Mobile menu

Pages in topic:   [1 2] >
Bare Website Translation Tool - A Better Google Translate
Thread poster: netEditr
netEditr
English
Jan 11, 2009

Here's a very barebones solution I've created to do my wesbite translations, the engine is Google Translate but I've extended functionality to be not as restrictive and there is no duplicated content. It works fine for plain text or HTML content so can be used for all sorts of text translations.

My solution may not be for everyone since it is very barebones but it works for those who are willing to learn how to use it.

Current Problem with Google Website Translations: If you translate from English > Spanish and you Ctrl-A, Ctrl-C, and Ctrl-V. You will see BOTH English and Spanish content when you paste into your target. Character limit only around ~1000 characters for most sites using google translator so most of your HTML content will be cutoff.

Solution: Filter out the English part, only keep Spanish and retain all HTML formatting. Character limit: Unlimited? (haven't test the real limit) but if you have a long web page, it should be able to translate everything fine. On rare instances, it may choke if your tables/divs are nested complicately and you'll have to separate the long webpage into smaller regions and translate each region one by one.

Here is the link:

http://neteditr.com/translate.html

Instructions:

1) Paste your PLAIN TEXT/HTML SOURCE (content between body tags) content in the LEFT TEXTBOX (Autodetects your language)
2) Select the target language above (Default is Arabic so change to the target language you want to translate to)
3) Content is automatically translated and you can see a preview below
4) If it looks good, click on the [ transfer ] button
5) All translated PLAIN TEXT/HTML SOURCE is transferred to the RIGHT TEXTBOX
6) Click inside RIGHT TEXTBOX, Ctrl-A, Ctrl-C and Ctrl-V in your target.

For extra human translation editing...

7) Go to main site http://neteditr.com
8) switch to HTML source mode
9) Ctrl-V in there
10) Turn off HTML source mode and preview see if it's ok
11) If you're unhappy with Google's translation, you can edit until you are satisfied
12) Once the translation is OK, switch to HTML source mode, copy source and then paste in your target.

If you need to translate many articles, refresh the page (F5 / Ctrl-R ) everytime you finish translating an article. As I said it is very barebones

Final note: You have to go directly to http://neteditr.com/translate.html as it is not really meant for public use, it's a quick hack I put together to be able to translate all my web articles quickly.

If you have any problems or suggestions to make this tool easier to use, click on the [ about ] link on the main page http://neteditr.com and leave a message there or leave a message here. When I have time I will work on it a little more.

Chris

[Edited at 2009-01-11 04:17 GMT]


Direct link Reply with quote
 
netEditr
English
TOPIC STARTER
Extra info Jan 11, 2009

When translating website content, please only translate article content only. Do not include any javascript code or linking to other websites or else this tool will may not work correctly. Only paste pure HTML formatted article textual content that you expect people will read.

Basically you will only select all HTML content within body / div tags and remove any javascript coding before pasting into the LEFT TEXTBOX

Any feedback and suggestions would be appreciated.

Chris

[Edited at 2009-01-11 04:37 GMT]


Direct link Reply with quote
 
netEditr
English
TOPIC STARTER
Small tip Jan 11, 2009

A small tip using Firefox and Firebug

Those who use Firefox/firebug know that you can hover the mouse over elements in a webpage and be able to immediately access and edit the attributes of the element.

Now lets say you find that this discussion that you are reading is something you want to paste into your blog in another language. You want to retain the exact HTML formatting so you don't want to just copy/paste from the webpage and you are too lazy to view in source mode and find the HTML code.

1) Enable Firebug Inspect mode
2) Hover the mouse over the SPAN/DIV of the discussion you want to copy
3) Click on the area you are hovering
4) The firebug console highlights the SPAN/DIV layer
5) Right-Click the highlighted SPAN/DIV layer
6) Copy InnerHTML

And there you have it, the exact HTML source is in your clipboard and you can paste it in for translation. This demonstrates how easy it is for you to apply this tip to translate your own webpages.

