Tip:
Highlight text to annotate it
X
>> Welcome, everybody. Today's TechTalk is on Writing for Translation or World-ready
Authoring. Our speaker is Bjorn AUSTRAAT. He--has dual degrees in Translation and Interpretations
and taught at the Monterey Institute, and has consulted for 15 years to companies doing
Internationalization of their Websites, of their software, and documentation. So please
also remember that this talk is being recorded and save your Google specific questions till
the very end after we stop filming. Thank you and here's Bjorn.
>> AUSTRAAT: [SPEAKING IN FOREIGN LANGUAGE] Okay. Yeah. So how did that feel?
>> [INDISTINCT] >> AUSTRAAT: [INDISTINCT] Yeah.
>> I understood TechTalk. >> AUSTRAAT: Edelweiss, Yeah. So that's what
its like for sometimes for translators or even speakers for--of a Foreign language to
consume things that weren't written or designed to be consumed in that language. And even
though translators are of course highly trained individuals that have degrees and--or you
know bilingual as much as you can possibly be bilingual, sometimes it's very difficult
for them to understand the source copy because of certain speed bumps, that I call them speed
bumps. Translation speed bumps that may exist in your collateral, or on your documentation,
online help in you user interface. And today's talk, even though Google of course is already
very world-ready, because you translate already into dozens of languages and you have 50 or
60 interface languages you can select for the Google Search--the main Google Search
interface. There are still some things that could be finessed a little bit. And, so today's
talk is not so much about, "Oh my God, this is horrible and here's--you need to completely
change what you do." But this is really like a coaching session to help you improve slightly
on the--on the great work you already are doing. So what is World-ready Authoring? It's
really about increasing the portability of your output, cultural and linguistic portability.
And because in modern, high-tech businesses, English is just another language it does really
pay huge dividends to design and author and think about translations before you even write
the original. And, so this why--I--the second bullet point or a sub bullet point it says,
Translations starts before you write the original. Which seems counter-intuitive but it really
becomes clear when you think about translations and localization as an--it's not a feature,
its not something you can tack on at the end, it's like an Architecture. And if it's not
architected into the whole document and Electronic Collateral Equation process it's going to
be very expensive later to fix it. Just like in a house if you don't put the pipes in at
the beginning and--or you have to tack on an elevator later it is usually a lot more
expensive, than if you had planned for those things in the beginning. So the US English
typically in a multi--in a national multi-lingual, global company like Google has to strike a
balance between the hippest possible ethno-centric, culturally super-duper, for American audiences
version, and a more linguistically portable version which is taking into account the 20
or 30 or 40 other languages that are still ultimately going to be consumers of that interface.
So it touches really everything the layout, the composition, the style, the Font choices
and so on and I will go in to much more detail in my presentation on what are some of the
elements of world-ready authoring. Well, some of the benefits of world-ready Authoring.
Well, if you have 40 languages and if you fix it once in English, you'd have to fix
it 40 times that's very simple multiplication math that if you remediate it once you don't
have to then fix it over and over and over again in the translated copy. The translation
quality goes up and the source text becomes easier to understand, also for US consumers.
Oddly enough writing for translation, good writing for translation is actually typically
good writing and just period. And, so how many of you had read Strunk and White? Yeah.
Old school. Strunk and White, you know page four, the clarity and conciseness that also
benefits translations is actually one of the hallmarks of good English writing as well.
And one very important, very bottom line centric aspect of writing world-ready--copy and world-ready
to utilize, is that it typically takes about 20 minutes to fix one of those or remediate
one of those problems. And the 20 minutes include by after--you know, we place of call
to the translator or to engineer then I have translate engineering use into English, and
then I have to communicate with the translator what that--was to--what that button label
meant and then get it translated and fixed and updated, and God forbid, it was UIElement,
so now I have to retake screenshots and very quickly one simple mistake can add up to 20
or 30 minutes of fix time, so even if you have just two per page on a hundred-page document,
in 40 languages it's actually quite good change of money to go through that and make it a
word-ready copy and that--you can save by doing that, if you do it once at the beginning
of the process. Do you all know what a locale is by the way? Because I mentioned that, its
one of those translations technical terms, what's the locale?
>> Its one of the reason [INDISTINCT] used to describe which dialect you're using. Because
for example at the [INDISTINCT] things got sense [INDISTINCT].
>> AUSTRAAT: Correct. So, the answer was said. It's a particular code that it tells you what
language and geographical location this is designed for. So for instance Spanish, Spanish
versus Mexican, Spanish. And there's also this great construct that all translators
that are really, you know, trained translators love which is North American Spanish, just
a great--it's a construct that doesn't really exist but for convenience is used to cover
the Spanish speaking population in North America. And if you're--I've lived in Los Angeles,
there's another language now, it's called Spanglish and it's truly an emerging language
that probably in a 100 years is going to be, like, Dutch, you know, it's just going to
be another language, where church is, for instance translated in churche, and bakery
is the bakarea and so things interesting, you know, a new language emerging there in
Los Angeles. So I mentioned there's quite a few components to world-ready authoring.
