Basic WinAlign shortcuts:
[Ctrl] + [J] = Join Segments
[Ctrl] + [L] = Split Segments
[F2] = Edit Segment
Joining SegmentsYou can easily join two segments without having to use Quick Edit or Advanced Edit. Press the [Ctrl] or the [Shift] key and use the mouse to select the segments you want to join. Select Join Segments from the Edit menu, or right-click and select Join Segments from the context menu or use the [Ctrl] + [J] shortcut.
Splitting a SegmentTo split a segment, double-click directly in the segment you want to split and position the cursor where you want to split the segment. (You can also select the segment you want to split by pressing [F2] to toggle to Quick Edit mode and using the mouse or the arrow keys to position the cursor at the point you want to split the segment.) Select Split Segment from the Edit menu, or right-click and select Split Segment from the context menu. You can also use [Ctrl] + [L] shortcut. WinAlign divides the segment into two segments at the point where you placed the cursor.
WinAlignIntroduction to WinAlign: what it is, when to use it, its limitations
Using WinAlign to align Word documents (demonstration + hands-on)
• Creating a new alignment project in WinAlign
• Setting up the alignment project (adding files)
• Aligning the source and target files
• Reviewing the alignment
• Editing the source and target segments
• Saving the alignment project and exporting the alignment results to a text file
• Importing the alignment results into a translation memory
• Tips to accelerate the alignment process (to avoid aligning each segment manually)
WINALIGN WinAlign is the alignment tool in the TRADOS suite. WinAlign helps you reuse previously translated material by creating a translation memory from source and target material and outputting it into a format that can be imported into Translators’ Workbench.
WinAlign examines source and target language texts and links sentence pairs it feels belong together to form segment pairs or translation units. After you have reviewed and edited the pairs, you can export the results to a text file, which can then be imported into a new or existing translation memory and then used in any of the Workbench editing environments.
Exercise A Aligning Source and Target Files This section covers how to set up a basic alignment project in WinAlign and how to align the source and target data. We will also discuss how to export the alignment results to an external text file.
To align source and target files:
1 Launch WinAlign from the Start menu on your computer (Start > All Programs > TRADOS 6.5 > Translation > WinAlign).
2 Create a new project by click on the
New Project button or selecting
New Project in the File menu. The
New WinAlign Project dialog box is displayed. This dialog box contains several tabs.
3 In the
General tab, specify the following information for your project:
• Click
Source Language to open the
Select Language dialog box where you can select a source language for your project, in this example English (United States). Click OK to confirm and to return to the General tab.
• Click
Target Language to open the
Select Language dialog box where you can select a target language for your project, in this example your A or B language. Click OK to confirm and to return to the General tab.
• Specify a name for your alignment project in the
Project Name box. Let’s call it
Alignment_Exercise.
• From the
File Type drop-down list, select the relevant setting for the files you wish to align. The default setting is Word Documents (*.rtf); however in this case we will choose
Word Documents (*.doc). WinAlign remembers the file type you used when last creating a project and displays it by default when you go to create another project.
4 Click on
Browse under
Intermediate Results Files to select where you want WinAlign to save intermediate alignment results. It is generally a good idea to save your files in one directory. We are going to save it to the desktop.
5 In the
Files tab, source language information is displayed on the left and target language information is displayed on the right.
• In the source language area on the left, click
Add to select the source files that you wish to include in your project. Selected source files are added to the list. In this example, select the sample file,
intro_en.doc.
• In the target language area on the right, click
Add to select the target files that you wish to include in your project. Selected target files are added to the list. In this example, select the sample file,
intro_xxx.doc.
• Click
Align File Names. This command automatically connects each source file with its target language equivalent. If the automatic alignment of file names is not correct, reconnect files manually by following the instructions in step 9.
6 In the
Structure Recognition tab, select
Ignore from the
Structure Recognition Level drop-down list. This switches off structure recognition and generally produces better results in files without any structural elements. However, at home you can define the structure recognition depth by selecting a setting from the
Structure Recognition Level drop-down list (WinAlign supports a structure recognition depth of four levels. This means it can align all Heading 1 to Heading 4 parts of a document pair. The default setting is Level 2). You are now ready to start aligning the selected source and target files.
7 Click OK at the bottom of the
New WinAlign Project dialog box to go to the main project window in WinAlign. Maximize the windows.
8 In the main project window, source file details are displayed on the left and target file details are displayed on the right. To start the alignment, go to the
Alignment menu and choose
Align Project. During alignment, a progress bar is displayed. When the alignment is finished, the progress bars will have reached 100%. Click OK.
