ApSIC Xbench Forum

How to adjust which file appears first in the results?

Hello, thanks for all the good support.
I use xBench often at work, and it’s very useful to use QA reports to compare old and new versions of a single project (since changes appear as Inconsistency in Target).

However, I observed that there is no evident pattern at which file’s segments appears first for each of the items caught by the xBench.

For an easier explanation, suppose that I am running files named “Edited version” and “Old version”. Both have identical sources since they have been extracted from the same file, only that the old version was extracted first before we started post-translation work for the purpose of comparison before and after review/editing.
We would run the QA check, works fine and all of the edited segments appear well.
But in the results, sometimes the segments of the Edited version appear before the ones in the Original (as we would like), and sometimes vice-versa (which we don’t want).

But the issue is that the display order does not affect results only - it affects how the segments appear in the exported QA results, and especially, how the report shows the edited history, like the example in the image.

Ideally we want the edited history showing the Edited version as the final, current translation - its segments being in green rows and the edited history showing the final version in blue text.
But we couldn’t find a consistent way to control which file appears first in the results. We tried editing file names, moving files to a subfolder, etc. - but no method seems to be reliable since what worked for one file didn’t work for another.

If there is a way - even a workaround - to control which file’s segments appear first for each error, could you kindly inform?
Or if no method is possible, at least what pattern xBench uses so that we could try to think of a workaround on our own?
Thank you!

1 Like

In the Xbench installation directory there is the command line ApSIC Comparator tool ccl.exe.

This tool will allow you to compare two versions of the same file.

You can run it from the command line to see the arguments. You basically put the unreviewed files in one a folder structure and the reviewed files in another folder structure. The file names in each directory structure need to match.

So something like…

ccl -unrev c:\from-translator -rev c:\from-reviewer -newandfuzzy -100matches -output comparisonreport.html

… will generate a report that will include all changed segments in new+fuzzy+100%, but not in locked or ICE matches and save it to file comparisonreport.html.

Although it only generates an .html file as output but you can open that .html file in Excel or Word if you wish to add comments to the changes. You can use for example the -openreport Excel switch to open the comparison file with Microsoft Excel.

Thank you for the detailed reply! We tried running the program, but the generated report is empty.
Are only certain file formats supported by chance? We tried using sdlxliffs but the report still shows no segments.