And one of the really important ones that is going to get really expensive and hard
to fix is layout. So for instance translated text is almost always longer than the original
text and the text swell is going to be between 40% and 300%. This is and actual German word,
I looked it up its on wikipedia, you can check this and it says, [SPEAKING IN FOREIGN LANGUAGE].
>> [INDISTINCT] >> AUSTRAAT: It starts with Beef. [SPEAKING
IN FOREIGN LANGUAGE] is labeling, [SPEAKING IN FOREIGN LANGUAGE] is monitoring, [SPEAKING
IN FOREIGN LANGUAGE] are the tasks, [SPEAKING IN FOREIGN LANGUAGE] is transfer and [SPEAKING
IN FOREIGN LANGUAGE] is law. And, so German is one of those great compounding noun languages
that you can make arbitrary long nouns, and well, this is of course an extreme example,
German words typically are longer than their English originals, and as are Spanish words.
So because there are--there's going to be more words and they're going to be longer,
one of the things that are really hard to deal with is a tightly constrained source
space, could be a button, it could be tab writer, it could be columnar layout. And,
so if you... >> Pop up windows.
>> AUSTRAAT: What's that? >> Pop up windows.
>> AUSTRAAT: Pop up windows, right or the Google box, the info box that is constrained
by law and the powers that be, in marketing it can only 120 pixels tall, but if we cram
it full--a truck full of goodies in English, it's likely not going to fit in the translation.
So another example is [SPEAKING IN FOREIGN LANGUAGE] or laptop and some Spanish flavors
that's, [SPEAKING IN FOREIGN LANGUAGE] so if you had a great little Map it button on
your Google maps and then you do it in German, good luck. And of course this means anybody
wants to take a stab at--this means--this long German word? So it's the law governing
the supervisory tasks regarding labeling beef, that's pretty much it.
>> So it's the FDA [INDISTINCT] >> AUSTRAAT: It's a sub-chapter of a very,
apparently very meticulous German Government agency that deals with nothing but supervising
how beef gets labeled and there's a law about it of course. So as an example I went to the
Google--the public Google site to look for some examples of what--this actually already
in effect today, and you can see that this columnar layout here, setting preferences
fits nicely in the bulleted form, but then as soon as you translate it into German, [SPEAKING
IN FOREIGN LANGUAGE] it does wrap and the German words are longer as well. So here its
okay, it doesn't look pretty but its okay. Where it's not okay if it's starts to have
constrained space and gets truncated, which happens sometimes in user interfaces that
are not entirely HTML or web centric. For instance on a regular windows dialog box,
if you chop off a piece of a word you could change the meaning and so then this becomes
a real problem. By the way, a lot of the examples are German not because the German translation
team is doing a poor job, it's because it's my native language. So, I felt most comfortable
selecting examples from that interface. So, here's some recommendations because it's always
easiest to say don't do this but obviously you can't prove a negative. So, here are some
of the things that you should do if possible which is to avoid a narrow or columnar layout.
It aids the eye reading online that's why the columns are used in web--on the web especially
because it's a lot easier to read a narrow column on a screen than it is a wide texted
scrolls across. This is also the main reason why large-format newspapers use a columnar
layout, because it guides the eye and makes reading faster. However in translation it
becomes more difficult. So, you should assume that for column width, that you would go probably
typically 60% wider than you would have in English to accommodate other languages. And
you can quickly try this out by simply going to a German newspaper website, cutting and
pasting a chunk of text into your layout and see how many words wrap. In translation, it
will hyphenate--the translators will be able to hyphenate, but if you have a sea of hyphens
it's also not very pretty. So, a quick and dirty trick is to just cut and paste text
from a German website or from the German Google site and see what that does to the word wrapping.