Reviewing Alignment Results 1 To review the alignment results, select the file pair and choose
Review File Pair from the
Alignment menu or right-click or select
Review File Pair(s) on the shortcut menu. If you have aligned several file pairs in one batch, highlight any aligned file pair and select
Review File Pair as described above. This opens the alignment editor for the current file pair.
2 A new window appears with source and target file information. This window is known as the alignment editor. (NOTE: If you have aligned several file pairs in one batch, a separate alignment editor window opens for each file pair) The alignment editor is divided into two parts. The top part is referred to as the outline area, and the bottom part is called the segment area.
• The outline area contains general information about the source and target files for alignment.
• The segment area contains information about the source and target segments in the selected files. Each area consists of a source frame, a link frame, and a target frame. Maximize the alignment editor window before you review the alignment results.
3 In the segment area, check that the source and target segment pairs have been correctly aligned. If any segments have been aligned incorrectly, you can realign them manually as follows:
• Right-click on the green icon of the misaligned segment and click
Disconnect. The line connecting the source and target segments will disappear.
• To establish the correct links, place the cursor on the icon of the source segment that you wish to reconnect. Press the left mouse button, hold it down, and drag the mouse to the yellow icon that belongs to the correct target segment. Release the mouse button. A new line appears, connecting the specified source and target segments. You can connect more than one segment. For example, if WinAlign has split a sentence into two segments you can connect both segments by connecting or editing the segments using
Quick Edit or
Advanced Edit in Trados 6.0 or above (see below).
4 When you have finished reviewing the current file pair, go to the
Alignment menu and select
Mark as Finished. If your alignment project consists of more than one file pair, repeat steps 6–10 for all file pairs.
5 When you have finished reviewing the alignment of source and target segments, the alignment results should be exported. Go on to Exercise B
In case you do not have time to review the whole file pair at once, you can save the alignment project as an external file and continue the revision process later. All corrections that you have made to the alignment links will be preserved.
• From the
File menu, choose
Save Project As.
• Specify a name and location for your project file in the
Save Project As dialog box.
• Click Save. WinAlign creates a project file with a *.pjt extension in the specified location.
• To open this project at a later stage, go to the
File menu in WinAlign and choose
Open Project. In the
Open Project dialog box, browse to select the relevant project file and click
Open. You're done.
Connecting, Disconnecting, and Committing Segments 1 To confirm a (dash-)dotted link between a source and target segment, right-click one of the icons in the segment link frame and select
Commit on the shortcut menu. The (dash-)dotted line changes into an unbroken solid line, indicating that this is a 100% aligned unit.
2 To commit or disconnect selected units, select the alignment units you wish to commit or disconnect using [Shift]+[Click], right-click in the source segments area and select
Commit Selected Units or
Disconnect Selected Units from the
Commit or
Disconnect sub-menus.
3 If you are pleased with the alignment results, you can commit all the alignment units in one structure level by right-clicking in a free spot in the segment link frame and selecting
Commit All Units from the
Commit sub-menu on the shortcut menu.
4 If the segments are really messed up within a structure level that has not yet been committed, you may want to disconnect all the alignment units in that structure level and start again. To do this, right-click in a free spot in the segment link frame and select
Disconnect All Units from the
Disconnect sub-menu on the shortcut menu. This removes all links between source and target icons, so you can create your own connections.
Joining Segments You can easily join two segments without having to use Quick Edit or Advanced Edit. Press the [Ctrl] or the [Shift] key and use the mouse to select the segments you want to join. Select
Join Segments from the
Edit menu, or right-click and select
Join Segments from the context menu or use the [Ctrl] + [J] shortcut. To insert a space between the segments you want join, make sure that you select the
Insert Space between Joined Segments option on the
Edit menu. This option is turned on by default. If your text is in a language that does not require a space between joined segments, such as Japanese you can turn this option off.
Splitting a Segment To split a segment, double-click directly in the segment you want to split and position the cursor where you want to split the segment. (You can also select the segment you want to split by pressing [F2] to toggle to Quick Edit mode and using the mouse or the arrow keys to position the cursor at the point you want to split the segment.) Select
Split Segment from the
Edit menu, or right-click and select
Split Segment from the context menu. You can also use [Ctrl] + [L] shortcut. WinAlign divides the segment into two segments at the point where you placed the cursor.
