Generally speaking, when a client asks for an "uncleaned"
Trados file, he is asking for a file that contains the original
and translated text in the same file – a bilingual file.
An uncleaned file is generated using Word (".rtf" or
".doc") or TagEditor (".bif" or ".ttx").
An "uncleaned" rtf or doc file looks like this:
{0>You
will have the ability to electronically send a complete claims
review file (damage assessment report and digital photo images)
to your insurer.<}0{>A
oficina terá a capacidade de enviar eletronicamente
o relatório completo de análise do sinistro
(relatório de avaliação de danos e imagens/fotos
digitais) à seguradora.<0}
{0>This
one-time procedure gives the insurer everything they need
to quickly settle the claim and let you begin the repair job
sooner.<}0{>Este
procedimento único fornece à seguradora tudo
o que ela precisa para processar o sinistro e para que a oficina
comece a trabalhar o mais cedo possível.<0}{0>And
insurance companies will want to do more business with you
because you make their businesses even more efficient.<}0{>E
as seguradoras desejarão trabalhar mais frequentemente
com sua oficina, pois ela torna o trabalho da seguradora mais
eficiente.<0}
Déjà Vu can import pre-segmented Trados files,
so that you can translate it and get "uncleaned" Trados
files. As you can see below, a pre-segmented file has the same
format as an "unclean" Trados file. The only difference
is that the file does not contain a second language yet.
{0>You
will have the ability to electronically send a complete claims
review file (damage assessment report and digital photo images)
to your insurer.<}0{>You
will have the ability to electronically send a complete claims
review file (damage assessment report and digital photo images)
to your insurer.<0}{0>This
one-time procedure gives the insurer everything they need
to quickly settle the claim and let you begin the repair job
sooner.<}0{>This
one-time procedure gives the insurer everything they need
to quickly settle the claim and let you begin the repair job
sooner.<0}{0>And
insurance companies will want to do more business with you
because you make their businesses even more efficient.<}0{>And
insurance companies will want to do more business with you
because you make their businesses even more efficient.<0}
Colors are set inside Trados. In this
case, source text is blue; target text is green.
Codes ( {0>;
<}0{>; and <0}
) and source text will be formatted as hidden.
For Déjà Vu, the important code is
<}0{>. In this code:
0 means "No Match" –
Target text will not be imported into Déjà
Vu.
1-99 means "Fuzzy Match"
— Target text will be imported into Déjà
Vu as fuzzy
100 means "Full Match"
— Target text will be imported into Déjà
Vu as matches
In
most cases, once you have a pre-segmented file (rtf, doc, bif
or ttf), all you need to do is to import it into Déjà
Vu using the appropriate filter. See "How to import a Trados
file into Déjà Vu."
When
importing a translation for editing, it is necessary to change
Trados codes. To know how, see "How
to change Trados codes."
Older versions of TagEditor work with a bilingual file with the
extension ".bif". Déjà Vu X can import
bif and ttx files (even though its bif filter has some problems),
but Déjà Vu 3 cannot import ttx files. To work with
ttx files in DV3, see "Translating
TagEditor Files (bif/ttx) in DV3."
As a curiosity, both have a structure similar to "tmx"
files.
4
- I normally choose 100%, so I do no get fuzzy matches.
5
- Click Translate.
6
- Trados will generate a pair of files for each file pre-segmented.
One will have the extension ".bak" and the other will
have the extension "doc," "rtf", "bif"
(if your Trados is v 5.0 or older), or "ttf."
TIP
If you have received a
pre-translated Trados file from your client, there
is
a possibility that the file was not fully
pre-segmented. If this is the case, segment your file
as explained
above.
Trados will segment non-segmented text, but will not
modify what is already translated/segmented.
If
your Trados file is not totally imported, this is probably because
the file is not totally segmented. You should pass files through
the pre-segmenting process again. Segment your Trados file again;
see "Segmenting Trados
Files." Be sure to select Segment Unknown Sentences,
Don'tUpdate Changed Translations and Don't Translate Terms. This second pass will segment unsegmented
units and will NOT change anything that is already segmented.
After the second pass, import, translate, and export files normally.
See Importing Trados Files into
DV files.
Pre-segment your file
using Trados Workbench. It is possible to use the demo version
for this. If you have ".bif" files, skip the next
and last bullet.
Open TagEditor and save TTX files as BIF (in the TagEditor's
Save as window, just name the file <filename>.bif.) —
This can be done using Trados' Demo Version.
WARNING:
It is necessary to open TagEditor,
open your segmented TTX file, and (File>)Save As BIF.
If you just change the name (e.g., in Windows Explorer),
it will not work.
To save it as BIF, you have
to do it manually; it is necessary to type "bif"
(replacing ttx) in the file name box.
Create a TMX project in DV3, specifying BIF in the Source
file extensions box (Project Configuration window), then import
your files.
Translate.
Export as UTF-8
Open BIF file in TagEditor and Save it as TTX
There is a script that converts all TTX files inside a folder
in BIF files. The script can be found in the Files section of
the TW_users group. It is named SaveAsBIF.vbs.
You only
want to change Trados codes if you want to import another
translator's work. That is, if somebody translated the file
using Trados and you need to edit it in Déjà
Vu. If files were pre-translated or pre-segmented and no
one has worked on them, you can skip these
instructions. If the translator formatted target
text differently from source, you'll have problems exporting
it (e.g.: words in italic in the target language
where they were not in italic in the source language).
To
edit Trados files in Déjà Vu it is necessary to
change Trados codes ( <}0{> ).
When importing your file into Déjà Vu,
0 means "No Match" –
Target text will not be imported into Déjà Vu.
