CafeTran Espresso 2016 Ichiro - Update 20


CafeTran Espresso 2016 Ichiro - Update 20 is available for download. The update can be performed via the Drag & Drop procedure as follows:

1. Run CafeTran.

2. Download file from here (do not unzip or rename the file after downloading) and place it on your desktop.

Note: On Mac OSX, the Safari web browser may unzip the file automatically after downloading. To prevent this, just use the right-click menu and choose the "Download Linked File" command.

3. Drag and drop the downloaded file into the Project Dashboard and restart the program.

4. Drag and drop the downloaded file into the Project Dashboard and restart the program. Second time!

    The step 4 above is only for the users who missed the preview (Harbinger) and Ichiro updates with the new icon sets.

On Mac OS X, the Java 1.8.0_77 update broke the drag and drop function in CafeTran. The fix is available starting from CafeTran's Update 4. If you skipped all the updates from 4 to 19, please choose the downloaded update file via the Help > About > Install Update button instead of dragging and dropping it into the Project Dashboard.

What's new:

1. Maintained the font type and size when applying the custom formatting in Word documents (e.g. bold, italics).

2. Added the settings for the font size of tags: View > Font > Tags.

3. Further increase of the button fonts in the main interface for 4K (3840 pixels) resolution screens.

4. ProjectTerms glossary files are saved with the line separator default in the used operating system.

5. Total Recall can connect to MariaDB database (MySQL replacement for Linux systems).

6. Added the Quick search... menu item in the Edit menu to show and hide the Search tool bar.

7. The Matchbar becomes a dockable or floatable Matchboard.

The new Matchboard groups all the matches for the current segment including the matches from the connected Machine Translation resources. The results in the Matchboard can be both clicked and selected to transfer them to the target segment. The match type and the resource name are also shown there. The Matchbar, placed at the top of the Matchboard now, is still functional to quickly display the matches without the need for scrolling in the long list of the matching results. Note the convenient keyboard shortcut when typing in the target segment pane: double SHIFT + the first letter of the source match to narrow down the long list of the results.


8. The improved Auto-assembling algorithm calculating the weight of the match with the following order of preference:

- Context segments

- Exact segments

- Prefix segments (translation memories with the Prefix matching on)

Context fragments (Exact and Virtual TM fragments with the matched context)

- Exact fragments and terms

- Virtual fragments ('guessed' TM matches with the high number of hits)

- Other Fuzzy fragments and terms (regular expression matches, prefix and stem matches, 'guessed' TM matches with the low number of hits). 

If the matches for one word/phrase are of the same type, the set priority of the resource is taken into account.

Happy Monday afternoon!


5 people like this

The Matchbar seems a great upgrade unifying resources, and removes the need for constantly docking the MT (and possibly other tabs).  It's very neat for mouse users, but it's also nice to see a good shortcut (Shift x2 + letter) for those of us relying mostly on the keyboard.

Thank you also for looking into MariaDB! Do the additonnal instructions written in the comments of Update #19 still apply?

I have updated this article on MySQL database connection with MariaDB information.

Here are some pictures of the latest build:


Notice the arrangement of the 4 resources.

Here after pressing SHIFT twice and pressing 'B':


And this is the arrangement after a restart:


Notice that the first two aren't aligned correctly: they are joined horizontally now. Unexpected and undesired behaviour.

Notice the blue label "Stem term". However, the colour indicator in the Source segment pane (namely orange) is the same for exact and stem terms. Perhaps this can be optimised in a future build too?

Perhaps the blue label 'Assembled' could be changed to 'Context Match', since that's more specific?

I have successfully created a MariaDB database for CafeTran and set up CT to use it as per the instructions.

A new TR memory table has been succesfully created since, so apparently all's good.

Will test in the next days and report back if needed! Thanks again.

This Matchboard and the colours for different resources  was something I miss in CafeTran. Great new feature, thanks Igor. 


One more thing: for CafeTran consistency reasons, shouldn't the brown table header in the glossary pane match the orange of the Source segment pane? Green is TM match, orange is glossary match.
Very good move in the development of the software.

However, is there any chance to number the results of the matchboard so that they can be inserted with e.g. ALT 1 and seq. (just like in the case of F 1 and F 2 window)

In my opinion the blue colour indicating the source of result is definitely too bright and all the colourful boxes are too wide at the expense of content of actual results (especially in vertical orientation)




Alt on Windows and Mac too? Not sure if that's possible.

The new matching algo is like this:


In previous builds, there was a switch to let glossary entries have priority over TM fragments. Looks like this switch has been removed (or I'm getting blind).

Anyway, I'm having this (for me) unwanted behaviour currently:

TM contains a wrong translation (singular instead of plural), but nevertheless this fragment is being auto-assembled:


Matchboard (which BTW is a great way to spot 'who's responsible for a certain AA result'):


I'd rather have the correct plural, present as an exact match in my Active Glossary with priority set to Medium, used for auto-assembling (glossary takes priority over TM fragments).

Is this still possible, Igor?

Should/could/would this:

- Exact fragments and terms

be changed to:

- Exact terms

- Exact fragments

Duh, the Exact term is already listed above the Exact fragment in Matchboard, doesn't that implicate that it should have been used in AA?

Is this still possible, Igor?


"If the matches for one word/phrase are of the same type, the set priority of the resource is taken into account."

So just lower the priority of your translation memory.

1 person likes this

Very nice. Thanks!

IK: So just lower the priority of your translation memory.

Settings are everything.


Login to post a comment