Start a new topic

Using the same TM with various projects

Up until recently my work has been so varied that I have not really had much use for TMs. However, I am now doing a batch of files with a lot of repetitions. The problem is that when I finish one file Cafetran does not offer me any help at all with the following file other than the usual suggested (and usually very useful) machine translation.

Is this because I am not understanding a vital piece of Cafetran architecture? I have always clicked on New Project every time I start a new file. Should I be doing something else? Is there a new file within this project procedure I do not know how to access. 

Grateful for any pointers you might offer,



Welcome to CafeTran!

If you have all the files that you want to translate available at the time that you start a new project, you should copy these files to one folder and load this folder via the checkbox Folder in the Dashboard.

If you receive these files by and by, you can add them to your project by copying them to your project folder and add them via Project > Add Document.

In both cases, you should keep using the same TM, either the automatically created ProjectTM or any other, manually created TM.

Berni: Up until recently my work has been so varied that I have not really had much use for TMs

You may be mistaken here. TMs are always useful, if only to look up terms you encountered before. Please see here and here for more information.


Thanks. I knew it must be possible, but I didn't know how and as usual work takes preference over a learning curve for how to use your software ;)

@Woorden, your links would be great if I had a mac. Can you use the same principle on the PC to create a TM over several projects?

I used to use Metatexis as my main CAT Tool and I really miss MT's easy way to choose which TM you wish to use. Again, there probably is a way for me to select a TM when opening a new project in CT, but I haven't discovered it yet. CT may be user friendly for techies, but from the programme's point of view I am not a friendly user ;)

CafeTran is very flexible in the TM department, it lets you choose the TM approach you prefer or change it ad hoc for a given project.

woorden's suggestion for a TM setup has been a great inspiration. Thank you woorden! I'm on Linux, but the method he describes works fine on (or can be adapted to) any platform, and CafeTran works much the same on Windows/Mac and Linux.

You can you the "Big mama approach" to aggregate segments from several projects.

I don't know how Metatexis handles TMs, how about you describe what worked for you and see if we can help you replicate or figure out a method that suits you even better in CafeTran?

Having at least the Project TM active helps not only with internal repetitions but also for consistency and concordance.

If you create a TM that you wish to use on several projects, you can tick it (or add it by drag and drop on the dashboard or though the menu if already started) when creating the next project you want to use it on. 

You can also rename your project TMs in a way that will let you know to which project they belonged and keep them in a TM folder by language pair and other criteria, then open all the contained TMs as a single TM (this is done in the Dashboard, by ticking the option and adding the folder or in Memory menu).

You can also specify the Project name, Client name, Subject field and other metadata in the Project configuration (or add such metadata afterwards, CafeTran let's you do maintainance work on TMX files), so that you can even filter for example only TMX segments related to a specific field, client or project.

You can use Total Recall to create building very big TM tables and reuse them easily.

TMs come with options, and those options are important to understand in order to make TMs work best for you. woorden's pages give very good hints on what settings to mind.

What do you find difficult in choosing the TMs you want to use?

What CafeTran does not necessarily do is organize your files for you. All project TMs have the same name for instance, so you have to figure out a way to make your TMs quickly findable in a File manager kind of way if you want to reuse the Project TMs.

Just some thoughts from a friendly user

@Idim - Thanks for the long series of suggestions and help. I'll reply in depth when I finish the current project. :)

Better late than never.

Here is my description of how Metatexis handles the issue of TMs.

When you start MT the easy way to do things is allow the Wizard to help you through the process (you can skip it after you get used to the procedure and use specific orders. But the Wizard is really useful).

The Wizard asks you if you wish to translate the current text (open in word) with the help of MT. You reply that you do. It then asks you to define the source and target languages.

After that it asks you to either open an existing TM or create a new one. In the creation process you can give your TMs easily identifiable names. So, for example, as my Main working pair is Catalan to English - my Metatexis TMs had names like CAT2ENG_GEN (for general stuff) CAT2ENG_ART for the specialised art work I often get, etc.  Easy to remember and easy to call up from the existing TM list.

TM also offers you the opportunity to call on two TMs - useful when general and specialist texts meet in a semi-specialist text.

Finally you are asked to name the resulting text that you will be saving.  (I think that process also involves defining the path to where it can be found too).

You can also define at this stage or later, whether you ant the programme to treat everything between full stops as a translation unit, or between full stops or semi colons or colons and full stops etc.  There are exceptions you can choose too at this point.

That's it. You are ready to go. From that point on you are presented with the standard two boxes showing the original and the translation. If there is anything of use in the TM that will be offered at this point.

What I miss in CT is the flexibility of naming TMs, calling up existing TMs, etc. There probably is a way to do this. But I have yet to discover it. Though following the advice above I did manage to use the same TM for the entire project, which saved me a lot of time and repetition. So thank you for that people!

In CafeTran, say you translate a Word file (not an external bilingual file).

- Choose the language pair on the Dashboard.

- Quickly select/import your resources: (TMs, Glossary, Web researches, etc.), it’s just a matter of ticking/unticking, dragging and dropping, or simply choosing the Project template (per project, language pair, client, with all the appropriate resources already selected) *Project templates are a new feature, be sure to check it* 

- Drag and drop the file

- Give a meaningful name to your project (organization)

- If you want to change the Segmentation, before continuing, click on Preferences and change it then.

- Launch the project.

Nice and easy. Flexible.

My suggestion for TMs. Use more than one, with specific settings each.

- Use a Project TM for each project (this is the default behavior). 

When you have finished the project,  you can rename ProjectTM to the project name (date, etc. example: "20170525 client project EN-FR.tmx" and copy it to a specific folder where you store your TMs (per language pair, subject, clients, etc, it’s your call). For example, create two folders, one for general stuff and one for art.

- Reuse your previous Translation Memories:

Option A: Use the Big Mama approach (see woorden’s explanations for settings and the specifics).

If you want, you can create two, one for general and one for art: just tick which you want to use in the Dashboard and voila.

Option B: Create an ad hoc Big Mama by opening your general and/or art TM folder AS A FOLDER (I explained how in previous posts), which opens all the TMs stored in it as one big TM. Do so in Read only mode to avoid mixing.

Want to create a Big mama of past work? Open TMs as a folder, save memory as... and use that memory.

Option C (esp. if you have many TUs): create one or more Total Recall translation tables and at the end of each project, simply import the TM to the Total Recall TM. Then, it’s just a matter of loading the resulting TMX from the Total Recall database when you create a new project. A little more advanced to set up, so leave it out for now. 

Thanks for the detailed guidance idim. When we have finished the exam period at the University where I work I will put your advice into practice when I do my next translations.

Have a good summer!

Login to post a comment