homepage Welcome to WebmasterWorld Guest from 54.161.192.130
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Pubcon Platinum Sponsor 2014
Home / Forums Index / Code, Content, and Presentation / HTML
Forum Library, Charter, Moderators: incrediBILL

HTML Forum

    
Creating HTML mark-up for languages you don't know
vkaryl

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 1:32 am on Oct 30, 2004 (gmt 0)

tedster, are you fluent in Polish?

If not, how do you go about creating a website in Polish?

[edited by: tedster at 3:31 am (utc) on Oct. 30, 2004]
[edit reason] split from this thread [webmasterworld.com] [/edit]

 

tedster

WebmasterWorld Senior Member tedster us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 8969 posted 3:50 am on Oct 30, 2004 (gmt 0)

Well, it's a heck of a process, let me tell you. One of my clients is taking an English language site and creating 14 different translations (for starters). No single person is going to have this fluency -- the translations range from Bengali and Chinese, through Hebrew and Polish, to the key Western European langauges.

And so we've put together a translation team - two people per langauge. One person is the translation professional, and the other is the proofreader. And when the first 14 languages are done, many more are projected.

So we first created the English language website. Next the translators create Word documents which we converted to pdf downloads for the first 14 languages - we have to ensure that our team has the correct fonts and embeds them in the pdf. We also create separate directories on the server for each language that have the correct HTTP headers, and make sure the pdf documents have correct metadata.

The pdf's go back to the proofreaders, especially for line break issues and such, although they also have an eye out for wrong shifts in meaning. The proofreader and the translator hash all that out.

Now we go through the website and create a list of words that we need translated just for the HTML version - file names, graphic text, text links, etc. The translators handle that.

Next we take the English HTML and change the pages' charset meta tags and lang attributes. After that, we create the new graphic images and then paste in the translated text.

Final step - back to the proofreaders, and possibly the the translators, for last minute tweaks. Then place the translated HTML online.

It's certainly a project the likes of which I've never tried before. If we stumble over anything more, I'll be sure to share. So far, we've been saved a lot of trouble by the technical knowledge of our translators.

Each language has its own unique demands - such as hyphenation conventions in those long, compound German words. And there are apparently conventions about which pictographs should and should not be near each other in polite Chinese.

[edited by: tedster at 7:25 pm (utc) on Oct. 30, 2004]

tedster

WebmasterWorld Senior Member tedster us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 8969 posted 4:27 am on Oct 30, 2004 (gmt 0)

I should add one other important step we take - we look for sites in each target language run by major newspapers and universities in that language. Then we review their HTML and HTTP headers to see what common practice is.

Google's "language tools" have been helpful in finding authoritative websites in each language.

Mr Bo Jangles

10+ Year Member



 
Msg#: 8969 posted 8:31 am on Oct 30, 2004 (gmt 0)

Having just gone through this with our site for a couple of languages (a few more to go) I can also attest to the fact that there is a bit in it, and cost. We advertised on a translater's freelance site for those who wanted the work and got *many, many, MANY* responses - a LOT of prices - then we just picked one of the cheapest.

Then of course there's the pain of doing updates to the translated pages when you update the main language page - inevitably I think the secondary language pages will 'lag' a bit *_*

vkaryl

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 3:37 pm on Oct 30, 2004 (gmt 0)

Wow. tedster, that's amazing! Talk about having to work out your deck of cards FIRST!

I'm considering a job for a financial wiz (basic page with links to pdfs to make his 300 different forms available for download and printing), and while I have a clear idea what to bid for the English setup, he also needs it in Spanish - which is the "home" language of about 30% of his market. He's providing the translation, btw, and the only graphics I'll have to worry about will be his logo, which he also will provide in the Spanish version.

I was needing some background on how much time this could take and how complicated it could get, so I know whether to just bid by the hour for updates, or bid a maintenance setup on a yearly figure (I tend to not like to do that anyway - too many possibles for "off the wall/off the charts" stuff to bite you you-know-where). Would you say that using includes for portions of pages which would need regular updating would be a good fit for a single "other language" setup?

