Today, iStock becomes truly international. We are thrilled to announce that our whole site is now available in English, French, Spanish, and German. Choose a language in the top right-hand corner. The site will always be in your preferred language when you log in, until you change your preference again from your control panel.

The financial side of iStock has gone international as well. Clients can complete transactions in US, Canadian, and Australian dollars; Euros, Japanese Yen, and British pounds. Just set your currency preference, and whenever you use the store, your totals will all appear in your native currency.

Finally, and possibly with the biggest implications, keyword searches on iStock are now available in 12 different languages. You can search in UK or US English, French, German, Spanish, Portuguese, Brazilian Portuguese, Russian, Italian, Japanese, Polish and Greek. Just enter the keyword in any one of these languages, and the engine will find any image in the library with that keyword. Searches in Japanese and Cyrillic characters are also supported. More languages will be released soon.

This is all possible because of a massive overhaul to our image tagging system. We have rebuilt and reprocessed our entire library using Getty Images' patented controlled vocabulary system. It means more than languages - everything about the system, from how contributors enter keywords, to how the engine returns results, to how you choose and sort between results, is clearer, more focused, and easier to use. It's a massive leap forward for how clients find images at iStock.

Entering Tags and Terms


There are a few more steps to the new system for contributors, but they will ensure that every file is more accurately tagged. Once you get the hang of it, you'll love it. There are a couple of new words to learn:
-A tag is a keyword or phrase.
-A term is a tag with a specific meaning assigned to it.

Step One: Is it a word we recognize?
Enter the tag. If the word is misspelled, or otherwise unrecognized, you will be offered several choices as to what we think you meant. Choose one of the suggestions, or accept the tag as you entered it.
Step Two: What language is it in?
That's right, contributors can keyword in any one of the 12 supported languages. When you enter a tag, you may need to specify the original langauage if the engine recognizes multiple possibilities.
Step Three: What does it mean?
Now you will be asked to choose between possible meanings for the tag. You can choose as many as are appropriate. For an image of an orange, you will be offered Orange - Fruit, and Orange - Color. There will be lots of times when the possible meanings are varied, and some will not be appropriate at all. Choosing the right ones will ensure your file turns up in the right searches.


Once you have chosen at least one term for each tag and saved the file, those tags become searchable. Your tag is automatically translated into all the supported languages.

So now a user does a search for orange. They will be asked if they meant the color or the fruit, and they have the option to see the results of both. They will get all the images with those terms attached, regardless of the language the term was originally entered in.



Questions?

We're sure you have a lot of questions. We took a guess at what some of them might be — follow the links below to forum discussions with our VP of Development, Aaron Springer.

Do contributors need to re-keyword their images?
No. Don't panic. All 1,000,000+ files already in the library have had their keywords translated into tags, with at least 1 meaning assigned. Contributors can add more terms if needed, and correct any terms that we may have gotten wrong. Double check potentially ambiguous files to make sure that they are termed the proper way.


There must have been a lot of words in the old library that aren't in the controlled vocabulary, though. What about them?
The old library contained thousands of words which the system simply couldn't understand - drastic misspellings, specialized naming conventions, or words that just weren't in the vocabulary. Rather than change or drop any of these, we made the system learn them. So none of your images will lose their unique tags.
None of these unknown tags from the past are currently translated, though. On the File Edit page, we've highlighted these unrecognizable tags. Please fix them if they were at any time misspelled.

Can we still do Boolean searches?
The Boolean operators - and, not, or - are available in English. The controlled vocabulary can't contain connective words (the, a, and... words with logical functions that don't denote anything). Our international users can use numeric symbols: + or - for 'and' or 'not'.


Can we search for phrases?
Phrases are tags, and are searchable. Contributors add phrases as words separated by spaces. Clients search for phrases surrounded by quotation marks. In the past, words separated by commas included by the contributor could be found by a search within quotation marks. Not anymore. The phrase is now either there, or it isn't.

What about ImageManager?
Right now, the system is going to automatically assign meanings to the tags it receives from ImageManager. Until there is a way to assign terms right in the IM application, you should go in after the file is uploaded to double check potentially ambiguous terms, just in case the wrong one is picked. Also, watch those spelling mistakes - if your tags are misspelled, IM will make the system learn the misspelling, rather than correct it.


What about Categories?
Each file still needs to be included in at least one category.

Obviously, you've got more questions. Please read this discussion for the latest information on what's going on.
Disambiguation, Search and Uploading - Questions Answered


Images
Global generation by nico_blue
Confused by amybuzz