Segments and Match Types in BWB
When you open the Browser Workbench (BWB), the target is populated with ICE and 100% matches. As you translate the task, save to translation memory, and reapply the translation memory, the number of 100% matches will increase.
In the Browser Workbench, you also can place segment comments and placeholders, split and merge translation segments, work with terminology database entries for translation segments. You also can lock segments.
Each segment in the Browser Workbench is in one of a number of states. The following table explains segment match states. The color in the first column is the border color of the segment.
| Color | Match Type | Description |
|---|---|---|
| Purple
| In-context (ICE) match | The segment has a 100% match from translation memory that automatically populates the target. An in-context exact match occurs when a segment has been translated before, in the same segment of the same asset. When a segment has an in-context exact match, the target segment has a purple bar to its left. When a locking workflow is used, there is also a lock icon that appears, and the translator cannot modify that segment. |
| Blue
| 100% single exact match | The segment has 100% match from translation memory that automatically populates the target. A 100% match occurs when a segment has been translated before, whether in the same asset, or in a different asset with the exact same segment. When a segment has 100% match, the target segment has a blue bar to its left. |
| Double blue
| Multiple 100% distinct exact matches | The segment has multiple distinct 100% matches from translation memory that automatically populates the target. A 100% match occurs when a segment has been translated before, whether in the same asset, or in a different asset with the exact same segment. When a segment has 100% match, the target segment has a blue bar to its left. When a segment has multiple distinct matches, the target segment has a double blue bar to the left, meaning the translation is different. By default, this feature is disabled because it affects performance during 100% matches. |
| Striped blue
| Repaired 100% match | This state reflects a segment that has been repaired up to a 100% in context match. A 100% match occurs when a segment has been translated before, whether in the same asset, or in a different asset with the exact same segment. Translation memory can "repair" a match by adding or removing placeholders or numbers, thus creating up to a 100% match. When a segment has 100% match, the target segment has a blue bar to its left. When a segment has a repaired 100% match, the target segment has a dotted blue bar to its left. |
| Gold
| Fuzzy match | The target segment contains a match from translation memory that is less than 100%. Fuzzy matches are not automatically populated into the target. You can choose to populate all empty targets in the document with the highest scoring match. When a segment has a fuzzy match, the target segment has a yellow-orange bar to its left. |
| Striped gold
| Repaired fuzzy match | The target segment contains a repaired match from translation memory that is less than 100%. Fuzzy matches are not automatically populated into the target. Translation memory can "repair" a match by adding or removing placeholders or numbers, thus creating a fuzzy match. You can click the translation memory match button for the segment (pages icon) and populate the target on a segment-by-segment basis, or you can choose to populate all empty targets in the document with the highest scoring match. When a segment has a fuzzy match, the target segment has a yellow bar to its left. When a segment has a repaired fuzzy match, the target segment has dotted yellow-orange bar to its left. |
| Green
| Manual translation | The segment has been manually altered by a user. When a segment is manually translated, it has a green bar to its left. |
| Dark red
| Auto translation | WorldServer has determined the segment does not contain any translatable content. The source content was copied to the target side and the segment is considered to be translated. The rules for determining whether the content is translatable or not can be customized through the SDK and translation memory properties. |
| Blue-green
| Machine translation | The Machine Translation marker indicates the segment has been populated from a match in a machine translation application. |
| Black
| Repeated Segment | The repeated segment marker color denotes the first occurrence of a segment that is repeated within an asset or across multiple assets within the same project. Translating this occurrence can provide a translation entry in the translation memory that can be used to translate the other repetitions. |
| Black
| Repetition | The repetition marker color denotes the segment as being a duplicate of a repeated segment. The actual segmented that is marked as being repeated may be located in the same asset or another asset within the same project. |
| Gray
| Not translated | The segment may or may not have fuzzy matches available, but does not have 100% match and has not been altered by any user. |
Filtering the Segments Displayed
Figure 1. Drop-down list for filtering segments displayed
By default, all segments are displayed. Select a different filter option from the menu to display only that segment type (for example, if you select All non-translated, the Workbench will refresh to display only the segments that WorldServer determines not to translate, but to instead copy the source into the target segment). You can also search for specific text using this menu.
