Start a new topic

Q: Is there any documentation for Preferences > Memory ?

 Hello dear CafeTran Users,


I'm would like more information about the Preferences > Memory options. Some entires are self-explaining, but with others you don't know exactly how the program will behave when changing them.

I have changed the 'Java Memory Size' to 6144 because I have 8GB memory and want to ensure fast processing, and it IS fast with this, no doubt.


The program behavior I would like to achieve by changing these values (if possible) is:

I would like CafeTran to concentrate on subsegments of 2-words and more when searching and displaying fragment hits / fragments with context alone and skipping 1-word expressions. Also I noticed that not every 2-word expression will be displayed, even when contained in the memories (yes I have checked the Fragment checkbox and the 'Fuzzy & Hits' choice), for example if a sentence in the memory contains the expression 'online ressources', CT will display a match for 'online' and one for 'ressources'; but I would like CT to skip such 1-words and display 'online ressources', even if it doesn't find the translation for it, so he may display the segments 1 | 2 | 3 etc.. with the context containing 'online ressources'.


Is this possible, and is there some documentation explaiining these options?


Thanks very much,

Mike


Mike: I'm would like more information about the Preferences > Memory options.


I wrote an article about my preferences.


...but I would like CT to skip such 1-words and display 'online ressources', even if it doesn't find the translation for it


I don't think this can be achieved by fine-tuning the settings. In AA, CT always inserts (if enabled) the longest match, even if the resource doesn't have the highest priority assigned to it, but if the word are non-sequential, they will be displayed separately. For that reason, you should add as many fragments as you can to your Memory for Terms/Fragments. And not to a TXT glossary, since in that case, you won't benefit from any hits of the individual words in the glossary.


H.

Hi Mike,


 Also I noticed that not every 2-word expression will be displayed,


Try lowering Edit > Preferences > Memory tab > Subsegment hits threshold (from default 3 to 1). Then, CT should display the longer hits with even one occurrence.


Igor    


Thank you all very much, this is all very helpful for me.
@woorden,
thank you for your article reference, I won't miss reading it.
When I said above 'concentrate on 2-words and more', I of course mean words that come in sequence, if not in sequence they generally would be of little help (in the past I wrote a lot of posts about this in the Trados forums to point out why this priority in the search is so important and useful, so they would correct their 'Auto-Concordance' feature, but to no avail.)

Thanks very much Igor, your suggestion works perfectly, I now get indeed all 2-word strings and longer terms,  and also the longest possible string (as long as they are in the memories of course). Also I notice that the more occurences of such strings I type in my translations, the better CT will 'guess' the translation, a simple look in the matching results is then enough because when typing it will autocomplete.

And about auto-typing:
I notice that when typing a word sequence the second time it's often already in the auto-typie suggestion. Very good and time-saving!

I haven't worked with CT for long yet, but that's going to change because I am very impressed by the whole design of CafeTran when it comes to matching results and the fast way to insert/use them in your target segment, it makes working very efficient and a pleasure.

@CafeTran Training,
than you for your link and also for your recent response in the Proz forum. Yes, I'm aware that CT can handle and import TMX files. I will soon look deeper at these workflows.

Greetings,
Mike

 

Login to post a comment