If it wraps in the middle of a word, it's typically going to be hyphened ultimately
and not a super great layout. So, white space is another very important thing especially
if one-to-one page correspondence is important, which sometimes it is because you're going
to produce a manual that is four or five languages, they all have to be in the same page layout,
the same page breaks. So, the graphics don't get messed up. So, if you don't leave enough
white space this will never happen and there's going to be very expensive lay outing that
has to happen to make that fit. And especially size sensitive collateral, quick start guides,
package inserts, spines of books, spines of boxes, spines of a Google appliance. And I
know that you put text on these sometimes, right? So, very, very important that you consider
that this is going to be--if it gets localized or translated, that you do have enough room
for the--without an excessive amount of wrapping and buttons and UIelements of course, the
worst offenders. Because buttons have a really hard time growing. In typographical conventions
one thing that English native speakers like to use a lot is capitalization for effect,
safe search, personalized search et cetera, right? Safe search is not only capitalized,
it's spelled together and what is known among encoders as the Hungarian or sometimes known
as the Pascal Notation. And so, if you see variable names sometimes or you see somebody
would write int, and then, you know, some upper case letter and so on, that means typically
it's a variable of the type integer and so for convenience people like to glue those
words together, capitalize the important words and that has a special effect, because it
creates the effect of the trademark or a branded name. Of course, in Asian languages and we
have some native speakers here, there is no capitalization because it's--their idiographic
languages, right? So, they just use the character symbols, so that is loss entirely. Also if
there are no spaces between words then putting two words together doesn't make a whole lot
of a difference. Also small font sizes may make characters really--a lot more difficult
to read than they are in English. For example, you see the first example--the first bullet
point here you have the regular A and then you have umlaut A with the two dots on it
and then you have an idiographic character which I have no idea what it means. Does anybody--can
anybody translate what that character means? >> Help.
>> AUSTRAAT: Help. I must have copy and pasted that from the website. So, at the same font
size note how it is a lot easier to read the regular, the English letter, the umlauts are
a little bit harder to make out and then the Asian characters are almost illegible because
it's much--it's a dense idiographic character. So, I have to go up in the font size to make
the Japanese or Chinese legible and--one second. And I can keep a smaller font size. So, now
imagine if I had densely packed small text in English it's going to get 40% longer after
translation and I have to bump up the font size by one click to make it legible. So,
now I have a real text wide problem, right? Because now I'm going to have suddenly two
pages worth of stuff and I have one page to fit it on unless I want to use the, you know,
pharmacological package insert font .6 with the reading glassed for free. Yes?
>> Is there any specific minimum size font that we should be using?
>> AUSTRAAT: Is there a specific minimum--yeah, I need to repeat the question for the audio.
So, the question was, is there a specific minimum size font, and the answer is no because
it depends on the language but I would experiment with as the lay outing person who creates
the original, I would experiment with different languages to see if the legibility is maintained
in that font size and if not leave more white space because then ultimately one of the devices
that the foreign language GDP person may have, that's the publishing person may have, is
to increase the font size to make it legible. So, it's just something to keep in mind as
part of the world-ready authoring approach. Yes?
>> Is bold generally just for emphasis in character based languages or is it not used?
>> AUSTRAAT: Is bold use for emphasis in character based languages? It is used and it is--sometimes
quotation marks are also used and underlining are used. On the web, underlining carries
a special meaning as in I'm a link, so that's a problem, right? But bolding is usually a
safer thing to do because it will carry well into languages that are not using the capitalization
or where you cannot use capitalization because it's meaningful like German.
>> And italics are pointless? >> AUSTRAAT: Italics are sometimes pointless
in Asian languages typically italics are used for different cultural contexts such as comics
and so emphasis which is--in English italics indicate emphasis may indicate a total cultural
shift or an inappropriate cultural shift in some Asian locales. So it's a tricky device
to use. So, here's some examples from the Google website, what is personalized search,
it's capitalized here for effect because it's a particular feature of Google and in German
that is--effect is entirely gone because now [SPEAKING IN FOREIGN LANGUAGE] is a regular
adjective, it's not capitalized and if I write that in it's German translation I would never
know the personalized search is a feature that you want to draw attention to because
that capitalization is really what that does, it draws attention to this word.
>> So, how would you suggest the translator to deal to with that, in other languages?
>> AUSTRAAT: That's--so the question was how would you deal with that in other languages
and the--there are multiple ways to deal with it. For instance bolding, a different font
type which is also tricky in Asian languages as you can imagine. And if it's a trademarked
term you may capitalize both words to indicate that this is a non phrase that belongs together
and it's a trademark term. If it's not trademark then you can't use that symbol of course and
it's a tricky issue to--it's usually marketing decision what is done with that, how important
is it, because also if you--if you capitalize the P here it will draw attention to itself
maybe unnecessarily so because it would be it--but on it's face value it would be a typographical
mistake. If it's not a trademark term capitalizing the P here is wrong it's an--it's a hard fast
error. >> [INDISTINCT]
>> AUSTRAAT: Correct. So, some recommendations again scaling font sizes to work well with
accented characters is one important consideration if you can't--if you don't want to do that
or can't in English just leave enough white space or use some of the other tools in the
arsenal to deal with the very, very likely text swell that will occur. Quotations marks
and bolding where feasible, because bolding carries well but again keep in mind you have
small Asian text and you bold it becomes illegible, you have to--need to scale it up a little
bit. And if you want to do yourself a favor, avoid creative fonts. I did one project in
a past life when I was working for [INDISTINCT] Translation Services and we localized the
Timon and Pumbaa CD-ROM games. Timon and Pumbaa was The Jungle King, you know, many moons
ago and they used a great and clever font called The Timon and Pumbaa Jungle font which
was old grubs and leaves and sticks and things like that. Of course it didn't contain a single
umlaut. So, then we sent a fontographer for weeks and made umlaut dotted O's and, you
know, strike through and Ñ's and something like that out of grubs and trees to get that
localized properly. >> Have you used coconuts?