It's absolutely amazing to me that just as I need some information, WebmasterWorld pops up with the goods! Thanks, ted. You're a wonder!

encyclo

WebmasterWorld Senior Member encyclo us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 8969 posted 3:51 pm on Oct 30, 2004 (gmt 0)

vkaryl, Spanish is easy - because it is a western European language and as far as I know, you can use the same charset (ISO-8859-1). That, along with the translation difficulties, is the worst hurdle. You could also use UTF-8, which covers just about anything - it may well be the best choice for a new site.

Is the site going to be completely bilingual with identical pages available for both languages? The best way I've found is to make a best-guess detection of the user's preferred language on the home page, and leave the option to switch languages at any time. I find it so much better than a splash page just to choose the language. If you can store the value in a cookie for later, that's a good idea too.

It is important also that the "switch language" link on any one page leads to the equivalent page in the other language, not just back to the home page.

Bilingual sites are not too bad to manage - especially if you speak both languages - but multi-lingual sites can quickly become a nightmare!

vkaryl

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 3:59 pm on Oct 30, 2004 (gmt 0)

Hi encyclo - nah, my Spanish is of the "two years in school" variety - THIRTY YEARS ago.... I can get a word or so, but sentence construction? No way....

Yes, I had figured that the character set wasn't a problem - that was the first and maybe only easy thing. Haven't got into the "this page in English, change to Spanish brings same page in Spanish" setup with him yet, though that's how I think it should work.

He does want a page just to choose language first. That's up to him, and we've already had the "search engine" conversation - he doesn't seem much concerned with that, since his business is so huge already with no website at all, therefore no seo/sem etc. I want to see a link on each page which allows change to the other language - sort of a "fail-safe" option....

I'm thinking this will be fun, if I can get the right approach ironed out ahead of time....

mifi601

10+ Year Member



 
Msg#: 8969 posted 5:15 pm on Oct 30, 2004 (gmt 0)

I have a couple of bilingual sites and have found, that people DO NOT use the 'this page in name your language'-link. Usually they have found the one in their language to begin with if they come from the side. A link to the main page of the other language is really all they use.

This is useful insight, since I had the link going to the equivalent page automatically, and sometimes your other language editor is not available to update/create that site.

If you are planning to do a CMS this become an issue.

Now each language editor can create pages with their own filenames (in their language!) and inform the other language guy of the new content. Whenever that person gets to it, the corespondiong page gets created/updated.

I have found this to be a much more fluid and also SEO friendly way.

vkaryl

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 5:39 pm on Oct 30, 2004 (gmt 0)

mifi601 - thanks for the useful insight. That may help to convince the client that a splash/pick-a-lang page isn't necessary. He's not "unsavvy" when it comes to the net, he just hasn't ever done a website - doesn't even have a "family" sort of page, so hasn't any idea....

bill

WebmasterWorld Administrator bill us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 8969 posted 3:06 am on Oct 31, 2004 (gmt 0)

I really want to re-emphasize some very important steps that tedster outlined here. A lot of people think that translation is just one step in the process. They completely miss the proofreading/copy-writing stage. Translation is a step where someone will change your content from language A to language B. However, a lot of translation doesn't exactly read well in the target language.

I've seen quite a few documents which have been translated from Japanese & Chinese that would still not pass for a well written document. In many cases they are not very understandable (and I'm not talking about machine translations either). You need to make sure that you are taking the extra step to ensure that the document is well written in the target language as well. Unless you are using a top rate translator often you need to take steps to have this extra work done.

You not only need to make sure that the finished work reads well in the target language, but you also need to make sure your message meets the market. For example, if you're targeting a Japanese audience you may not know that direct references to your competition are frowned upon. You'll never see an ad on Japanese TV where they espouse the virtues of brand A over brand B. However, if you go to the US it's hard to not see a commercial where they don't do this. These may seem like small market differences but it could change your content entirely. You need to keep these things in mind, and having your translated materials proofread and rewritten can help tremendously.

