Common Use Cases
First of all, CAT application does not perform translation independently or automatically (as in machine translation). CAT application is a tool used by human translator, and all the translations are done by human. In CAT, translation is performed segment by segment, as the program brings next segment after the previous one has been completed.
CAT program automatically searches for the same or identical segment in translation memory for each segment in the text. Translator translates sentence by sentence, while CAT program independently provides already existing translations found in translation memory, or updates translation memory with new segments, as translator translates them.
Example:
|
Custom Report Tutorial |
This sample text contains a lot of sentences which resemble, and sometimes even repeat each other. There are two general match types between segments in CAT. These are "exact match" and "fuzzy match".
А. Exact-match
It is evident that the same segments (which are highlighted with the corresponding colors) are translated identically. Translator equipped with translation memory does not need to return to the previously translated segments to check how exactly they were translated. Instead of this translation memory independently provides a complete human translation of a segment. This means that translator does not have to spend time translating one same segment twice.
If translator has already translated the following sentence: " Foreword: the Concept of Custom Reports” (it is highlighted in blue) in the very beginning of the text, CAT program will independently translate this sentence which appears in the text as introduction title.
B. Fuzzy-match
Unlike exact-match segments, fuzzy-match segments can also appear in a text. In the previous example these are the following sentences:
|
Projetex supports custom reports created with the help of FastReport editor. |
|
Projetex supports custom reports created with the help of FastReport |
|
Projetex supports custom queries created with the help of QueryBuilder editor. |
These segments are translated almost the same, although they do not fully match. CAT tool will determine similarity between the sentences and, when the translator comes to the next similar segment, will suggest one of already existing translations pointing that source text match is incomplete and marking elements, which differ in the two matching segments.
C. Second use of translations (translation recycling)
Among the three sentences mentioned above translation of the first segment left in TM can be used to translate second segment in the list as well. As for the third segment, memory of similar segments already accumulated in TM will also be useful for translator, who will only have to edit nearly completed translation. This is the concept behind the translation recycling.
As translator processes more and more segments, TM will provide increasingly efficient variants of translation even for different (at first glance) source segments.