>> AUSTRAAT: Coconuts. Yeah. Well, you know, there are so many foreign characters if you
ever want to see what's really in each font go to your start button and type and run type
"char map" and then scroll down and you'll be surprised how many different character
there are, that you need to recreate out from coconuts and trees and grubs to get something
localized. >> Do you have any international [INDISTINCT]
section they have a print out of all the extended character sets and it's something like 15
or 20 feet long, and three feet wide. It's really worth looking at.
>> AUSTRAAT: Really. >> [INDISTINCT]
>> AUSTRAAT: Yeah. And the reason why Unicode fonts are so large is because they contain
every single one of those characters. >> Is that a problem?
>> AUSTRAAT: Arabic, et cetera, et cetera, every possible accent of characters, strikethroughs,
you know. Every possible mutilation to a character that you can imagine is in there. Now, a very,
very, very important topic near and dear to my own heart, Glossarization. And this is
one of the core ways you can influence translation speed as a technical author. A document by
and large cannot be translated until it comes with a glossary. Okay. Very important. You
cannot just look it up. That is--if you ever say that catch yourself and say "no. oh, I
learned that". You can't just look up in the dictionary or online. You can't even Google,
just Google it. Why? Well, because without context, words have no meaning, very little
meaning, or many, many possible different meanings which is even worse for a translator.
And for example, the word set. Just a small sample of all the different ways you could
translate the word set into Spanish. It could be [SPEAKING IN FOREIGN LANGUAGE] if it's
golf clubs. [SPEAKING IN FOREIGN LANGUAGE] if its kitchen utensils. [SPEAKING IN FOREIGN
LANGUAGE] set of teeth. So you can imagine that this is just a subset because set, is
like looking up Smith in the phonebook pages, and pages and pages of translations. Time
depending again on the context, right? What time is it? [SPEAKING IN FOREIGN LANGUAGE]
How much time does it take? [SPEAKING IN FOREIGN LANGUAGE] How many times did something occur?
[SPEAKING IN FOREIGN LANGUAGE] So imagine you're a poor translator at 6:00 PM on Friday
and somebody's dumping a translation task on your desk of string tables. Have you--do
you know what string tables are? String tables are basically all the words that belong to
a piece of software and sometimes they pull all of those out into one separate file. And
so now, here you are sitting and trying to translate the word time or set. Well, it's
impossible. Example, glossaries are also great for translating the same word consistently
over time. And I found this particularly delightful example. So we have [SPEAKING IN FOREIGN LANGUAGE]
which is translates as home. This is both alive right now in your website, by the way.
Google [SPEAKING IN FOREIGN LANGUAGE] That's help. [SPEAKING IN FOREIGN LANGUAGE] That's
actually wrong. >> [INDISTINCT] Google as a noun. I'm sorry,
her, her. You're using >> AUSTRAAT: Were?
>> [SPEAKING IN FOREIGN LANGUAGE] >> AUSTRAAT: More ways to Google.
>> Yes. So you're using it as a verb. >> AUSTRAAT: Is it supposed to be a verb in
English? >> No.
>> AUSTRAAT: No, this could be a noun. Yeah. [SPEAKING IN FOREIGN LANGUAGE] is just--is
the standard German, all roads lead to Rome. So it is--it is a noun. So--but as you can
see even if you have the best translators in the world which I'm sure you do. Overtime
there are so many different ways you could translate the same word into different ways.
That a hard glossary becomes a real control on creativity and a welcomed control, because
this is confusing ultimately because I click--this is literally clicking back and forth between
two pages and getting two different Nav bars. >> [INDISTINCT]
>> AUSTRAAT: Yeah. These are--this is one link, one click away. These two. Yes?
>> I'm a little bit confused how can we can create a [INDISTINCT] glossary if I had to
define like the word set [INDISTINCT] all of a sudden...