The first group of filters in the drop-down list are based on the type of match found during the leverage process. The bottom group is based on the segment status.
| Filter | Description |
|---|---|
| All segments | Shows every segment in the asset. |
| All except ICE | Shows every segment except those for which an ICE match was found. This is a good filter for a translator to use when working, because usually the translator can skip ICE matched segments. To be characterized as an ICE match, a match candidate must satisfy the basic criteria for ICE matching—the source text must be an exact match, and must share a common usage context with the translation memory entry providing the match. When a segment has an in-context exact match, the target segment has a purple bar to its left. When a locking workflow is used, there is also a lock icon that appears, and the translator cannot modify that segment. |
| All except ICE and 100% match | Shows every segment except those for which an ICE match or a 100% match was found. 100% matches are less dependable than ICE matches, but you might want to exclude 100% matches as well as ICE matches, to see only those segments that need the most work. |
| All 100% matches | Shows only segments that are 100% matches. This does not include ICE matches. You might want to inspect 100% matches to see how dependable they are. |
| All repaired 100% matches | Shows only segments that have been repaired to be 100% matches. Repairs are in-memory changes; they are never automatically stored in TM translation records. If the translator agrees to the repairs, the translator can choose to add new translation records to the TM when the asset being translated is saved. |
| All auto-translated | Shows only segments processed by the translation memory auto-translation mechanism, which copies the source to the target when it determines that the source does not need translation. For example, if a segment consists only of punctuation, it will be auto-translated. Auto-translated segments have a dark red color status bar. You might want to view the auto-translated segments to see if in fact you agree that they don't need to be translated. |
| All non-translated | Shows only segments which have no status (showing up with a gray bar). This is the initial state of a segment that is not an ICE or 100% match. Viewing only these segments will show you segments that have not be translated at all. |
| All matching text | When selected, this filter includes a search box, in which you can enter text for which you want to search. Only segments that include this text are displayed. |
| All with comments | Shows only segments that have comments attached. This is helpful for finding comments from the reviewer to the translator, or vice versa. |
| All with no review status | Shows only segments that have no review status assigned to them. |
| All pending review | Shows only segments that have been marked to indicate that they have been translated and need review. If a segment is marked as Pending Review, then it is the reviewer now "owns" the segment. |
| All reviewed | Shows only segments that have been marked to indicate that they have been reviewed, and are ready to move on through the workflow and be added to the translation memory. |
| All rejected | Shows only segments that have been rejected and should go back to the translator. When a translator has had an asset returned by the reviewer, this is a good filter to show the problem segments. |
| All with errors | Shows only segments that have errors. When a translator (or editor) has an asset sent back from the reviewer, this filter lets them view only the segments marked with errors. It is a good idea to enable "Show Segment Details" in the workbench toolbar when using this filter. |
The 100% matched count shows the number of segments in the asset with content from translation memory as 100% matches. This includes ICE matched segments as well as single exact matches (segments with blue and purple bars). Normally this information is applicable only the first time the asset goes through segmentation. If you change a 100% matched segment manually and save the changes, the segment then will not be included in the 100% matched segment count because the content of that segment is now coming from user-entered data rather than the translation memory. However, if you choose Save and Update TM, the count is refreshed to reflect a new 100% segment count.
Update functionality is useful for refreshing the counter after translation memory is reapplied to a segment or asset. If there are more 100% matches found after the user reapplies translation memory, then the 100% matched counter reflects this.
Note that although the counter includes both ICE and 100% matched segments, the segment filter shows 100% matched segments only (blue bars) and does not show ICE matched segments.
In a large file (where there are over 100 segments on a page—which is possible if you have View all pages selected), the system stops checking TM and TD matches per segment, to limit the number of queries being issued. In this case, a question mark is placed on the match indicator icon and you need to click on the item to see whether there are any TM or TD matches.
Highlighting Differences in Fuzzy Matches
To help you see the differences between your source segment and the fuzzy match found in the translation memory, WorldServer provides a Highlight Differences feature, which is off by default. When you enable it and click the Translation memory matches icon to the right of a fuzzy matched segment pair, the Matches for Segment <N> dialog window appears:
The gray highlight indicates that the TM entry contains the word “chapter” and “to” but the source segment does not. The yellow highlight means the TM entry does not contain “section” or “you” but the source segment does. If the fuzzy-match score is high enough, you could paste the target entry into your translation’s target segment, and then use this information to adjust the target translation accordingly. In this example, you would simply replace "Este capítulo" with "Esta sección".