If there are any problems during translation process, remove any complications of javascript and deeply nested layers before translating.

[Edited at 2009-01-11 05:36 GMT]


Direct link Reply with quote
 

Nicole Schnell  Identity Verified
United States
Local time: 13:55
English to German
+ ...
Dear software developer netEditr.com Jan 11, 2009

I am pleased to see that you discovered a new marketing channel by advertising your product in our forums. That's much cheaper than paying ad space like anybody else.

However, I am still struggling with the concept of how this is possibly related to professional translation.

Unfortunately I had some spare minutes and I tried out your software by inserting a paragraph of the text I am working on right now. I was impressed - in terms of hilariousity, that is, as it made me roll on the floor with mirth.

Gibberish machine translation for translators? Please explain how come that you consider this appropriate.

TIA!

Nicole

PS. Moderators may kill me


Direct link Reply with quote
 
netEditr
English
TOPIC STARTER
Glad to amuse you :) Jan 11, 2009

Just being generous to share a tool I use myself.

Sorry if it doesn't work for you but when doing translations, it has helped me translate things faster if you do not have a strong command of the target language. At least it provides the meat and bones in a few seconds and you can post edit it to make it more sensible afterwards since you have most of the keywords translated already. All you have to do is restructure and make it flow better.

The main feature that is not heavily emphasized is when you need to maintain complicated HTML formatting of the document. That is the main reason for this tool that I cannot find anywhere else online. For one phrasers, it's not much help. But if you have a heavily CSS designed HTML document, it comes in handy. If you deal with a lot of HTML documents and need to translate it, you'll understand what I mean. But again, maybe most of you just translate and give the translated text to the HTML designer to paste it back in... anyhow, I do all of the translating and HTML building and I know what works efficiently within my workflow.

Given that it is built upon Google Translate, it will only get better as the aggregrate sum of all global translators keep building up the quality of the Google Translate engine.

Feel free to translate with it in the future whenever you need a laugh Maybe one day it'll surprise us all by returning correct translations... as for now, it's good for retaining HTML formatting and laying out some foundation to work upon.

Chris


Direct link Reply with quote
 

Lawyer-Linguist  Identity Verified
Portugal
Local time: 21:55
Dutch to English
+ ...
Indeed it is Jan 11, 2009

netEditr wrote:

As I said it is very barebones



But I've seen worse.



[Edited at 2009-01-11 16:57 GMT]


Direct link Reply with quote
 

gianfranco  Identity Verified
Brazil
Local time: 17:55
Member (2001)
English to Italian
+ ...
Some objections about this approach Jan 11, 2009

netEditr wrote:
Sorry if it doesn't work for you but when doing translations, it has helped me translate things faster if you do not have a strong command of the target language.

Sorry, but we translate only into languages for which we have a strong command, indeed a very strong command. Preferably we translate almost exclusively into our native language and some translators, occasionally and in very particular circumstances, may translate into a non native language of which they have anyway a near-native command.
A tool to help translating faster into languages for which we do not have a strong command is a completely wrong approach.

netEditr wrote:
At least it provides the meat and bones in a few seconds and you can post edit it to make it more sensible afterwards since you have most of the keywords translated already. All you have to do is restructure and make it flow better.

I disagree. A machine translation does not provide much of any practical use, your so called meat and bones.

The first draft, generated automatically, is as far from a good translation as you can get, and the devil is in the detail. The details indeed is all that matters, with the first automatic draft being not much better than a space filler.

netEditr wrote:
The main feature that is not heavily emphasized is when you need to maintain complicated HTML formatting of the document. That is the main reason for this tool that I cannot find anywhere else online.

For maintaining the HTML, XML, or any other tagged structure, we do have plenty of efficient and reliable tools, but they are not barebone. In fact they are quite reliable and well tested.
I don't see any need of additional automatic translation features on top of a good editor for tagged files.

netEditr wrote:
Given that it is built upon Google Translate, it will only get better as the aggregrate sum of all global translators keep building up the quality of the Google Translate engine.