>> AUSTRAAT: I'll get in to that, Glossarization recommendations. Yes? So here some recommendations,
how to make a good glossary? Include all, any and all companies and industry and products-specific
terms. Even if you think they are absolutely obvious because one thing that is not obvious
is that as soon as you work at any place, not just Google, any place for more than a
week you'll start to wear that place's verbal uniform. And verbal uniforms work like this
they are used as a social marker. If you speak a lot of technical terminology that only your
peers understand, it raises your social standing, you may have noticed this in talking to some
people. Especially if you know a lot of acronyms these, are like metals on the verbal uniform.
And it's great as a--it basically accelerates communication that has its place. Doctors
talk in technical jargon amongst each other because it's more efficient to communicate
this way. They can say this particular nerve, Latin word. Instead of saying, well the nerve
do you know that runs from here and all the way down here and comes out your ankle, because
that's not very efficient. However, as soon as you leave your social group then it becomes
manipulative or ineffective. And so my favorite sentence when I ever ask a technical person
to explain something to me is to say, "In a language a 12-year old could understand
and in 25 words or less what is a compiler?" And that makes great glossary entries. When
you set up like this mentally. So since we have a little bit of time. In a language a
12-year old could understand and in 25 words or less what is a compiler? Anybody want to
venture a definition? Yes. >> It translates a language that a human can
understand to a language that a computer can understand.
>> AUSTRAAT: It translates a language a human can understand, or some humans, into a language
that computers can understand. That's great, that's a great definition. That's a very good
definition. And that makes sense, right? It's like a translation device. And my grandma
asked me this question few years ago and she loved baking. So I said it's like an oven
and you take the ingredients for a program. And you baked them and what comes out is still
the same ingredients but it has different functions now. So the compiler is like an
oven which made perfect sense to my grandma. So depending on your target audience coming
up with a good glossary definition is hard but it's a huge time saver. Because that--those
translators will not call you or will not waste hours, and hours and hours doing terminology
research when they could be translating. Yes? >> Given your baking example, are you going
to talk about idioms and metaphors? >> AUSTRAAT: Yes. So include all A's, I's
and A's, acronyms, abbreviations and initialisms, very, very important. Try looking up DB in
an acronym dictionary. Again like smith in the phone book. Lots and lots of different
meanings decibel, database and so on and so forth. Dun and Bradstreet, there are many,
many possible meanings for DB even in the same context. And at the very least a glossary
has to have four columns, a source term, a target term, you can leave that blank because
nobody's expecting you to actually find the translation for that term. But when you submit
a glossary for translation it should contain the source term, which part of speech is it,
very important. Remember the set example or the time example, is that noun, verb, adjective,
pronoun, non-phrase, none of the above and a comment field. And that--the comment field
is where you can spend four hours writing comments for a glossary and save yourself
per language 20 hours times 40 languages real big *** for the buck. This should be making--making
a glossary a robust glossary should be part of your translation process which I'm sure
it already is. But it's the responsibility of the source authoring team to really contribute
to that process. The translation team will create a glossary but you can really make
it a lot better and more useful by contributing to that process. There was one question in
the back there. >> So just to be clear this glossary in use,
is it going to be for translation team not for the for the target audience [INDISTINCT]
>> AUSTRAAT: Good question. So the question was, is this specifically for the translation
team or for the target audience? Both, because the target audience benefits tremendously
from a good glossary. >> So it seems to me that [INDISTINCT] if
your target audience is [INDISTINCT] >> AUSTRAAT: Correct. So that--again the question
was, does the target audience influence what goes into the glossary? Yes, it does. So the
most expansive possible version of the glossary will be for the translation team and for a
novice audience let's say. Somebody just bought Google Mini and wants to just plug it into
the network and they may need to look up what the, you know, feeds protocol is. And is that
it's own word and what is that mean? Even though they just bought the search appliance
and they are moderately technical just not in your way. You have a question? Okay. Now
to your earlier point, of course, translation is not just replacing words as you well know.
Because otherwise machines could do it and they still kind of suck at it. But we've been
away from perfect machine translation for 10 years now since the 1950's. So that's,
you know we're making progress we're still 10 years away. And there's a lot of cultural
content as well that--that influences and profuses language. For instance, if you simply
say the word Super Bowl you--most people in this room would know what that is, Right?
And what--what are all the things that pertain to it? What are some of the words that go
with Super Bowl? Football championship, snacks, barbecue, party...
>> Point spread. >> AUSTRAAT: Point spread. Very good. ***,
you know, et cetera, right? So the superbowl is a cultural institution and a sports game.