Inserting an Empty Segment You can insert an empty segment above or below the current segment. This is especially useful if WinAlign has split a sentence in two and you would like a clean memory. To insert a segment, select the segment above which you would like to insert an empty segment. Select
Above Current Segment from the
Insert Segment sub-menu on the
Edit menu, or right-click and select
Above Current Segment from the
Insert Segment sub-menu on the context menu. WinAlign inserts an empty segment above the current segment and you can add text directly to the empty segment.
Editing Segments using Quick Edit There are four ways to access Quick Edit:
• Double-click directly in the segment you want to edit.
• Select the segment you want to edit, and press F2.
• Select the segment you want to edit, and select
Quick Edit from the
Edit Segment sub-menu on the
Edit menu.
• Select the segment you want to edit, right-click and select
Quick Edit from the
Edit Segment sub-menu on the context menu.
In each case, a cursor appears in the segment you want to edit. You are now in Quick Edit mode. You can move, copy, add, and delete text at will.
You can easily cut, copy and paste text to and from source and target segments in Quick Edit mode. To cut text, highlight the text you want to cut and select
Cut from the
Edit menu or use the standard [Ctrl] + [X] shortcut. To copy text, highlight the text you want to cut and select
Copy from the
Edit menu or use the standard [Ctrl] + [C] shortcut. To paste text, highlight the text you want to cut and select Paste from the Edit menu or use the standard [Ctrl] + [V] shortcut.
You can undo and redo actions in Quick Edit as well. Undo and Redo apply to actions carried out in one segment only. To undo an action, select
Undo from the
Edit menu or use the standard [Ctrl] + [Z] shortcut. To redo an action, select
Redo from the
Edit menu or use the standard [Ctrl] + [Y] shortcut.
NOTE: To undo all actions and exit Quick Edit mode, press [ESC]. When in Quick Edit mode you can use the arrow keys to move around the segment. The up and down arrows jump to the first and last line in the next segment if you have already passed the first or the last line in the current segment. You can also press the [F2] key to toggle Quick Edit mode on or off.
Editing Segments using Advanced Edit Advanced Edit allows you to edit tags as well as text. If you are editing segments with complex RTF formatting, it is a good idea to use Advanced Edit. Complex RTF formatting features such as index entries in Word documents and jumps in
online help files are usually denoted by curly brackets. Advanced Edit ensures the formatting in these segments are untouched.
There are two ways to access Advanced Edit:
• Select the segment you want to edit, and select
Advanced Edit from the
Edit Segment submenu on the
Edit menu.
• Select the segment you want to edit, right-click and select
Advanced Edit from the
Edit Segment sub-menu on the context menu.
• Select the segment you want to edit, and press [Ctrl] + [I].
WinAlign displays the text as RTF code in the Advanced Edit box. Make your changes to the text. Click OK.
Exercise B Exporting Alignment Results When you are satisfied with your alignment, you can export the results. You have two choices of export format: Translator’s Workbench Import Format or Translation Memory Exchange Format (TMX). TMX is the standard exchange format for translation memories; however, the default setting is Translator’s Workbench Import Format. If you select Translator’s Workbench Import Format, you determine which additional information (project attributes, names, or dates) should be generated and WinAlign then creates the import format for Translator’s Workbench.
You can create a translation memory or a termbase from the WinAlign export file. You can import the alignment results into a translation memory in Translator’s Workbench or into a termbase in MultiTerm, and start re-using your previous translations immediately.
Setting Export Options In order to export alignment units in Translator’s Workbench Import Format, you must first define what project information should be exported in addition to the alignment units before you can start the actual export. These settings are defined in the
Export tab of the
Edit WinAlign Project dialog box.
Some examples of export information include: Creation User, Creation Date, Text Field ID Code (for example, a specific project name), Attribute Field Client (e.g., Trados), and Attribute Field Domain (e.g., software, pharmaceuticals, etc.). To review your settings:
1 Select
Project on the
Settings menu and select
Export tab.
2 Before specifying project information, define which alignment connections should be exported using the
Export Threshold slider.
3 If you set the Export Threshold to 1%, WinAlign exports every alignment unit except those that were disconnected or not aligned in the first place. If you set the Export Threshold to 100%, WinAlign only exports units that you have committed.
4 By default, WinAlign exports the project as a Translator’s Workbench import format (TXT) file. This format retains the formatting information. To export, without formatting information, select
Export as Plain Text.
NOTE: It is a good idea to add content to the text and attribute fields that you define. If you do not, this may create problems when you go to merge the export file into a translation memory.