Snappy_fish

10+ Year Member



 
Msg#: 8969 posted 9:39 am on Oct 31, 2004 (gmt 0)

I run a translation company dealing with the English and German languages and I can confirm that there is a great deal of work involved in getting it right. Unfortunately, especially in the English-speaking world, there is a great deal of ignorance about language, translation and, ultimately, the culture of the target audience.

The usual objective of having a website translated is to make money by selling the product or service to consumers in another country. Indeed, it's never been easier to tap into new markets. But getting the translation right is just as important as getting the programming right. Translators are not computers! I know of companies that will spend thousands getting their copy just right, hold countless meetings discussing how the message is to come across and pay a great deal of attention, as they should, in projecting their business in a professional way. Yet for all this effort, when it comes to translating the site into another language, its often the cheapest translation provider that gets the work provided, of course, they can finish 20,000 words over the weekend! The foreign reader is presented with poor and unconvincing copy at best, gobbledygook at worst.

On the other hand, good translation makes it possible for you to convey your thoughts, ideas, hopes, dreams, ambitions, experience and anything you want to people who share a different language and a different culture. And when the translation is good - and some translations are very good readers will understand and take in the message as if it had been written in their own language by someone who shared their own culture! A good translation can make the difference between winning or losing a contract, between encouraging and preventing a sale and ultimately between being taken seriously or not.

Good translation is the ultimate marketing tool its just that not many people know it!

mincklerstraat

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 9:52 am on Oct 31, 2004 (gmt 0)

The value of good translation is indeed not to be underestimated. If you don't fall in the category of a site that can spend lots of cash and time on translation, however, it's essential that your translations have links to the page in the other languages available (esp. English). This is also a great feature even if your translations are really good, since people do surf in on foreign-language pages trying a search in that language when they might actually prefer another language. Having to go back to the home page and find the same page over again is a huge, huge pain. Unfortunately very few sites are built to this specification, and it does make the programming more complicated.

MovingOnUp

10+ Year Member



 
Msg#: 8969 posted 12:36 pm on Oct 31, 2004 (gmt 0)

I think we've all read poor translations where the translator obviously doesn't understand our language. While they might make good humor, they certainly don't work to promote a product.

I think you might be missing one of the major factors here. You can build a page that looks perfect, but how are you going to respond to email inquiries? I'm assuming that you won't have anyone on staff who is familiar with all these new languages.

Mr Bo Jangles

10+ Year Member



 
Msg#: 8969 posted 12:40 pm on Oct 31, 2004 (gmt 0)


I think you might be missing one of the major factors here. You can build a page that looks perfect, but how are you going to respond to email inquiries? I'm assuming that you won't have anyone on staff who is familiar with all these new languages.

I have found that you can machine translate e-mails using, for example Systran, and then machine translate a reply to the e-mail - *and* include a prefix line apologising for 'machine translation' - I have had excelent communications using this methodolgy.

MovingOnUp

10+ Year Member



 
Msg#: 8969 posted 1:36 pm on Oct 31, 2004 (gmt 0)

I just used Babelfish to translate your message from English to German and back to English again to demonstrate how broken computerized translators can be:

I found the fact that you can to work on translate enamels also, e.g. Systran and works on then translates an answer to the email - * and * you include a prefix line, for 'machine translation' apologizes; - I have excelent communications with methodolgy this had.

mifi601

10+ Year Member



 
Msg#: 8969 posted 2:06 pm on Oct 31, 2004 (gmt 0)

LOL

I found that with the Spanish market it was especially difficult.

You really have to know the target COUNTRY. Just any Spanish speaker, as good as s/he might be won't do!

Some very embarrassing, albeit funny, mistakes can happen during translation.

That's why I try to have an 'editor/writer' for content in the first place and do not place so much emphasis on 'translation' - you cannot translate culture.

lexipixel

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 4:54 pm on Oct 31, 2004 (gmt 0)


I found that with the Spanish market it was especially difficult.

