Forum: SDL Trados support
Topic: Fragment repetition count without translation memory
Poster: Emma Goldsmith
Post title: Segment or fragment?
[quote]Mar Jiménez Quesada wrote:
Does the software include in the repeated fragment count the first instance of a repeated fragment [/quote]
By default, a segment is a standalone unit that usually ends with a full stop. So a sentence is the most common-sized segment.
A fragment is a chunk or group of words within a segment. It isn't a standalone unit.
In the above explanation there are 4 segments. "a standalone unit" is a fragment that appears in two of those segments.
I expect you're referring to segments. In that case, where there are four identical segments, Studio considers the first as new content and the other three as repetitions.
If you're talking about fragments, then if the fragment is long enough (supposing it accounts for 80% of the segment), then it will be picked up as a fuzzy match. To get Studio to account for repeated fragments (fuzzy matches) within a file and without a translation memory, simply mark "report internal fuzzy match leverage" in the Analyze Files batch task. Again, Studio will consider the first segment as new content, and subsequent "fuzzies" as such.
In Studio 2017, fragments are processed in a new way, which I discussed in this post:
[url removed]
Topic: Fragment repetition count without translation memory
Poster: Emma Goldsmith
Post title: Segment or fragment?
[quote]Mar Jiménez Quesada wrote:
Does the software include in the repeated fragment count the first instance of a repeated fragment [/quote]
By default, a segment is a standalone unit that usually ends with a full stop. So a sentence is the most common-sized segment.
A fragment is a chunk or group of words within a segment. It isn't a standalone unit.
In the above explanation there are 4 segments. "a standalone unit" is a fragment that appears in two of those segments.
I expect you're referring to segments. In that case, where there are four identical segments, Studio considers the first as new content and the other three as repetitions.
If you're talking about fragments, then if the fragment is long enough (supposing it accounts for 80% of the segment), then it will be picked up as a fuzzy match. To get Studio to account for repeated fragments (fuzzy matches) within a file and without a translation memory, simply mark "report internal fuzzy match leverage" in the Analyze Files batch task. Again, Studio will consider the first segment as new content, and subsequent "fuzzies" as such.
In Studio 2017, fragments are processed in a new way, which I discussed in this post:
[url removed]