But if you say, I went to a superbowl party, that means a lot that you don't have to explain
because you all share the same culture. Of course, if you're in a culture that for some
bizarre reason doesn't have the game of football. And not in that form, they played with the
smaller ball and that it's round and, you know, this soccer thing that's going on right
now. Then adding something that's culturally relevant is going to really cause a huge translation
problem. The World Series is the same thing. Do you remember the subway series? This was
a few years ago where two teams from New York played each other in the World Series but
you could go between the two stadiums using the subway so not so much world in a technical
sense of the word. Constructions like great move, not. You're not using those. I'm just
saying don't. Potluck party, what does that entail? That's everything from suburban living
to a particularly informal way that American's interact with each other so potluck party
is simple straightforward word. If you have to explain to somebody who has never been
to the United States, you could spend easily a few pages explaining what that is. Hang
ten, et cetera, et cetera. So even straightforward words like school can be many different things.
School in Austria, for instance, is a lot more about academics. There's no school sports.
This caste system of jocks and cheerleaders, doesn't really exist like that. Clothing is
less relevant, after school activities are practically nonexistent while in the United
States, the high school is the focal point of social life for most--for most people growing
up. And also take into account that when you translate something, you also have to take
into account legal and technical requirements for other business cultures. For instance,
here's an example of what is--why is it that you can get the number of steps in the statue
of liberty in under a second but waste so much time--and so on and so forth. So, of
course, when I translate that, I would have to take something that is relevant to my target
audience that is maybe different from the example of statue of liberty because that's
a uniquely American example, not the end of the world. And that this is, you know, a professional
translator will know what to do because they're trained in doing this very specific task.
I'm just saying that this is another--one of those tiny speed bumps, that's going to
slow down the translation process a little bit. And here's an example of--these are--this
is a legal disclaimer that has to do with intellectual property and it's a good translation.
It's accurate. What I don't know without being a lawyer, a German or Austrian lawyer is does
this mean anything legally in that target audience. I don't know. Maybe this is reasons
for a lawsuit. Because I have now violated the statue of free information sharing from
1869, you know, I don't know. So again, translation is a lot more than replacing words even if
you're the perfect legal translator in this way--in this particular place, you would need
to culturilize it as well. Here are some of the worse offenders of translation speed bumps.
They're called modifier strings. For instance, I'll give $20 to the person who can tell me
unambiguously what the first bullet point means. I've got it right here $20. Anybody?
What does that mean? Yes, volunteer. >> Programs that work faster.
>> AUSTRAAT: Programmers that work faster. >> Programmers that work faster.
>> AUSTRAAT: Programmers that work faster or it could be faster programs?
>> No. >> AUSTRAAT: Is it developing faster or the
program's faster? >> [INDISTINCT]
>> AUSTRAAT: What's that? >> Accelerated applications don't make any
sense. >> AUSTRAAT: Accelerated applications. Well.
>> Yes, yes they do... >> AUSTRAAT: If the programs run faster, that'd
be a good thing, right? I'll have to keep those, sorry. Relational database index schema,
that's one of yours. Enterprise performance management software solutions, great. But
my favorite is from an actual government report and the government is the worst offender when
it comes to acronyms, initialisms and technical jargon followed closely by engineers. The
Commission was impressed by the--this is an actual none-made up example but a Test Project
Command Module Reaction Control System Engine Oxidizer vapor inhalation damage recovery
results. Translate. The best dictionary in the world plus Google will not do you any
good in this. This is not translatable as it is. Sometimes it's unclear, the modifier
strings are shorter but they are loosely correlated in the text such as, "You must send the art,
topography and color specifications to us." So are the specifications modifying the art,
topography and color or just the color. We have striven to integrate system development
and sales and support personnel on a global level. Again, non-translatable as it is. So
what do you with these... >> You write for the clear [INDISTINCT]
>> AUSTRAAT: Yes, and we'll get there in a second. Another problem with getting things
out of context is that sometimes words are spelled the same but they mean different things.
Because they are either homonyms or homographs. Meaning they're pronounced differently but
spelled the same, or they mean different things? For instance, set--we already saw an example
with the set. Searching the database is an example of something that out of context is
very difficult to understand what that means. Is it a message--a status message. I'm searching
this database right now, or is it a chapter that tells me how to search the database.
Searching the database. Here's how do you it in four easy steps. Record or record have
different meanings depending on how it's pronounced but, of course, on paper or in electronic
format, you don't know what that is. Telegraphic English, empty database. Does that mean I'm
just supposed to empty the database, the database is empty. And invisible plurals. In Spanish,
I would need to know how many servers are in the rack because otherwise, I can't translate
it as I would in--well German, I could just make one of those great compound pronouns
and let the user figure it out. Program update, how many programs are getting updated. Database
update in your case, how many databases are actually getting updated? Because the translation
would be different depending on--if it's a singular database or plural. Yes?
