Best Practices

Statistics

There are some discrepancies between the number of characters entered and the number of characters that we send to MT providers, it depends on the processing filters selected.

There are a few options for how the statistics are calculated, they are described below:

The translation template selected does not have any processing filters. In this case, the number of characters entered in the console is exactly the same as we send to MT providers as the text is not processed additionally.
The translation template selected has the processing filters that do not require tag processing: Remove Tags, Space keeper, Tag spaces cleaner, Tags encoder. In this case, the number of characters entered in the console is exactly the same as we send to MT providers as the text is not processed additionally.
The translation template selected has an HTML tags keeper filter applied. In this case, additional text preprocessing is required, as the request that the console sends to the MT provider includes additional characters responsible for better handling of the tags, thus there are discrepancies between the number of characters entered in the console and the number of characters in the request we send to MT providers.

Two scenarios:

For text that does not have tags, 17 characters are added.
For text that contains the tags, 93 characters will be added. For example, for the following text (<a href=\"#\">Привет!<spanclass=\"mob\">En</a>) 93 characters will be added. The text contains 81 characters, and the console will send the request containing 174 characters.

The translation template selected has an HTML keeper ver. 2.0 filter applied. In this case, additional text preprocessing is required, as the request that the console sends to the MT provider includes additional characters responsible for better handling of the tags, thus there are discrepancies between the number of characters entered in the console and the number of characters in the request we send to MT providers. This filter is better suited for tag processing helping to provide a more natural translation due to the complex algorithm applied, still in beta-testing.

Two scenarios:

For text that does not have tags, 41 characters are added.
For example, for the following text (<a href=\"#\">Привет!<spanclass=\"mob\">En</a>) while the text contains 81 characters, the console will send the request containing 55 characters, so the number of characters in the request the console sends will be less than the actual number of characters entered.