Sentence patterns coming into CafeTran


The next update of CafeTran will see a new enhancement to auto-assembling called "Sentence patterns". It will allow translators to create translation memory segments with variables in the following form:

All the leaves are {1} and the sky is {2}. = Todas las hojas son color {1} y el cielo es {2}.

Then CafeTran will be able to use terms in glossaries or fragments in translation memories to replace the variables with the found entry, creating a complete translation.

The feature will also let the user set the default exact match for the pattern such as:

All the leaves are {1=brown} and the sky is {2=gray}. = Todas las hojas son color {1=caféy el cielo es {2=gris}.

The order of variables is not fixed so the function may be really useful in auto-assembling when the translation of a sentence has the variable lexical elements in a completely different order.

I call this new improvement "Sentence patterns" as suggested by a user but it would be interesting to know an alternative (or perhaps a standard) term for it.


4 people like this

I'm dreaming.

What's the difference with subsegment matching?


>What's the difference with subsegment matching?

I think: the fact that you can 'hard code' the content to be inserted.

@Igor: Sounds like a very nice feature!

Hans Lenting

Okay, so what's the difference with a termbase entry like:

today annonced, that

(used in just about every press release)


Auto-assembling with subsegment matching alone builds the target sentence keeping the order of subsegments found in the source sentence. Sentence patterns give a translator a finer control over autoassembling as she can define precisely her own subsegments order and placing in the target segment.


OK, but in this example, what would be the diffrence between Sentence Patterns, and termbase entires like:

all the leaves are

and the sky is

It looks like a novilty that would fit perfectly in my workflow, except that I don't understand it.


Wait a minute, the translator (who is a "he" in my case, and politically inccorect at that) doesn't have to do anything to achieve the result. Is that the bonus?


Having one sentence pattern, CafeTran will create the correct target translation no matter what stands for the variable part in the source sentence. So the colors of the leaves and the skies in our example can be changed in the source sentence, and yes, the translator does not have to do anything if he has a glossary or translation memory of colors.


But, but, do you have to mark the sentence pattern in anyway? And if so, again, what's the difference with adding fragments to your termbase? If not, it's subsegment matching, the fuzzy way?

Mama Cass. Dead. Dammit. What a voice. Had to listen to a number of songs because of this.


You have to mark it if wish to point precisely where in the target segment, the variable part/fragment should go. As I said, the patterns give a translator a full control over auto-assembling. The translator steers the placement of the fragments in the target sentence.

Monday, Monday :)


IK: Monday, Monday :)

It is. But it never rains.

But I still don't get it. How do you mark it? {1}? Where, in the TM? How do you do that?


You will be able to mark it as shown in the announcement. The patterns can be set either in txt glossary files or TMX memories. I will provide the full description of the feature in the Solutions section along with the release of the update.




Sound like a great feature - and it's also great that you can use it with both TMX files and TXT glossaries.

MBr: Sound like a great feature

But, but, you don't even use AA, do you? Because you concentrate on legalese...


Login to post a comment