>> Can you explain why [INDISTINCT] >> AUSTRAAT: Right. So I'll actually get to
that on the next slide. It's basically, you've left out a keyword either the database is
empty or how to empty the database? And so now I cannot translate this out of context
because I won't know what the intent was. Right? Here's a great example from your Website.
Google Search appliance Feeds Protocol Developer's Guide. Specifying sculpting radius values.
Food for thought? So how do you solve these guys? And this is again, a great benefit to
you English users as well. For example, accelerated development of applications. To our first
example, the $20 example. Or if it's not, that--if that's not what's meant then obviously,
you can just reshuffle it. Google Search Appliance Developer's Guide for the Feeds Protocol,
is that right? Is that what is meant to be said there in the previous, in the previous
noun phrase. That's right. So look, we incurred almost no text swell. Just a little bit but
it's a lot clearer in English as well. Translation comments is another huge time saver especially
if you do have to export stuff into string tables or into some disembodied form, adding
a little comment marker like button label here and it means save, makes this word instantaneously
translatable. >> Well, what's wrong with searching the database?
>> AUSTRAAT: What's wrong--nothing. >> Then how is it possible?
>> AUSTRAAT: It resolves--By doing this, you resolve the puzzle of searching database.
The puzzle is searching database which could mean how to search the database or I am searching
the database right now. And by either using one of those two forms, it becomes immediately
clearer what this sentence means. Searching database, not so much. And this is also--it's
basically, this technique, expanding an elliptical English, we're going from empty database to
database is empty while you've sacrificed four characters two spaces and is, you've
added a lot of meaning. And glossarizing invisible plurals are very, very important. Server rack
contains one or more servers. Okay. No problem. Program update updates three programs that
maintain the database, for instance. Last but not least, graphical elements are innocuous
and I showed this to--this slideshow to Joanna before I presented it here and she said "What's
wrong with the mailboxes?" And absolutely nothing is wrong with the mailboxes except
those are breadboxes in France. Because mailboxes don't look anything like that in Europe, right?
And great, Gmail is not using mailboxes anyway, it's all the envelope symbol which is universal
which is a good--a great use of iconographic design that is culturally portable. This means,
a okay. Anybody from Brazil here? Means something slightly different in Brazil like really bad.
>> You can see it a lot in the World Cup >> AUSTRAAT: In the World Cup, yeah. Great
job striker for the other team. What's wrong with the Yin and Yang symbol? What does it
mean in this country? Here. Yoga studio, something like that. Meditation, you know. But it's
actually a pretty significant religious symbol to Buddhist people so, using that frivolously
might not be the most culturally portable thing to do. The mailboxes, we already talked
about, this means "stop" in American English. It means "high" in Iraqi Arabic. This means
"come here" in Thailand. This means "victory". This means something like this, in England.
And the Red Cross and the Red Crescent, they have a problem now, because now--the Israelites
have wanted to have a red star of David for a long time but nobody could agree to that.
So, they're actually going to change their symbol now to a red diamond which is going
to be the universal symbol for the Red Cross and Red Crescent. And then, you could put
your own cross or crescent inside, per country, but they had to come up with a culturally
portable symbol so people wouldn't be offended. If you write somebody's name in Korea in red
that means that they're either dead or you wish them dead. So, that might not be the
best thing to do. And green in Arabic-speaking countries of course, is a religious color.
And I've checked this, I went to Al-jazeera and I hope the homeland security department
is not no--hot on my trail now, but I went to several Arabic-speaking websites to see
it. And they also use green in their user interface. And I--since I don't speak Arabic,
I couldn't find out if that was always with a religious context, it's just something to
keep in mind, that it might be religiously connoted, have somebody check it to make sure
it's okay. So, for instance, here's some examples from the Google interface. Of course, Google
is famous for it's seasonally appropriate Google logos and there's no reason or whatsoever
to stop those. I'm not suggesting that. It's just a great example of a culturally specific
iconographic design. And I know that Google Australia has different symbols depending
on their holidays, and Germany and so on. Yes?
>> What about the color? I mean, do we have--is there any safe color [INDISTINCT]
>> AUSTRAAT: Are there any safe colors? No. No. There aren't any safe colors. Black, gray
and white, but white means death too, you know, white ladies and so, and so. There aren't
really any safe colors. This is an example from--this was I think sketched up where roads
or Google Maps, I forgot, where the road symbol looks like an American highway sign. Of course,
that wouldn't mean much to a user of a foreign culture. And very, very, importantly, concatenation.
Have you ever seen things like that? Or written than worse. Unable to do something with something,
unable to move the directory, generate the report, delete the file. Very efficient in
English, because you just plug-in words. Of course, it doesn't work at all in foreign
languages because the gender of the report might be masculine, so now, in other words,
we have to change to fit that. So, anytime you'll see concatenated sentences like that,
assume as a default position that it will not be translatable. And they will cause huge
rework and pain and suffering in the translation department and in the engineering department
because you will have to rewrite your code, and that's expensive.
