in the best way. So I just hung up and went on with my evening.

When I got back home at around midnight, I checked my email before going to bed. There was a job request from the agency that had called me. I emailed back saying that I could take the job, and within five minutes had a response thanking me. I guess the agency was really desperate to place that job! I think it still doesn't warrant calling someone at 10pm, though.

I very much doubt if the person from that agency reads this blog, but if you do, sorry for the manner remark. But you really shouldn't be calling translators at night. Email is fine any time.

I was happy to take on this project, partly because Murasaki is something of a legend in the Japanese rock scene (and especially here in their home prefecture of Okinawa), and partly because it would be a good chance to test out some ideas I've been having about getting simple, clean websites up quickly and efficiently.

There was a firm deadline for getting the site up: a major Japanese heavy-metal magazine (Burrn!) was going to feature Murasaki, and they were going to list the band's website. My strategy was therefore to get something simple and working up by the deadline, and then add to the site gradually.

The Tools

django

I'm partial to cherrypy/jinja2/SQLAlchemy as my web stack, but for this project I chose django for the website backend, because django is known for making bog-standard stuff very easy.

The built-in admin interface was a big win for django, because it allowed the band to get in and start adding information right away, as I worked on the site design and structure.

A major annoyance, however, was the inability to smoothly evolve my database. django won't modify database tables for you if you add a field to one of your models, and since I'm growing the site organically, I'm faced with either mucking with the SQL by hand to modify my tables, or violating YAGNI by adding all the fields up front, even if they're not used now (and might never be used).

A good example was lyrics for song tracks. At first, I didn't add a lyrics field, because we weren't sure if we were going to publish them. But then the band's manager decided that it would be OK, so I had to modify the database to add the field. The next issue is samples for tracks; the band thinks that it would be a cool idea, but hasn't worked out if they want them. So that will be another modification to the database if they're added.

I think that django's insistence on making database evolution as hard as possible is wrong-headed and violates the Pythonic principle that we're all consenting adults. Modify the damn table for me, and let me deal with the consequences!

I'll note that although there are a couple of django evolution apps, the ones I checked didn't support sqlite databases (which I'm using because the database is currently around 100 KB, and I don't expect it to ever pass 1 MB).

jQuery UI

I chose jQuery UI for the site layout. The Theme Roller application for designing a custom theme was a huge win; I was able to get exactly the look and colors I wanted in minutes.

One small issue I had was that the UI is a bit too coupled to the JavaScript end. For example, I decided to use the tabs widget for the navigation menu, but out of the box, this widget makes you load all the content into the page, and switch using JavaScript. I had to muck about in the CSS file in order to have the tabs link to actual URLs.

I also benefited from a couple of other jQuery components, such as jQuery galleria for the photo gallery (although I'll probably move to something else as more photos are added, and I add a tagging feature).

WordPress

I used WordPress to create two blogs for the site (one English and one Japanese). Yes, there are blog solutions for django, but WordPress is ready to go out of the box, and in this case (a tight deadline and budget), practicality definitely beat purity.

I created themes integrating the blogs into the site design using Themes Press — that was $10 very well spent! I doubt that even an experienced wordpress hacker could create an integrated theme for less than $10 of their time.

Overall Impressions

I think that the combination of django and jQuery UI can make designing a simple site very quick and painless, although django has some issues with organic site growth, and jQuery UI couples the CSS and JavaScript a bit too tightly for me.

There were a couple of other minor issues (such as when my approach to internationalization differed from django's), but they were probably due more to my lack of familiarity with the frameworks than inherent limitations, and I found workarounds for all of them, so no big deal.

It was also a good experience for me to step away from my more familiar combination of cherrypy/jinja2/SQLAlchemy to use django. I think that in cases like this one, where you have non-programmers contributing a fair amount of the content, the advantages of the free admin interface outweigh the loss of flexibility from using an all-in-one stack like django.

A couple of weeks ago, someone close to me died after a long fight with cancer. When we both lived in mainland Japan, we'd talk about moving to Okinawa, and living a more relaxed life.

One day, I packed up my family and moved here. Since he'd have a lot of trouble finding a job with equivalent pay in Okinawa, his plan was to wait until retirement, sell his house, and use that and his retirement to supplement his income. He died 3 years before reaching retirement.

I learned from him that life doesn't always go as planned, and we've got to do things when we can. I feel fortunate that my job allows me to live where I want, and have a lot more control over how much I work, when I work, and what kinds of work I do.

I've always been a kind of impulsive person. I think that mulling over big decisions too much makes us unhappy about the outcomes. But I've made a renewed commitment to get out and try new things; to stop waiting until the time is right, because it never will be just right. And we don't have all the time in the world.

Here's the project URL: http://bitbucket.org/ginstrom/mailer

Apparently you can't open write access to everyone, so if you'd like write privileges, then please let me know your account name.

About mailer

mailer is a wrapper around the Python email and mimetypes libraries.

Link to broadcast feed

Kudos to Mike Sekine and the others who made this possible.

If you translate between Japanese and English, but didn't get a chance to go to IJET, then by all means check it out.