I strongly doubt about this statement.
Good terminology, respect of style guides, stylistic consistency, adherence to proprietary terminology or client preferences will never come out of Google Translate, and that is what we do every day.

netEditr wrote:
Feel free to translate with it in the future whenever you need a laugh

Thank you for sharing this tool then...


One question: how much do you know about professional translation?
Your approach seems to indicate "a barebone nothing"

bye
Gianfranco



[Edited at 2009-01-11 19:02 GMT]


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 22:55
Member (2006)
English to Afrikaans
+ ...
How about uploading instead? Jan 11, 2009

netEditr wrote:
Here is the link:
http://neteditr.com/translate.html


1. I'm not sure if Google's terms of use would allow you to use their service in a way that makes it appear as if you're using your own service. Especially now that the URL is public, if someone complains, you may find your domain banned.

2. If the object is to simplify HTML translation, how about uploading files instead? The workflow I have in mind, is this:

* The page has a form with which the user selects and uploads the file.
* The web site has a CGI script in the backend that grabs the uploaded file and puts it somewhere where the URL is known.
* The script then sends that URL to Google Translate.

So my idea is that you upload a file to a site, and then give the URL of that page to Google Translate. To simplify things, use a web form with upload function and a CGI script that saves the uploaded file in a specific location.

In this way, an entire page is translated and the user can save it where he wants to.

[Edited at 2009-01-11 17:15 GMT]


Direct link Reply with quote
 
netEditr
English
TOPIC STARTER
Thanks for your warm concerns Jan 11, 2009

Samuel Murray wrote:

1. I'm not sure if Google's terms of use would allow you to use their service in a way that makes it appear as if you're using your own service. Especially now that the URL is public, if someone complains, you may find your domain banned.



Thanks for your warm concerns but it is Google's intention to allow developers to use their tools and extend upon it. Do you see domains being banned because people are putting Google Maps and adding features to the map on their websites?

Thank you for your feedback and I'll get back to you about your suggestions later.


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 22:55
Member (2006)
English to Afrikaans
+ ...
Here, give it a try Jan 11, 2009

Samuel Murray wrote:
2. If the object is to simplify HTML translation, how about uploading files instead?


I took the tutorial at:
http://www.sitepoint.com/article/uploading-files-cgi-perl/
and adapted it so that the uploaded file is sent to Google Translate. I gave the landing page an meta-refresh redirect.

Try it out:
http://www.vertaal.org/file_upload.html

In this temporary demo, I've limited the size to 50 KB and it only translates EN-ES. Obviously you have to click to remove the frame before you save or view the source.


[Edited at 2009-01-11 19:51 GMT]


Direct link Reply with quote
 
xxxMarc P  Identity Verified
Local time: 22:55
German to English
+ ...
Nice idea, but... Jan 11, 2009

netEditr wrote:

At least it provides the meat and bones in a few seconds and you can post edit it to make it more sensible afterwards...


This seems to be a common misconception among software developers: that even if machine translation is substandard, it at least means that some of the work has been done.

Given that I produce multiple (four) drafts of a translation and that the first draft accounts for around 30% of the total time, it's easy to see how the misconception arises. The machine translation system produces the first draft for me, so it should logically save me 30% of my time (assuming that it is fully automatic or at least almost instantaneous).

Unfortunately, the logic is flawed. During my first draft, the actual work of typing only accounts for a fraction of the time needed. Most of the time is spent on thinking about what the source text means, researching associated subjects, and to a lesser extent, how to phrase the meaning in the target text. In progressive drafts, this emphasis shifts, but what also changes is my understanding of the source text. By the time I have completed the first draft, I have an understanding of the source text that I did not have when I began translating the first sentence. If the first draft is produced automatically for me, though, I'm still as unfamiliar as if I had never set eyes on the text. All I have done is postponed the task of familiarizing myself with it to a subsequent draft, which will become correspondingly longer.

...since you have most of the keywords translated already. All you have to do is restructure and make it flow better.