Unless you specify a Creation User, WinAlign specifies Align! as the default creation use. This ensures that Translator's Workbench (and you) takes the alignment penalty set in Translation Memory Options in Translator’s Workbench into account.
Exporting the Results 1 We are now ready to export the results. Click OK to return to main WinAlign window. Select
Export Project from the
File menu.
2 Select a name for the file and click
Save (save it to the Exercise E folder).
3 A progress bar is displayed during export. When the export is finished, the progress bar will have reached 100%. Click OK.
4
Send me this .txt file in an e-mail (NOT the .pjt file). This .txt file is used to import the results into Workbench.
Importing Alignment Results into Workbench You can import alignment results into an existing translation memory or into a new memory. This section covers how to import the alignment results into a new translation memory.
To create a new translation memory and import alignment results:
1 Launch Translator’s Workbench from the
Start menu on your computer (Start > All Programs > TRADOS 5.5 > Translation > Translator’s Workbench).
2 Choose
New from the
File menu. In the
Create Translation Memory dialog box, select the source and target language for your new translation memory (English US to English UK in this exercise) and click
Create. Note: The source and target languages in the translation memory must correspond to the languages in your alignment project.
3 Specify a file name and location for the new translation memory. Let’s call it
MultilingwebTM and save it to your
Desktop. You have just created an empty translation memory, which remains open in Translator’s Workbench.
4 To import the alignment results go to the
File menu in Translator’s Workbench and choose
Import. It is not necessary to define text and attribute fields (such as ID Code, Client, and Domain) in the Translation Memory setup before import. Translator’s Workbench adds these fields automatically if you accept the default Import setting,
Add to Setup.
5 In the
Import dialog box, under
Import Mode, select
Large Import File. Click OK to accept the remaining default settings.
6 In the
Open Import File dialog box, browse to select the export file (*.txt) that contains your alignment results and click
Open.
7 Translator's Workbench imports the segment pairs into the translation memory where they are stored as translation units. When the import is finished, the import results are displayed in the Workbench status bar. Your translation memory can now be used to translate future projects.
Exercise C (if we have time, otherwise do this as your homework) 1 Go to http://www.un.org/apps/pressreleases/ and choose a text in the Press Releases to align that is at least 3 pages long. Then save the corresponding A or B language text.
2 You either can save the texts as HTML files or you can click on the file, copy the text and paste it into Word.
3 Align it in WinAlign following the steps above
4 Export the results into a .txt file and send the .txt file to me.
Homework Find a text and translation that is at least 3 pages long and align it in WinAlign. Export the results into a .txt file and send the .txt file to me by March 6, 2007.
Miscellaneous Tips Structure Recognition When aligning source and target texts, WinAlign utilizes the fact that documents are usually structured and divided into various sections. For example, when a document is created in Microsoft Word, it usually contains structural elements such as Heading 1, Normal, etc. The structural elements in the target text generally have the same formatting as the source text. Other text formats such as SGML/HTML or DTP programs such as FrameMaker use tags for this purpose. WinAlign uses this information to create a structure tree for the source and target documents. Even when the document pairs do not have a clear structure, WinAlign uses information such as font sizes and carriage returns to perform structure recognition.
Identifying Segment Pairs Once the structure has been determined, WinAlign begins aligning the individual segments. A segment can be a sentence, title, footnote, list element, caption, or any other textual unit. WinAlign considers a large number of factors during alignment, which helps to produce a high number of matching segment pairs.
The Alignment Editor Once the first alignment run is finished, WinAlign displays the alignment results in the alignment editor, where you can manually edit the suggested segment pairs. You can combine, move, delete, or even change a segment’s content to correct typos in the alignment editor. By joining or deleting segments you can influence the final results. The text is displayed with its original formatting. This helps you visually identify segments at a glance, which saves you from having to carefully read each individual segment. For example, if a segment pair has a few bold terms you can easily identify whether or not it is a match.
In addition to allowing you to edit the segment pairs, the alignment editor also displays the overall outline of the documents in a tree diagram, which enables you to reassign text blocks at this level.
The source and target segments are all connected with dotted lines or dash-dotted lines. This means that WinAlign assumes they are reliable alignment units. The more reliable WinAlign feels they are, the closer the dots are between source and target segments. After you confirm an alignment unit, the dotted line is replaced by an unbroken or solid line and the alignment unit is “committed.” It is a good idea to spot-check structure levels for misalignments and correct them as necessary. If nothing is misaligned, leave the structure level as is. However, if alignment goes wrong at some stage, it is useful to disconnect some alignment units, make new connections as needed and save your changes. Once you have edited the structure level, work your way through all remaining outline levels of the source and target text. Start from the top level and work your way to the end.
