When I use Task > Insert all EMs these CMs and EMs are inserted (as should). However, they are not skipped when I use Command+Alt+Down.
CT propagates only the identical segments both textually and tag-wise. There may be a tag difference so the program can neither auto-propagate nor skip even though the text appears the same. I am working on the version where the user can have a better control on tags, that is, see (optionally) both the boundary tags and adjacent tags.
Exact matching is when there is 100% match in your TM and then such a TM match is auto-inserted. Propagation is about identical source segments in the source document (TM is not involved at all).
Here it is, the message that I've sent:
True. But the translation of segment 123 is fed to the TM. Arriving at 129 it should have been inserted to this inserted sent.
When I pause and use Insert all EMs, 129 is inserted from the TM. It is there when I return from segment 1 to 128.
However, when I use Cmd+Alt+Down, CT doesn't jump over 129.
This only happens in these Help&Manual mqxliff projects.
This is such a specific combination that I'm afraid that nobody can confirm this.
CT does not jump over a segment when it detects a tags inconsistency to let the user correct it. The new build will let you inspect the segment boundary tags and adjacent tags to see why exactly it cannot jump over.
I wanted to try out Exact Matches, so I started a new project, after I had just finished exporting it. I used the EXACT same document I had used initially.
It only did about 50% of the Exact Matches, but I expected it would do 100%. Why?
The ones it didn't do, do have tags. However, they were the same tags as I had just dealt with, nothing different, since it was the same document.
Any ideas what I need to do? I want to get the most out of exact matching, obviously.
Was "Processing tags" checked in the TMX Memory you used? [You can verify that with a Right-click>Options on the TM]
If not, you can open initial project, create a new TM with this option ON and try again.
It is a good option to have enabled for Project memories. Just a side note: Total Recall does not store those.
Thanks for your advice. Changing my BIG MAMA to "process tags" worked like a charm!
Where do I learn what the standard features I need are? I hate bothering people with all these questions to just get started.
Is there any reason I wouldn't want "processing tags" clicked? I guess, I'm wondering why the choice, if you know.
So, I want BIG MAMA to have Processing Tags clicked, then, right? I didn't. Now, I do. Is that your advice?
When I go into the original translation, and I click on PROJECT CONFIGURATION. Neither ProjectTM or Project Glossary are clicked, although I am very confident that I clicked them when I started this project. And the projects folder has a project TMX. Any idea why the project configuration would have them unclicked (after I've exported at least)? Is there ever a situation that I wouldn't want projectTM or glossary? Again, wondering why I have the choice. When I click on preferences for the ProjectTM, it does have processing tags clicked...would you recommend that?
Thanks again for your advice, Jean!
CafeTran does not force a specific workflow upon the translator, it can be used in various ways, when it comes to TMs (and glossaries). This strength can sometimes become daunting when trying to figure out which workflow to use.
In a very simple/default workflow, ProjectTM and Project Glossary are ticked with every new project. This creates those resources within the Project folder that is populated when you create a new native Project. By native project, I mean one where CafeTran handles the original file format.
I don’t tick Project TM and Project Glossary (and maybe they get unticked on their own) when I work on external bilingual files or packages that come from other CAT tools. In that mode, CafeTran does NOT create a Project folder. It edits the bilingual file at its original location.
What I do is simply create a TM (that I simply call Memory) and eventually a Glossary (that I call Glossary) after opening the external bilingual file. To make things simple, I save those in the same location with the bilingual file.
For the Big Mama approach, I think you should check what Hans (woorden) has to say or comment: http://cafetran4mac.blogspot.fr/2015/04/the-big-mama-approach.html and http://cafetran4mac.blogspot.fr/2015/03/tmx-files-approach.html
I use a modified workflow that suits me well:
- Systematically create a Project TM (in native and external projects as described above). In this TM, I usually enable Processing tags. Why? Because they
After completing each project, I copy paste the ProjectTM in a folder that aggregates all my TMs for this language pair. I rename the ProjectTM using a naming convention like “20171027 client project language-pair.tmx”. This helps keep track of where each TM comes from.
This folder can be used to create a Big Mama anytime: In a Project, go to Memory>Open memories folder and select the said folder. Then, you can save the TMX.
- Instead of the Big Mama TM, I use Total Recall, which aggregates all my TMs. Each time I finish a project, I import the TMX to the Total Recall table for that language pair. The Total Recall database is ticked by default, and a TMX is created from it (you can tweak its options in the Dashboard) on the fly. Note that this TMX does NOT have Processing tags, since tags are removed when imported in Total Recall.
- I also use a Big Papa TM, for fragments (see below).
Here is what I think about Processing tags: it is a good option for a ProjectTM, not so much for a Big Mama. Why? Because Processing tags is generally applicable on a document level. Stripping the tags provides a more universal reusability, as it can apply to segments with same text and similar tags. You win some and you loose some.
If you save Processing tags in the ProjectTM, you can always create an ad hoc Big Mama (Memory>Open memories folder) with that option ON, so that your resulting TMX is ready with the tags in place.
Now glossaries, termbases are another chapter. I won’t go into too many details today, but be sure to check the Big Papa approach by woorden.
My take: I use a Big Papa TM and attach it in all projects, only enabled for Fragments. Basically what woorden suggests.
Then I use a glossary for importing or exporting terms from client glossaries, or for terms applicable for a specific project. The project Glossary can also be used for (disabling auto-assembling in all other resources and) fixing the MT output with the option to Team auto-assembling with machine translation (but this can also be achieved with termbases = TM for Fragments, fragments being words, expressions, etc.). But this would require to talk about auto-assembling… not today!
This forum has many posts which can give interesting insights about how to use CafeTran. Tooltips, the Solutions (knowledge base) section of this site, some user created manuals, etc., can also be helpful.
CafeTran is in constant development and it is difficult to keep track of all features and changes, a manual would require much effort to be kept up to date. A complete, official manual is something that is lacking.
I thought of creating a resource that outlines various aspects of a CafeTran workflow, but it will take some time to finish as I just started.
In the meantime, ask away, as you see there are many active users eager to assist, along with Igor of course!
I hope this helps.
Thank you so much for your time in doing this. I will try to put this information into practice and find my own best workflow, as you suggest. Take care!
Jean: I thought of creating a resource that outlines various aspects of a CafeTran workflow
You make me very curious.
Hans CafeTran Wiki
As per subject line.
Why is this?