1-99 means "Fuzzy Match"
— Target text will be imported into Déjà
Vu as Magenta
100 means "Full Match"
— Target text will be imported into Déjà
Vu as Green
I recommend to replace 0 codes with
something >0. I generally replace them with 10. See "How
to change Trados code."
Once this is done, create a Trados project, import files, and
work normally. Export finished edited file normally. Sometimes
there are some formatting problems in the target language —
check the final document carefully.
First get the Trados TM to be used
if the agency has one already in place and wants you to take
advantage (or disadvantage:-))) of it.
Then open Workbench, and open the
received Trados TM. If you received a Trados TM in text format,
create a new Trados TM and import the TXT TM received into
that TM. If you didn't receive any Trados TM, simply create
a new and obviously empty Trados TM. (And keep it somewhere
for future similar projects.)
With Workbench open, in Translation
Options, tick **Copy Source to Target**.
Click on Tools>Translation. Make
absolutely sure to tick **Segment also unknown sentences**.
Then click on Add, add all the files that you want to segment,
and finally click on Translate. At this point all your added
files will be segmented by Trados (following Trados rules
for segmentation, of course).
Close the Translation window and
close Workbench. Open Word if you like (it's not mandatory)
and see that your files are indeed segmented. See also that
Workbench created a .BAK file for each of the segmented files
that is simply the original file with a different extension.
Now open DVX, create a project,
and go all the way (do not use the Import Wizard) until you
arrive to the Project Explorer window. Locate all your present
files to translate and drag them to the lower right pane.
Click on each the dragged files in turn, and you will see
a Properties window to the right. Start clicking on Filter
until you see appear Trados Workbench to the right of it.
Repeat this procedure for each of the Trados files to be
imported.
Finally, import all your files.
Just in case, import them one by one so if there is some trouble
with a file you can easily pinpoint it. You will see that
DVX will show you in the target cells all sentences (if any)
already present in the Trados TM. All other target cells will
be empty.
Translate normally.
Proofread and export.
Be happy.
If you need to join/split segments
in a Trados project, see my recent message
to Marco about this feature.
The TWB filter for DVX has always been different from the one
that DV3 used and unfortunately Office 97 doesn't support the
new structure of the code. So you can't import/export TWB files
with DVX if you only have Office 97.
A native Trados memory is a set of five files (extensions: tmw,
mwf, mtd, mdf, iix). It is not possible to work in DV using those
files. Trados can export its memory to two formats: txt and tmx.
These are the formats that can be imported into Déjà
Vu. If you do not have a Trados license, you should ask your client
for a txt or tmx Trados exported file.
Go to File > Open and open the memory you want to export.
In the Open Translation Memory window, select "Exclusive"
(bottom of the window).
Select File > Export
Select the first field you want to export in the "Field
List" and click the "Add" button.
Select the second field and click "Add", and so
on.
You should select at least the Source and Target fields of your
memory (in the window above, English (United States) and Portuguese
(Brazil).
Click "OK."
Select if you want to export as TMX
or TXT, select the directory where you want to place the exported
file, and click on "Save."
To import a Trados Memory (TWB) into Déjà
Vu, you should first export it from Trados as TMX or TXT (see
above). You cannot do anything with Trados memories while they
are in their original format (*.tmw, *.mwf, *.mtf, *.mdf, and
*.iix)
Trados exported memory (".txt" or
".tmx") contains codes you do not want to import
into Déjà Vu. To delete them:
Download
this macro and install it in Word. This macro will
clean most of the codes
Open your ".txt" or ".tmx"
file in Word and run the macro above
Save the file as ".txt" or ".tmx"
(and not ".doc")
If DV3, open Déjà Vu 3's Database Maintenance
module.
Open your memory.
If exported as TMX: Go to File>Import, select
TMX and be sure language codes are in the form XX-XX
(and not XX_XX). If necessary, type language codes inside
the appropriate fields. Those fields will accept typing.
If exported as TXT: Go to File>Import and select
"Trados WB database."
NOTE: Trados memory
is a txt file, but will NOT be imported into
DV as txt, but as Trados WB Database.
If your Trados memory was exported from Trados 7, when importing
it into DVX, choose UNICODE (UTF-8) instead of the default (Western
European - Windows). In this case, it is not necessary to clean
the memory.
This is a very handy aplication that allows you to open a TTX
file inside Word and to save a bilingual RTF file as TTX: TTXpress.
Additionally, converting a ttx file in an RTF allows people using
DV3 to translate those files, instead of going through the process
of saving them as ".bif"
files.
Years ago, there was this application called ITP FILTER. Probably
some old users of Trados still have it. This filter allows to
save many file formats (html, Ventura, FrameMaker, etc.) as Trados
RTF files.
Joining
cells in a Trados TTX Project (DVX 7.0.238) Jorge Gorín
Since
DVX 7.0.266, it is possible to join and split cells freely
in Trados TTX files.
Do thoroughly what I do and you'll be systematically successful:
Do your join/split in the DVX source segment.
Reproduce EXACTLY your join/split in the Trados source
file. This means do all the necessary edits in it (by means
of copy/cut/paste, including required segment delimiters)
as to have an exact correspondence between the cells in your
Trados project in DVX and the segments now present in your
Trados source file. (All these operations should be made in
real time. That is, immediately after you do a change in the
DVX source cell, do the corresponding change in your Trados
source file.)
Keep translating in DVX until
you arrive to the next cell where you have to perform a join/split,
and again reproduce this join/split in your Trados
source file. In case no other join/split is required, keep
translating until
you reach the end of your file in DVX.
Make sure to send the whole file
in DVX to the MDB (this is obviously
crucial, so big disclaimer here if this is not performed as
required).
Close your Trados file in Word,
and even close Word just in case.
Reimport your file into your
DVX project.
Pretranslate.
Adjust any small discrepancies,
then proofread and spellcheck.