Re-aligning the File Pair After correcting misalignments and confirming new ones throughout the whole file pair, you can run a second alignment on it. This is referred to as re-alignment. During re-alignment, WinAlign takes into account any changes made when editing the initial alignment result. For example, WinAlign uses any committed alignment units as reliable anchor points. To re-align the file pair, right click in the segment link frame and choose the
Align Structure Level command from the Project window shortcut menu. You can also select
Align File Pair on the
Alignment menu in the Project window.
Formatting Significance WinAlign can take various tuning options into account to achieve optimum alignment results. For example, if formatting is identical across both source and target texts, it can be a useful aid to the alignment process. Set the tuning options in the Alignment tab. For the purpose of this example, follow these steps:
1 Select the
Alignment tab.
2 Move the
Formatting Significance slider towards the
High value. The more you slide it towards High, the more WinAlign takes internal formatting changes inside source and target segments into account when assessing the
reliability of a segment pair. Since internal formatting (for example, bold on/off or font changes) is used in the sample file pair, you can drag the slider further into the High half.
3 Accept the other default settings.
To highlight internal, external and non-translatable tags in the source and target segments, select the
Tag Text command from the
View menu. This makes the translatable text very obvious and reduces the possibility of you deleting or editing tags inadvertently. You can choose to display partial, complete or no tags.
Aligning HTML files WinAlign provides direct support for HTML alignment, including any HTML derivatives such as Active Server Pages (ASP), Active Server.NET (ASPX and ASPC), Java Server Pages (JSP), and Include files (INC). Tags are presented in the same way as they appear in WinAlign. For example, the anchor tag appears as
or the bold tags appear as and in the alignment editor.
In general, no major file preparation is required. Check the files for alphabetically sorted lists. If they are present, re-sort the lists so that the target files match the sequence of the source files.
WinAlign uses an independent algorithm that enables structure recognition based on standard tags(h1, h2 etc.).
To edit an INI file for a HTML document, specify the file as an XML or SGML file, in the File Type drop down box on the Project Settings General tab. This enables the DTD Settings… button, which you can use to access the DTD Settings dialog box.
Adding HTML files for Alignment
If your HTML files are spread over several subfolders, use Windows Find Files in conjunction with the Drag & Drop support in WinAlign. To add your source and target files to the alignment project:
1 Select the Project command on the Settings menu to open the Edit WinAlign Project dialog box. Click the Files tab.
2 Select the Find submenu on the Windows Start menu and select Files or Folders. Use the Browse button to locate the root folder of your source HTML files. In the Named box, type *.htm; *.html. Click Find Now. Windows searches all folders containing HTML files and lists them in the lower part of the Find Files dialog box.
3 Drag and drop all desired files from the Find Files dialog box into the WinAlign source or target file list as required.
Alignment Strategy
When creating a new WinAlign project of HTML files, define your general settings in the General tab of the New WinAlign Project dialog box as follows:
• Choose HTML Files from the File Type drop-down list in the General tab of the New WinAlign Project dialog box.
• In the Alignment tab, define the weighting of the various tuning options according to the specific document features of your alignment pairs.
• In the Structure Recognition tab, you need only set the structure recognition depth. WinAlign automatically identifies and assigns the HTML heading tags (for example, H1 and H2) to the structure recognition Levels 1 – 4. This is why the levels are de-activated when HTML has been chosen as a file type.
• Run the first alignment of a representative file pair and spot-check the result. If necessary, change the Alignment and Structure Recognition options to suit your alignment pairs. Repeat this procedure on the representative file pair until you are satisfied with the initial alignment result. For example, it may be better to set the structure recognition depth to Level 2 instead of Level 4 even if your HTML documents have four heading levels.
• After adjusting the Alignment and Structure Recognition options, run the alignment of the whole project. After the alignment, check, edit and run re-alignments as necessary. Once you are satisfied with the project alignment, export the results.
Troubleshooting Tips
• If you are trying to align files and WinAlign keeps crashing on you. Save the files as text only files (.txt or Plain Text) in Word and align those files. The formatting is probably to blame, and you really don't need the formatting for your translation memory.
• If that doesn't work, try breaking the files down into smaller chunks. Sometimes WinAlign has problems with really large files.
http://appling.kent.edu/ResourcePages/Courseware/Terminology_ComputerApps/winalign_handout.pdf