The Language I/O Help integration for Salesforce.com automates the pull of Salesforce Article content from the knowledge repository for translation. It then marshals that content through translation services and finally it imports translated Article content back into Salesforce. This documentation will cover the process of requesting translations within Salesforce as well as an overview of how the rest of the process is automated.
The Language I/O HELP integration for Salesforce uses the existing translation request functionality provided by Salesforce. However instead of requiring the Salesforce user to manually export content for translation and import the translated content back in - as is currently the case if you attempt to leverage translation memory and standard translation tools - Language I/O automates these steps.
As show in the above screenshot, the Salesforce user selects the English articles that are ready for translation and hits the Submit for Translation button, provided by Salesforce by default.
At that point, a window pops up, prompting the user to select the languages into which the article should be translated, as shown in the next screen shot.
Once the desired languages are selected, it is important to ensure that the correct translation queue is set next to each language, such as “Translations Queue” in the example. Language I/O will be configured to query this queue to determine which articles are ready for translation. It is not necessary to have a separate queue per language, only that you use the queue that has been agreed upon by you and your Language I/O project manager.
When you hit OK, your work as a content manager is basically done and Language I/O picks up the work from here.
When a user requests a translation of an article via the process described in steps 1 and 2, a new draft translation article is created for each English article that was flagged for translation. Language I/O will look for all articles that are awaiting translation inside of the named queue.
Our linguists, however, need the latest version of the English content associated with these draft articles. So during the pull process, Language I/O stores the ID of the foreign-language drafts (so Language I/O knows where to push translations) but Language I/O pulls the English content from the English, Salesforce Source Article.
Language I/O queries the specified Salesforce queue, pulls all Articles from Salesforce slated for translation and pushes the content to our linguists or machine translation services (depending upon your agreement with us) for translation.
Once the translations are complete, Language I/O pushes the translated content into the draft articles sitting in the queue from where they were originally pulled. We can configure Language I/O to leave the new translations as draft versions in the queue, leaving the content manager to publish them. Or the articles can be immediately published.
In addition to automating the pull and push of content during the translation process, Language I/O also automates the re-writing of links to image files, related articles and websites so that a translated article links to translated resources.
Take the below article for example. If all Language I/O did was to translate the content, the Japanese version of this article would still pull in a screenshot with lots of English labels.
When Language I/O is preparing to push the English content to the Japanese linguists, it can be configured to first re-write links to images such as the Cirrus audio recording image above so that a localized version of the resource is pulled into the translated article. In this example, it will re-write the path to the English image file so that it instead pulls the same screenshot from a path designated to store Japanese localized versions of the image files, as shown in the image below.