>> [INDISTINCT] >> AUSTRAAT: Concatenation is, for instance,
I will have a table of words like, move, generate, delete, update, and I have another table of
words called record, file, report, directory, and then, I'll stick them into this template
as needed. And I'll say, take the sentence template unable to [INDISTINCT] something
[INDISTINCT] I'm able to move the directory, delete the file, update the record and--because
I have so many different possible permutations. With eight words, I can make eight factorial
permutations. It's a great time-saver in English and is an absolute translation-killer, because
it does not have to translate, period. Your report...
>> [INDISTINCT] >> AUSTRAAT: Right. You have to--you may have
to change the location. Other words may have to change, because now, it's a different article,
because suddenly the noun is masculine, not feminine, and so on and so forth. Another
great time-saver is your reports are being printed or your report is being printed if,
you know, you see this all the time, if the count is less than one, you know, don't put
an "s" there. Of course, in German, not so much. So, here's some recommendations for
concatenation. Avoid, it--definitely avoid it. Do not concatenate words into strings,
and evil companies use concatenation, which Google is not, right? There are ways around
it. For instance, you could make it slightly less slick in English by saying, number of
reports printed, X. That is translatable, no problem, because I have taken pieces from
within a sentence that are meaningful and I've isolated them into one meaningful unit
and a variable. So, that is translatable. If you ever want to really get somebody--some
translators' blood-boiling, ask them if they--if you should concatenate. You will get a good
response. So, here, in a conclusion, some interesting translation facts. It often takes
as long to translate something as it did to write the original. And to you as technical
writers, this may not be very shocking, but to some managers, it's flabbergasting. Why?
Don't you just sit there and like a parrot, you know, just type it in a foreign language?
No. As a technical writer, did you sit there and write it in English? Well, if you don't
type fast, huh, it takes you a whole day to do a single page. Slow typists, all of you?
No, because technical writing is different from just typing it, right? And translation
is different from just typing it into foreign language. Translation researchers typically
translate about 2500 words at most a day. You can do more if it's really--at the end
of a large translation project and you have already totally assimilated the glossary and
you're totally in a flow, but by and large, count on being about 2000, 2500 words. Optimized
source text reduces the number of occurrences tremendously, of these translation speed bumps
that take up to 20 minutes to fix. And nobody in the world, including yourself, will ever
read the text as carefully as a translator, because they have to understand every single
sentence. No user, no boss, no reviewer, no editor, will ever read your text as carefully
as a translator. So, they can be your allies and by making their life better and easier,
they will be able to be more productive which obviously helps Google as a company. Translation
is a complex and creative process and it's well beyond just looking up words, in which
we explored a little bit of that today. And the number of qualified translators is actually
quite tiny. How many translators do you think there are in the United States? Professional,
you know, certified translators? >> Five thousand.
>> AUSTRAAT: Five thousand? More? Five thousand would be a lot less than [INDISTINCT] surgeons.
There are actually a lot less than [INDISTINCT] surgeons. Yeah. The American Translator Association
I think has about 13,000 members. That's not a whole lot. And there are other translators
that are not members but it's a really small number. Typically, translation is a training
that's between two and four years plus experience and people work constantly to keep their languages
fresh. It's a highly--I taught translation at the Monterey Institute, I can tell you,
there's a reason why there is a two-year graduate program to become a translator. There's a
lot to learn in addition to having two languages. Okay? And that's pretty much it. Any questions?
No? No questions? Yes? >> [INDISTINCT]
>> AUSTRAAT: So, the question was... >> [INDISTINCT]
>> AUSTRAAT: ...do you avoid idioms and metaphors? Some. And I would avoid idioms and metaphors
if they have no apparent translation such as, something that is so uniquely cultural,
like, super bowl, or pot luck, or World Series that is going to be--it's going to take a
translator's footnote this big to explain. If it's an--a standard phrase like, early
birds catches the worm, there is a great German and Spanish and Japanese translation for that,
it's going to say something entirely different but the meaning is going to be--it's going
to be possible to carry the meaning across. So, no. You don't have to make your source
text so bland as to avoid all references to something cultural. You may want to glossarize
it and you may want to draw the translator's attention to it, what you meant to say with
this metaphor. What is it that you really want to communicate by saying super bowl or
the bottom of the ninth and the bases are loaded, which if you're not from a country
that plays baseball, doesn't mean much, right? So, it's not--it's not an--I'm not advocating
making the text blander, but just keep it in mind that this is going to be something
that's going to bubble up in the translation process that needs to be handled and managed,
but it's okay because these are professionals.