Misconception number two: a lot of time is spent looking up words in dictionaries. If they could be inserted automatically, it would save time. Indeed it would, if the right words were selected, but for almost any word there will be a choice, and the chances of a machine making the right choice are slim. Again, the procedure is wrong-headed: understanding the source text is a prerequisite to choosing the right words, not vice-versa. Machine translation can be useful for providing a gist of a text, but if someone needs such a gist as an aid to translating it, they should steer clear of professional translation.

It would be nice in fact if the software would simply provide a drop-down list of possible alternatives to a selected word, but given the typical specialization in translation, even that is unlikely. I doubt that, used as a dictionary, Google translate could even compete with my own library, much less all the online resources that translators make use of.

What we have here is something like a machine that automatically takes photographs and mixes the paint according to the colours, and the argument from its inventor that this is a time-saver for artists.

The main feature that is not heavily emphasized is when you need to maintain complicated HTML formatting of the document. That is the main reason for this tool that I cannot find anywhere else online. For one phrasers, it's not much help. But if you have a heavily CSS designed HTML document, it comes in handy. If you deal with a lot of HTML documents and need to translate it, you'll understand what I mean. But again, maybe most of you just translate and give the translated text to the HTML designer to paste it back in...


Some of the website work I do is pasted back in by the website designers, it's true. This may be through ignorance, but it may also be because they prefer to have full control of the formatting; they may for instance want to make adjustments for difference in length between different language versions. I think there are some misconceptions between the different professions regarding each other's procedures and I have certainly experienced problems caused by such misconceptions. But generally, I prefer to work directly with the html, since it is in fact one of the easiest formats of all to handle (in addition to the fact that html is more often than not layouted by professionals, and therefore properly, whereas MS Word is often "layouted" by authors themselves). OmegaT is a perfect tool for translating html (and can be found online, being open source), but I am sure it is not the only one.

Marc


Direct link Reply with quote
 

gianfranco  Identity Verified
Brazil
Local time: 17:55
Member (2001)
English to Italian
+ ...
The only possible use is as space filler Jan 11, 2009

I totally agree with Marc.

The only possible use I can imagine for such tool is not for a translator, but perhaps for a webmaster who is looking for a way to please a customer adding to his services the possibility to provide some kind of "translated version" of a website.

Such automatically generated translation would be editable, but still 99% of the linguistic work would need to be done manually. The first draft is some kind of space filler, the pages would be there, the content could be more or less in the desired language, but to have it resembling anything decent a translator would have to go through the text and reword almost everything, using a proper editor for a tagged file.

Needless to say, the Google output is not very useful as a first draft for a translator, as there is no sensible criteria in the choice of words, it would contain many mistakes in structure sentence, grammar and consistency, it would not produce anything good in many subjects and generally would not save much time, for the reasons explained quite well by Marc.

BTW, I tried to get the translation of a simple HTML source page, no more than about 50 words within a simple set of standard tags, but it did not work at all. So, I cannot assess the quality of its linguistic output. Anyway, judging from whatever I have fed to a MT engine, Google or others, it has been only good for generic gisting purposes and never good enough to be used as a draft for a proper translation.

bye
Gianfranco




[Edited at 2009-01-12 00:02 GMT]


Direct link Reply with quote
 
netEditr
English
TOPIC STARTER
Agree Jan 12, 2009

The only possible use I can imagine for such tool is not for a translator, but perhaps for a webmaster who is looking for a way to please a customer adding to his services the possibility to provide some kind of "translated version" of a website.


And this phrase pretty much sums up the whole meaning of this topic. Speaking as a hybrid translator/HTML designer, you NEED space fillers to retain the format of a HTML document when designing in WYSIWYG design mode. On the same account, this is why you sometimes see Lorem Ipsum dummy text in HTML templates but in this case, it's not as useless since the content is still somewhat relevant, just not approriately translated. Think of the machine translated text as place holders where you replace it later with the official translations. And at the same time, you get to see what the page looks like in the target language.