You really have to know the target COUNTRY. Just any Spanish speaker, as good as s/he might be won't do!

Some very embarrassing, albeit funny, mistakes can happen during translation.

... reminds me of the story of how Chevrolet tried to market the "Nova" in Mexico.

Try translating "no va" from Spanish to English and you get "it does not go".

:)

tedster

WebmasterWorld Senior Member tedster us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 8969 posted 8:15 pm on Oct 31, 2004 (gmt 0)

The foreign reader is presented with poor and unconvincing copy at best, gobbledygook at worst.

Welcome to the forums, Snappy_fish. I know exactly what you are saying, here. Several of our translators (but unfortunately, not all) are professional copy writers in their language, and I rest much easier when that is the case.

how are you going to respond to email inquiries?

The email issue is a good one - my client is a global organization with native speakers in the languages where we are creating translations. They are not translators, but they will handle the emails just fine.

As English speakers, we've all seen instructions written for our electronics, clothing, food, and so on that were very poorly translated - and I hope the approach my team is using will avoid that kind of problem.

Some challenges have already come up where the the translation needs to be a bit "loose" and the content of two sentences must be intermingled in the target language. Our page builders sometimes inappropriately introduce a new paragraph break between those two sentences to keep the text blocks small on screen, and the flow of meaning is disrupted.

But that's the beauty of using a proofreader - we can catch those problem before final release. These sites are heavy on information, basically lead generation sites, not product sites. So the text must come across very well.

superpower

10+ Year Member



 
Msg#: 8969 posted 11:24 pm on Oct 31, 2004 (gmt 0)

A few years ago I was in charge of the i18n program for a large website. By the end I had an i18n process checklist that was about 50 pages long (not including the translator hiring process which was another big project). I'm not as up on it now (boy did I get out of i18n, what a pain) but here are some random thoughts from the perspective of high volume translation:

Depending on the complexity of your website there can there can be a lot of technical issues that may require additional engineering staff--like character encoding being supported by web applications you are using,double-byte issues, database, url browser compatibilities etc. Also there are issues such as how it's displayed (geoip-based?) and how new information gets translated and published (process, time-lag etc.), architectural issues. A large site usually has QA. Can QA handle foreign translations--do they know that there may be technical differences or process time lag when compared to English pages? Are there tech or other emails that are sent out that will now need translation? How long will that take each time one is needed? Is everybody in the loop and on the same page about this?

Do not just hire the cheapest translator. That's a bad idea. The amount of hassles created by our using cheap translators (round 1) cannot even be imagined. Translators should be well-trained with some sort of documentation (certification) or credentials--which is industry-standard for a professional firm. Just getting anybody who knows English and Spanish to translate your site into Spanish is not necessarily going to work. Translation is an actual skill. The translator has to know both languages fluently, has to make decisions based on the context of the words sometimes rewriting things, has to know the idioms and slang of both languages and other specialized internet words. I hired a translation company not the individuals, but I still insisted on interviewing the translators and vetting them to make sure they were up on music style (our website topic) and other specialized info.

Since we were a large website with a ton of data it required a team of translators on their end, which in turn required an internal process on their end and a project manager on their end. Since each translator has a different translation style a set of rules, a style guide, has to be put in place and approved. Then you need another style editor person to make sure the style was maintained (between translators), then finally a proofreader. Also there needs to be a feedback mechanism in place so errors can be quickly sent back to them and corrected. And make sure versioning is under control--are they working on the correct version of the document--have things been changed on the English site, perhaps an error corrected, whcih will not show up corectly inthe translated doc because they were working on an older version.

Then there is translation tone. We were targeting a young 16-35 demographic for music so the tone of the translation had to be more informal. Of course in many languages this is much more of an issue than in English. Round one we didn't think of it as much and the translation read like it was for a 60 year Spanish professor. Then as somebody else stated there are different styles of Spanish and slang in each country.

