Automatic Translation and Translation Memories
The future of multilingualism and the success of multilingual communication rely on the progress made in the field of Automatic Translation, or Machine Translation (MT). Unless we spread the awareness of the existence of MT, no progress can be made in this domain. Collaboration between writers and MT specialists is vital.
Web writers and authors should be encouraged to tailor their texts according to the basic needs of MT. Once the text is polished and prepared for MT, authors can then tell their readers that their pages are "Machine Translation-Friendly", or "Machine Translation Ready". For this purpose, adding a small label will distinguish your friendly site from other unfriendly sites and your MT-ready content from the rest of content.
It would be very useful to place such label on your front page if all your content is "MT-Friendly" or on the pages that you think they meet the following minimum requirements:
1. Short paragraphs
If Machine Translation system is using Google AJAX Language API, then your paragraph should be less than 500 characters (including spaces); the rest will be dropped!
2. Short phrases
Avoid long phrases as much as you can.
3. Clear phrases
Avoid ambiguous usage of linguistic components (subject, verb, object, etc.). Unclear phrases will produce erroneous translations. Use your knowledge of a second language to imagine translation scenarios; how this would translate into that language?
4. Correct spelling, no typos
Wrongly spelt words will not be recognized by MT systems. These are the "white socks" of your elegant site! Little extra effort will give your site an elegant look.
5. No slang or invented words
Words which do not exist in known dictionaries will be ignored by MT systems. No MT Data Base will contain all your vocabulary.
6. No highly elaborated terminology
Try to simplify the usage of special terminology as much as you can; use common synonyms when possible.
7. No acronyms or abbreviations at all
Acronyms and abbreviations can be translated into anything; they should be avoided completely.
8. Don't mix languages
Mixing languages in your text is confusing for humans; for the machine it can be a nightmare. If you have to include text from other languages, declare the language of that text in your HTML code.
9. Clean pages
Clean and correct HTML code will make the life of MT much easier; messy code can cause your translated text to break.