Start a new topic

How does Total Recall work?

I'm having troubles to understand how Total Recall works.


I've created one big TMX file from all TMX files that I have translated with CafeTran.


I've created a TR DB and imported the TMX.


I create a new project from a document that I've already translated (and the TMX for that project is included in the big TMX file, loaded in a TR table.


When the project is created, a preliminary matching takes place. When this is finished, I use Translation > Insert all EMs. Nothing happens.


When I use the Find dialogue box to search for any part of any source segment, scope set to TR, no results are returned.


I must admit: I understand nada.


For clarity and to increase the chance of matching in new projects, Total Recall removes tags from segments. So you will get exact and fuzzy matches in real time as you translate from segment to segment. The "Insert all exact matches"  command inserts only perfect matches when the program 'knows' tags positions in tagged segments. Currently, tags positions are stored in TMX files only. If you wish to re-translate the same document as you did in your test, please use the TMX memory (e.g. ProjectTM.tmx) that you created translating that document before.


As for the search, you should have set the scope for TM source/target since you search for the segments recalled to the working memory. 

Thanks for the confirmation.


I think that the merged TMX file is corrupt (I took the liberty to include some legacy TMX files from a client, I guess one of them isn't valid):


image



So I deleted the table, via the DELETE button:


image



The DB still takes 95 MB, I guess that I can simply delete this file now?


image



I'll start creating a new table and import only CafeTran generated TMX files, soon.

>For clarity and to increase the chance of matching in new projects, Total Recall removes tags from segments.


So, does CafeTran actually store tags? I always thought it only stores tag positions, in properties.

Can it be that this info is outdated?


image



No such tab on my CafeTran/Mac. The active DB driver is:


image


Can't remember if I changed it.

Created a new table. Loaded one already translated project including the corresponding TM. When I try to import it to the new table, I get:


image


Am I doing something wrong?

> Am I doing something wrong?


I don't know. You could submit a support ticket attaching a TMX memory file that you are trying to import into Total Recall.

It is a TM created by CafeTran. First I'll reboot the computer and try again. I take it that there's nothing wrong with the settings of the SQLite DB. Studio strange thing is that I was able to import the merged but obviously corrupt TMX but not to import a perfectly valid 13 K TU native CafeTran TMX file. Strange things do happen on this planet.

Hans Lenting: The DB still takes 95 MB, I guess that I can simply delete this file now?


No.


H.

If you delete that .db file manually, CafeTran will create a new one after restart.

IK: If you delete that .db file manually, CafeTran will create a new one after restart.


I know. But why would you do that?


H.

I know. But why would you do that?


Hans L. seems to have loaded some corrupted TM to the database. Instead of looking for those segments, he may just wish to start from scratch by deleting the database file.

Just delete the table(s) rather than the DB, I'd say. I deleted the DB with three databases once, but that was because I copulated the DB up. All by myself.


H.

I've created a ticket and attached a TM. What I've found out in the meantime: the dash in names (like all_de-nl) is represented as an underscore (the table appears as all_de_nl). I do hope that this isn't causing my issues.

Just noticed that a table name like 'all' isn't accepted. Perhaps because this is a reserved SQL instruction?

Login to post a comment