See how this keeps getting bigger with more variables and potential for complication? Granted on a small site many of these points are not necessary but you get the idea of what to look out for...

btw, check out the w3.org for more info and background from a technical perspective:
[w3.org...]

tedster

WebmasterWorld Senior Member tedster us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 8969 posted 1:16 am on Nov 1, 2004 (gmt 0)

An excellent post, superpower - thank you.

...and a project manager on their end

I'd like to underscore this point. The value of a strong person as the translation team project manager cannot be overstated.

Our translations manager is multi-lingual and has been a project manager for many years on a healthy variety of projects. He has saved us untold trouble and sometimes has been able to get his team to turn around their work well ahead of schedule.

Gusgsm

10+ Year Member



 
Msg#: 8969 posted 3:07 pm on Nov 1, 2004 (gmt 0)

As a Spanish speaker myself, may I advice you to make the translations into something that can be called "neutral Spanish"? It would be the equivalent of standard English. There are many many things in any variety of Spanish that may annoy the speaker of another variety. So a common denominator Spanish is the preferred language between "educated" Spanish speakers.

That, of course, does mot apply if your target is, let's say, Mexican teenagers and a "cool" music selling site. Then the language used is part of the game, as your audience wouldn't like a bit being spoken in formal neutral Spanish.

chadmg

10+ Year Member



 
Msg#: 8969 posted 4:58 pm on Nov 1, 2004 (gmt 0)

I think the translation part of this task has been covered very well. I'll throw in my 2c about the technical aspect. I used to work for an online English teaching school, so the whole world was the target audience. I think we concentrated on translations in 8 languages. Every piece of text was placed into a database. When images needed to contain text, we made 8 different versions, and put the URI for each image into the database.

So we created a web application called the Language Layer. Each section of text, or blurb, was placed into the database, and would appear in a queue for each team of translators. The translators were suppose to view the page the text was on to get the context of the blurb, and translate. If a blurb did not have a translation for the language a user has set, it would default to English. So you could be viewing the site in Japanese and have a paragraph or a button in English. If the blurb changed in English, the translations would get flagged, and the blurb would reappear in their queues. The English version would then display on the page until it was retranslated. As a web developer it was great, for a paragraph, I just wrote something like <p><?=getText(blurbID)?></p>.

I no longer work there, but this is what I've learned from that system. It's a good idea to develop a system where you only need to program for one site. You do not want to start teaching html to translators, and you don't want to have to make programming changes on 8 systems. Although, in creating your database application, you don't want to lose focus on how people will use it. We had a problem where we would create a blurb for a menu item such as "Community" and then we would use that blurb any place we needed it. However, that word may change in the translation depending on the context, and we didn't always account for that. Also, since the blurbs were constantly being tweaked for maximum effectiveness (damn salespeople) and we had tight deadlines, translators would not always look at the context of which the blurb was being displayed and they would just translate the paragraph on its own. I would recommend hiring a team of salespeople who are native to the country you are targeting instead of just straight translators. Then they can write text that fits the page and not just a direct translation from English.

If I were going to redesign their database system I would put more emphasis on the context than separating text into paragraphs. If a page had multiple paragraphs, I would put all of the paragraphs into one field. The problem comes with things like lists, tables, and links where html needs to be written but you don't want to allow translators to write html. Actually I may have redesigned the system to parse the html on the page and build a form that sort of resembled the look of the page but contained textareas and text input fields. Then the translators would translate a page at a time, instead of a blurb at a time.

aleksl



 
Msg#: 8969 posted 12:05 am on Nov 2, 2004 (gmt 0)

Since I run a decent-size bi-lingual portal...here's my 2 cents worth.

I would estimate, that to create completely workable, debugged and well-translated version of your site with support in second language it will take between 50% and 100% of effort you spent on building the original site.

If I were a large corp, I would approach a large or well-known web development agency in the native country to build my site. Believe me, this would be money well spent and an effort not wasted.

vkaryl

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 12:08 am on Nov 2, 2004 (gmt 0)

Priceless info you all have provided here, and I'm VERY grateful! I'll have a lot better position when interviewing the client now.... especially as this will be done over the phone for the most part - he's in SLC, I'm 280 miles south of there and the weather sucks rocks through a straw! He was planning to come down first of next week, but now has some fires of his own to put out....