From a professional pov, I'm sure most of you would rather work with the source because you don't want to lose anything during the machine translation. For me, I check both the source and machine translation, if nothing is lost then I continue my proofreading/re-editing process. Maybe by looking at both versions I waste more time but overall it helps me.

So why not just work with the original source with a tool that is considered professional? Personally I find it more satisfying to see everything translated beforehand, no matter if it's correct or not because eventually, I will review each sentence to see if it makes sense and meaning conveyed. It also helps me put my mind into the target language mode when I read everything in the target language even if it becomes ridiculous gibberish. Almost everytime I have to re-edit, re-phrase, and re-word so I agree that machine translation have many errors. This is the same job that I did before when working with human translators, they would send me their final copy and then I do final edits (I was an editor/proofreader in my previous translating career).

Not sure why it didn't work gianfranco, but I don't think you'll have much use for it anyway

Anyhow, each person has their own ways of working and sets of tools. And given the complexity of different softwares and features, sometimes even with a combination of different softwares you are missing a small feature that will glue everything well together. Overall, it has definitely helped me get my HTML translated quickly while I am using a CMS to manage and publish HTML articles.

Regards,
Chris


Direct link Reply with quote
 
netEditr
English
TOPIC STARTER
possible Jan 12, 2009

Samuel,

Ok, I see where you are getting at. I think it is possible to do something similar. I will look into this and add it to my todo list for the next version.

Although the engine is Google Translate, the translation process is a bit different from Google's official method. Which is why I could strip the source language (English) away from the targeted language (Spanish) in the rendered HTML page.

I'll update through this thread once this feature is added.

Regards,
Chris


Direct link Reply with quote
 

Nicole Schnell  Identity Verified
United States
Local time: 13:55
English to German
+ ...
"Pre-translated" text as filler Jan 12, 2009

netEditr wrote:

Speaking as a hybrid translator/HTML designer, you NEED space fillers to retain the format of a HTML document when designing in WYSIWYG design mode. On the same account, this is why you sometimes see Lorem Ipsum dummy text in HTML templates but in this case, it's not as useless since the content is still somewhat relevant, just not approriately translated. Think of the machine translated text as place holders where you replace it later with the official translations. And at the same time, you get to see what the page looks like in the target language.


This can turn a project easily into a nightmare. I consider this a horrible habit.

Example:

A huge website that I had been working on for more than one year (!) went live. It was one of the most beautiful projects ever, the design was gorgeous, the text brilliantly written. Then I looked at it - and I didn't know if I should cry or simply throw up. The thing was crawling with machine translated fillers that for whatever reason somebody "forgot" to replace ("Hey, it looks German! What's the problem?". Can anyone imagine how much work and how expensive it is to proofread such a gigantic website AGAIN? Of course the agency's client didn't pay for this stunt. It took weeks to fix it. Why on earth are IT people doing this? This huge and costly project turned the client into the laughing stock of the year. For the first time ever I wished that somebody got fired (sorry..)

Rant over.





Edited for typo.

__________________________

However, my point is: As long as you are not assigned with translating something - don't touch it.



[Edited at 2009-01-12 02:21 GMT]


Direct link Reply with quote
 
Pages in topic:   [1 2] >


To report site rules violations or get help, contact a site moderator:

Moderator(s) of this forum
Maria Castro[Call to this topic]

You can also contact site staff by submitting a support request »

Bare Website Translation Tool - A Better Google Translate

Advanced search


Translation news





Déjà Vu X3
Try it, Love it

Find out why Déjà Vu is today the most flexible, customizable and user-friendly tool on the market. See the brand new features in action: *Completely redesigned user interface *Live Preview *Inline spell checking *Inline

More info »
CafeTran Espresso
You've never met a CAT tool this clever!

Translate faster & easier, using a sophisticated CAT tool built by a translator / developer. Accept jobs from clients who use SDL Trados, MemoQ, Wordfast & major CAT tools. Download and start using CafeTran Espresso -- for free

More info »



All of ProZ.com
  • All of ProZ.com
  • Term search
  • Jobs