Not sure how the Spanish part of this will play out. Once made aware of the potential for blowing up in his face, he may opt for English-only. Or not. He's done very very well without a website, AND with only basic Spanish-speakers in his office, so we'll see. (The Spanish speakers btw are former LDS missionaries. And his Spanish-language clientele are pretty much within the Salt Lake Basin, with a few as far south as Orem, and a few as far east as Evanston WY.)

I'm as yet unsure whether he has plans to use a website as a vaulting horse to further expansion, or if it's simply an easier way to provide forms and documents for his current client-base....

Again, my thanks!

rencke

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 10:34 am on Nov 2, 2004 (gmt 0)

A method for smaller budgets, less expensive and less time consuming than the one described by Tedster above (but potentially producing lower quality):

If you check the site in my profile, you will find a row with 19 flags across the top of each page. We coded the English version ourselves, placed identical copies on the server for each of the 19 domains involved and gave the translators access to the server. They could then download the files and go to work on the text only. All translators were nationals of their respective countries and living in them (important, since expatriates tend to lose their language after a few years). So we got idiomatically correct translations in contemporary language.

If we had needed compelling advertising prose, then we could not have taken this inexpensive route and would have had to use local professional copywriters.

Remember: All internal links must be relative, so that the translators won't have to change coding. They must not use editors that write their own code, such as Front Page a.o.

bobince

10+ Year Member



 
Msg#: 8969 posted 2:38 am on Nov 3, 2004 (gmt 0)

In an ideal world, you've got a native speaker on-site, who already understands the basics of how the site works, and can edit their own branch of the site's textual templates directly, so they can see and test their own changes straight away. (CVS or similar is good for this.)

Unfortunately most of the time such a person is not available, leaving you stuck with the usual techniques of sending text files to external non-technical translators. For anything but trivial static sites, this doesn't work well. Without the context of the running site in front of them they're not going to get it right, resulting in long cycles of correction and testing. (Or, omitting testing stage, laughably awful site copy.)

I wish I had a solution, but I really don't think there is one; to do it right you need staff who understand both your site and the target language.

vkaryl

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 8969 posted 3:09 am on Nov 3, 2004 (gmt 0)

Yes, that was the thrust of our conversation today; the result is that the client is putting back the development of the Spanish version of the site until 2nd quarter 2005 at the earliest, and in the intervening space of time will be looking for those quality translators/testers/proofers that will be necessary to produce a quality Spanish-language site. He also seems now to be leaning toward a Cuban-Spanish site, and a separate site for other Spanish speakers (don't know about that one, though a large proportion of his Spanish-speaking clients ARE Cuban....)

In any case, I'm quite happy that this has sorted out this way. I'd like to get his site up and running the way we BOTH want to see it, without having a problem of the other nature waiting in the wings. Hopefully by mid-2005, he'll have done his research and will have some viable options for us to work with.

I can't begin to tell you all how valuable your input has been. No matter how long I've been designing websites, this particular thing was FAR out of my realm to discuss with any intelligence. Thank you all so much - you helped me come across as professional and logical. Bless you!

Gusgsm

10+ Year Member



 
Msg#: 8969 posted 9:57 am on Nov 3, 2004 (gmt 0)

<i>He also seems now to be leaning toward a Cuban-Spanish site, and a separate site for other Spanish speakers (don't know about that one, though a large proportion of his Spanish-speaking clients ARE Cuban....) </i>

That would be like making a version for Canadian English version and another for standard English. A bit useless and hard to understand.

But that is just my personal and untransferible opinion ;) (posted from Madrid).

Gusgsm

10+ Year Member



 
Msg#: 8969 posted 9:58 am on Nov 3, 2004 (gmt 0)

The absurdity of my last post only serves to reinforce the need of proofreading in any language, of course ;)

(Aaarrghhh!)

Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Code, Content, and Presentation / HTML
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved