Salt Hat Translation Notes

Writings in Swedish only

2012-01-04T00:08:00.000+01:00

Skriva

This posting is now removed

PDF to CAT and Back

2011-01-29T11:05:00.004+01:00

PDF - Portable Document Format

The file format PDF - Portable Document Format was developed by Adobe, and is an open standard. PDF files can be read (and in some cases edited) with Adobe Acrobat reader, a free application or with Acrobat Pro. The purpose of the PDF file format is to be able to produce readable documents from whatever application or paper document in a format that can be read on all computers and all platforms, without having access to the source files or having to buy the application. This also means, that there are no documents that are produced in Acrobat as source files. All PDFs originates from other source file formats.

The problem for the translator is, that customers often sends a PDF, and they are not aware of the fact that there always exists another document in the background. In most cases it is easy to obtain the source documents from e.g. FrameMaker, InDesign, AutoCAD or Word, which of course is more adapted to the translation workflow.

But Studio is able to translate PDF directly?

Well, yes I am aware of that the advertisements tell you that. The problem is that if you use the built in PDF converter in Studio, you cannot do a proper pre-editning, which I find is essential. Do feel free to try, it might work ok in some cases. But if you find your document full of rouge tags and in-line paragraph breaks, then follow my advice in this document and pre-process your PDF.

First rule of PDF translation - Do not do it!

Never accept to translate from PDF without first putting in a certain amount of asking, begging, investigating and extorting your customer to find the source files! Tell your customer that the pre-processing and post-processing of every PDF will add 1 h to the translation bill, and still the result will be of poor quality or a text-only file.

If this fails, and you do not have any other customers to work for, then it is actually possible to work also in PDF. In some cases, there actually are good reasons why the source files cannot be obtained.

PDF conversion to text

In contrast to converting other file formats to and from your CAT tool, the conversion from PDF is an irreversible process. You will never get back the same formatting that you started from. So never promise anything else than a plain text file without pictures and formatting when working with PDF. But, it is actually possible in many cases to keep the formatting somewhat close to the source, if you know how to. There are two main types of PDF:s, graphics based and text based. If you do a scan from paper and save as PDF, you get a graphic image of the text. If you make a PDF by electronically "printing" the result from an InDesign or a Word document, you get a better format, as the PDF will still retain the text as editable. In the following I will call the types graphic PDF and text PDF.

Conversion of text PDF

For this type of PDF you will need a good converter. I have tested several, and I now use Nuance PDF Converter Professional. PDF Converter manages to keep sentences in one piece (no paragraph breaks in end of lines) and also fixes most pictures, tables and columns. I always try to save in full graphic format first, but if that does look too ugly, then I save as text only.

Conversion of graphic PDF

If the text is a picture of text, you have to use OCR. There is some kind of OCR function also in the above mentioned PDF converter, but a real OCR program like ABBYY FineReader is much better to use for this kind of documents. In some cases I have also found, that the PDF converter cannot handle a text PDF very good, then it is better to use FineReader instead. Also in FineReader you have the option to try saving as fully formatted or plain text Word document.

Pre-editing of PDF

After the conversion, before translating a PDF, there is always some pre-editing to do. I have found it most time-saving to do as much pre-editing as possible before translation. If you save as plain text, some of the points below can of course be skipped.

1) Set Word to "Show invisibles" and look for misplaced paragraph breaks in the middle of sentences. Also check for new line characters and change them to paragraph or remove them. This will avoid having sentences split in several segments. I also recommend to search and replace soft hyphens with nothing. That reduces the rouge tagging.

2) Sometimes you can also get problems with language specific diacritics, e.g. åäö or ü. In such cases, select all and set a font that contains your specific letters. Note that some special symbols might change doing this, if other fonts are used. This must be restored after translation.

If you save as formatted text:

3) Select all of the text, select Home/Font/Character spacing and do the following settings:

Scale 100 %

Distance Normal

Position Normal

Kerning remove tick

This will prevent much of the rouge tags in the CAT editor.

Post-editing PDF

It is also a good idea to proofread your document and compare with the source PDF. In most cases you will find some deviations to correct, even if you have done a good post-editing.

Filtering in the Studio editor

2011-01-28T14:31:00.000+01:00

You can use regular expressions in Studio to filter what you want to see/edit in the Editor view.

If you do not see the Filter toolbar, use View/Toolbars/Display Filter.

You will now see a field called Containing:

This field can be used for writing search criteria and/or regular expressions. Write something in the field and hit return.

Try the following:

\d Shows all lines that contains numbers or numbers + text. Text only units are excluded. Same as [0-9].

^[0-9] Shows lines starting with numbers

^([-+] ?)?[0-9]+(,[0-9]+)?$ Shows +/- integers and decimal comma numbers

^([-+] ?)?[.0-9]+(.[0-9]+)?$ Shows +/- integers, comma and dot numbers [corrected expression]

^([-+] ?)?[.0-9]+(.[0-9]+)?$ Shows also headers like 1.2.3

^\d+\d+\d+$ Shows integers only, without sign

To produce the sign ^, copy from here or use Alt94

Regular expressions in Studio

It is also possible to use regular expressions for other functions in Studio:

- Filtering text files when defining a new job/adding files

- Defining segmentation rules

- In the Search/Replace function

Tagging text files in SDL Trados Studio

2011-01-24T09:59:00.001+01:00

Regular expressions is a general method for matching strings. In Studio, there are several ways of using regular expressions to change the behaviour of the system. I will describe one case here. This relates to my previous posting

http://salthatten.blogspot.com/2011/01/cad-to-cat-and-back-introduction-this.html

about translating CAD drawings with use of the converter TranslateCAD.

Note: Learn about regex
http://www.regular-expressions.info/quickstart.html

If you open the translatable file from TranslateCad, trans1 file in a text editor, it looks something like this:

-----------------

##000019##

FIRMA OY

##000020##

1:10

##000021##

1010

##000022##

Piirustukset ovat ohjeellisia.

##000023##

-----------------

The codes with numbers ##000XX## are created by the converter, to keep track of the reimport of texts. But you do not want to have them in Studio, of several reasons. They have to be untouched; and if you use Autopropagate in Studio, they are propagated to all similar lines every time you confirm a row, which is time consuming. A drawing typically also includes a lot of standalone numbers (measurements), that shows up as integers in Studio.

However, there is a nice function in Studio that allows you to define nontranslatable units when importing in text format. In the CAD case I have chosen to hide both the code and numbers/integers, so they do not show up in the Studio editor.

Workflow for text import

The best way to work with this would probably be to create a dummy project first in the desired language combination. Then you have a place for your experiments, without changing any parameters that will affect the global settings or any ongoing projects.

1) Go to menu Project/Project settings

2) Expand file types by clicking on the preceding +

3) At the bottom of the list, choose the file type Text.

4) Click on Inline patterns.

5) Click on Add rule... and paste the following into the field Opening:

##\d+##

That will make the tags into not translatable placeholders.

6) Click on Add rule... and paste the following into the field:

^([-+] ?)?[.0-9]+(.[0-9]+)?$ [corrected expression]

Note: the ^ symbol is not found on your keyboard, so either use cut and paste from here or use the alt code Alt 94.
Note: http://usefulshortcuts.com/downloads/ALT-Codes.pdf

The second rule will remove all standalone numbers. I am actually not quite sure what this expression does, I found it on a computer geek site. But it at least removes the standalone integers. Probably it will also remove numbers preceded by +- and with commas in them.

The above example, the trans1 from TranslateCAD will now look like this:

-----------------

FIRMA OY

1:10

Piirustukset ovat ohjeellisia.

-----------------

If you still would like to see the codes and measurements while working, use the Display Filter bar. If you do not see the Filter toolbar, use View/Toolbars/Display Filter. The display field probably says All segments; change that to All content. Now you will see a number tag on all rows that contains untranslatable tags only, the Tag ID is displayed. To see the whole tag content, click on Full Tag content on the QuickInsert toolbar. Now you can see all codes and numbers as non-editable tags.

Save and reuse your settings

When you are ready with the settings, you should save your work in a project template.

Go to Project/Create project template

Give the template a file name, e.g. FinnishToSwedishCAD and save. Now you can use this template for your present and future AutoCAD projects without having to do all the settings again.

2011-01-23T16:39:00.001+01:00

CAD to CAT and Back
Introduction
This article describes a workflow for translating the text in AutoCAD drawings without buying an AutoCAD license. It is also applicable on other graphical formats that can be converted to dxf (Drawing Interchange Format). The translation tool in this example is SDL TRADOS Studio, but the method, with minor modifications, is possible to use also with other CAT (Computer Aided Translation) tools, as the translatable format is txt.
Programs needed
CAD program
AutoCAD (developed by Autodesk) produces a file format called dwg. For transfer to other programs, Autodesk also developed a format called dxf, (Drawing Interchange Format). There are also other CAD programs on the market that can edit and convert the dwg and dxf formats. In this example I use a program called progeCAD Professional that costs € 291, developed by ProgeSOFT. If you can get your customer to save the drawings in dxf format, you actually do not have to buy neither AutoCAD nor progeCAD for the translation work. But, I would strongly recommend to buy a CAD program or a program that at least can read dxf, else you are unable to see the result of your work. And, you cannot verify the conversion results.
Conversion program
The conversion to and from the translatable text format (.txt) is done with a conversion tool called TranslateCAD from TranslationToSpanish.com and actually only costs $ 29 which is a very low price considering they also have free support for the program! The converter is a standalone program and not integrated in the CAD program, thus you only need a dxf file for input.
Text Editor
As there is a glitch in the back conversion to text from SDL Trados Studio, you also need a good text editor that can convert the txt format. For this conversion I use EditPad Pro that costs 40 €.
Workflow
The workflow is based on a real case where I translated 37 dwg drawings. The workflow starts with the source files saved in a folder named DWG_source
Step 1 - Convert dwg to dxf
The converter needs a dxf for input.
Open progeCAD professional and open the files.
Save As, choose AutoCAD 2010 ASCII.dxf (binary does not work in the converter)
Save the dxf files in another folder DXF_translate
Step 2 - Convert dxf to txt
Open TranslateCAD. In the left upper pane, choose the folder DFX_translate. Select all files in the lower left pane.
Push Extract text!
You should now see a Status window that shows
The target files were successfully created.
The converter has now produced two files for every source dxf. They have the extensions -trans1.txt and -trans2.txt. The translatable file is -trans1.txt, and this can be opened in Studio or another CAT application, or actually even edited manually in the text editor. If there are a lot of files it is a good idea to copy the trans1 files to another folder for translation.
Step 2.5 - Optional: Filtering with Regular Expressions
I recommend to use filters when importing, this is explained in a separate blog posting.
Step 3 - Translate with Studio
Optional: file merging
If you have many files, which is typical in a drawings translation project, it would be nice to have all drawings in the same file. Drawings also tend to contain a lot of repetitive standard texts, thus it is easier to handle in a single file. That is easy to achieve (but optional) in Studio. Use New project, which starts the project wizard. When you come to the Add files pane, first add the files, then select all files with Ctrl+A, push Merge Files. Name the file AllDraws.
The merge function will produce a single file of all drawings. Now open the file in the Studio editor. Note that it is till is easy to see in which individual drawing you are working, as the drawing name is highlighted in the list in the Editor pane to the left of the dual editing window. To expand the Editor window, if you do not see the file names, click the double arrow.
Note that Studio changes the scale figures, like Scale 1:10 to 01:10. (Do not ask me why). This leading zero must be removed manually.
Step 4 - Save As Target
In a merged file, you cannot save as target from the dual editor. Change to the Files pane, right click the merged file, choose Batch task/Generate target translations. That produces individual files for every drawing.
Step 5 - Convert to correct text format
This step is necessary only if you use Trados Studio. That is because Studio saves back the text files as Macintosh-formatted text, which cannot be converted in TranslateCAD! It is now you need your text editor, e.g. EditPad Pro. Open the trans1 files. Click to open the Convert menu. You now see, that the To Macintosh is greyed out, which means you have a genuine Macintosh file. Now click on To Windows (CR LF). Now the To Windows should be greyed out, indicating that the file is back in Windows format. Save the file, and exit EditPad. I suppose this could be done in any proper text editor. Note: If you by mistake use the Mac format, you do not get any error message in TranslateCAD. The error shows when you try to open the dxf in progeCAD!
Step 6 - Copy the translated trans1 to the source folder
First, make a copy of the folder DXF_translate, where the source dxf, trans1 and trans2 files are. Then, copy (and replace) with the translated trans1 files from your Studio project folder. The source trans2 and the translated trans1 must be in the same folder to run the back conversion. That is why it is important to make the copy of the source folder; if anything goes wrong you must be able to start again from the source trans1 texts.
Step 7 - Create translated dxf files
Open TranslateCAD, now use the right side panes. In the upper pane, locate the DXF_translate folder. You should now see all of the -trans1.txt files. Select all files and push Re-construct. It happens that you get the message "The target file is already open by another application. Please close it and try again." and that happens even if progeCAD is closed. I think that this error has something to do with the trans1 files, and a remedy that works is to delete all the trans1 files in the folder DXF_translate and the copy them again from your Studio project folder. You also might have an error message in progeCAD next time you open it after this error, but just click it away.
The converter produces new dxf files, that are named -trans.dxf, the source dxf are not overwritten.
Step 8 - Create the translated dwg
Open progeCAD (clicking on dxf does not work) and open the -trans.dxf files. Note that you have to change file type to dxf in the file format dropdown, else you will only see dwg files. You can now check your translation and formatting in progeCAD, and then Save As... and select AutoCAD 1010.dwg (AutoCAD 2007 seem to be the default).
Editing in progeCAD:
To move the drawing in the pane, push and hold down the scroll wheel on the mouse, and then drag. The scroll wheel zooms in and out in the drawing. And, you have all CAD functions available to edit the drawings if needed.
Thanks to
Paul Filkin, SDL, for help with regular expressions and text format.
Vicente Victorica, TranslationToSpanish.com, for fast support on the Mac file problem.

PDF-konvertering

2010-10-28T02:41:00.001+02:00

Översättning av PDF-filer i CAT-program
This text is in Swedish, please use Google Translate!
Mitt CAT-program Studio kan enligt marknadsföringen klara av att "översätta PDF". Det är givetvis en grov överdrift. Testa gärna denna funktion och efter att ha konstaterat att det inte blir bra, återgå till de metoder jag anger här.
Den viktigaste regeln angående översättning av PDF är att aldrig någonsin acceptera "enbart PDF"! Det existerar nämligen inga dokument som *ursprungligen* har gjorts i PDF-format med Acrobat. Syftet med att använda PDF är att enkelt kunna hantera dokument som gjorts med komplicerade och dyra program.
Om du har bönat och bett, hotat, övertalat och väntat i veckor, men ändå inte kan få originalfilerna så finns det faktiskt vissa möjligheter att arbeta med PDF.
Det finns två typer av PDF; den bästa typen, text-PDF, är textbaserad och görs ur andra elektroniska dokument från desktop-program som InDesign eller FrameMaker, eller från Word-dokument, genom att skriva ut dokumentet som PDF. Den andra typen, bild-PDF, är inscannade pappersdokument som alltså är bilder av text. De här typerna av PDF kräver helt olika behandling och olika program för att kunna översättas. Om du vill veta vilken typ av PDF det är så prova att markera ett ord i texten eller söka efter ett ord med Acrobats sökfunktion. Om det inte går, så har du en bild-PDF. Notera att det förekommer dokument där båda typerna finns, så testa på lite olika ställen om det är ett långt dokument som du misstänker kan vara hopklippt av flera.
Konvertering av text-PDF
Jag beskriver den enklaste varianten först; för text-PDF behöver du bara ett bra konverteringsprogram. Jag har provat flera, och enligt min mening är Nuance PDF Converter Professional det bästa. Om man har lite tur kan man till och med få bilder och formatering att lite grann likna originalet! Men lova aldrig annat en "oformaterad text" om du åtar dig PDF-översättning. Allt annat är en bonus. PDF Converter klarar ganska bra att hålla ihop meningar (inga stycktetecken i meningarna) och klarar även spalter och tabeller hyfsat. Alla konverterade PDF kräver dock för- och efterbehandling och en noggrann jämförelse med originalet.
Förbehandlingen bör åtminstone bestå i att skumma igenom dokumentet med "visa P" aktiverat så att de osynliga tecknen för ny rad och nytt stycke visas. Ta bort alla stycketecken mitt i meningarna. Gå sedan igenom alla radbrytningstecken (ny rad). De måste nämligen hanteras olika; ifall de finns mitt i en mening ska de tas bort, ifall de kommer direkt efter en mening ska de ersättas med nytt stycke-tecken. Det är givetvis i sin ordning att debitera kunden 1-2 timmar per dokument för konvertering, förbehandling och efterbehandling.
Konvertering av bild-PDF
Vissa PDF-konverteringsprogram påstår sig klara alla typer av PDF, men det är tveksamt om de verkligen håller vad de lovar. För bild-PDF är det bästa att istället använda ett renodlat OCR-program (Optical Character Reading) som verkligen klarar av att tolka textbilder på ett bra sätt. OCR-programmen finns i olika prisklasser, jag har fastnat för ABBYY FineReader som inte är alltför dyrt men ändå ger ett förvånansvärt bra resultat.
Ett OCR-behandlat dokument kräver en hel del handarbete för att bli översättningsbart, även om man bara ska använda texten. Ifall du lägger in ett obehandlat OCR-dokument i ditt CAT-verktyg får du sannolikt en väldig massa onödiga taggar överallt. Det beror på att OCR-programmet försöker arbeta så noggrant som möjligt och tittar på teckenavstånd, olika storlekar och typsnitt. I vissa fall konverteras även åäö för sig och får ett annat typsnitt vilket ibland leder till att man får osynliga ordbrytningar mitt i orden. Det i sin tur gör ordräkningen i CAT-verktyget alltför stor, och eventuella hittar mot gamla TM försvagas.
Förbehandling av OCR-behandlad text
Det mest radikala är naturligtvis att spara hela dokumentet som ren text och sedan öppna som Word-dokument igen, då slipper man alla formateringsproblem, men dokumentet kan ju bli lite tråkigt, och alla bilder försvinner.
Ifall du vill försöka bevara formateringen och bilderna ungefär som i originalet kan du försöka med följande:
För Office Word 2007:
- Markera hela texten och välj Start/Tecken/Teckenavstånd, och gör följande inställningar
Skala 100 %
Avstånd Normal
Läge Normal
Kerning ta bort bocken
Ifall åäö ser annorlunda ut än de övriga bokstäverna så bör du också ställa in samma typsnitt för hela dokumentet (Arial). Det innebär nämligen att det i CAT-programmet blir en ordseparator före och efter varje åäö. Ibland syns detta inte i originalet utan kommer fram först när du lägger in texten i CAT-verktyget. Ett säkert tecken på att du har åäö-problem är om CAT-verktyget ger en avsevärt (30 %) högre ordräkning än Word. Små skillnader förekommer alltid i ordräkningen mellan Word och CAT-programmet.
Det kan givetvis ha vissa nackdelar att ha samma teckensitt i hela dokumentet, ifall man använt sig av detta för specialsymboler. Dessa måste i så fall återställas manuellt. Samma gäller om du väljer att även ha samma stilstorlek i hela dokumentet, vilket även minskar antalet taggar i CAT-verktyget.
Generell förbehandling
Även om du får dokumentet i originalformat i Word kan det finnas anledning att titta lite på det innan du lägger in dokumentet i CAT-programmet. Det händer nämligen att skribenterna inte använder sina verktyg på rätt sätt.
Vanliga konstigheter:
- Meningar har delats på rader med stycketecken eller ny rad
- Tabelluppställningar har gjorts genom att använda flera tabbar i rad i stället för att ställa in tabulatorlägen och använda en tab
Så ibland kan det vara värt att göra en genomgång i förväg. Det är heller inget större jobb att börja om, ifall du redan har lagt in dokumentet i CAT-verktyget. Det är ofta enkelt att gå in i originalet, göra lite ändringar och spara dokumentet under ett nytt namn. Då kan du lägga in dokumentet i Studio med funktionen Lägg till fil. Observera att Studio beter sig lite egendomligt när man ska lägga till filer. Du måste växla till källspråket för att lägga till, sedan köra batchen Prepare without project TM, och sedan växla tillbaka till målspråket *innan* du klickar på filen för att öppna den i editorn. Ifall du klickar på filen i källspråket får du ett mycket mystiskt felmeddelande om "wrong language culture". Språkväxlingen görs i en liten ruta längst upp till vänster i Files.

Min termhantering förr och nu - Studio SP 2

2010-03-04T11:38:00.000+01:00

This article is in Swedish only!
Try Google Translate: http://translate.google.com/

MultiTerm har alltid varit Tradosprogrammens svarta får, men jag har ändå använt det flitigt, på grund av dess enda verkligt goda egenskap; att man får upp specialtermerna i en liten ruta i Trados och att källtextens termer markeras med en röd linje ovanför, ifall de finns i termbasen. Vitsen med att använda MultiTerm har alltså varit att lägga in specialterminologi i den, inte generella ordböcker, eftersom man ju då skulle få röda streck på varje ord och en alltför lång lista att välja från. Så det du ska lägga i MultiTerm är endast specialtermer eller kundtermer. Det var ganska besvärlig i gamla Trados att lägga till termer "on the fly" och MultiTerms funktioner för redigering och tillägg av nya termer var rätt klumpiga och komplicerade. Därför har jag använt Word-baserade tabeller där jag samlat kundspecifik projektterminologi under projekten. Sedan har jag vid årets slut tagit lite tid för att sammanställa årets termskörd och suga in den i MultiTerm via importfunktionen, till en stor ordlista. Jag har heller inte använt funktionen att kopiera in termer med knapptryckningar från MT till dokumentet, även om jag vet att det går i gamla Trados Workbench också.
Med Studio är förutsättningarna annorlunda, som gjort att jag ändrat mitt arbetssätt. MultiTerm är fortfarande ett ganska svart får, men det har ändå blivit vissa förbättringar, särskilt i kommunikationen med Studio. Det är också möjligt att jag nu har lärt mig att använda verktyget bättre, och upptäckt mer av dess funktioner, mycket tack vare den certifieringskurs jag gått igenom.

Det nya är att jag nu börjat använda ordlistorna mera direkt i översättningsarbetet genom QuickInsert och AutoSuggest, och jag använder nu också termverifiering via QA-funktionen. Därför vill jag gärna ha in terminologin i MultiTerm redan under arbetets gång i pågående projekt. Det är nu möjligt på ett relativt enkelt sätt. Eftersom MultiTerm och Studio numera hanterar hur många termbaser som helst på ett smidigt sätt, så gör jag helt enkelt en kundspecifik ordlista för varje återkommande kund som jag jobbar med, och bygger upp dem under arbetets gång direkt i MultiTerm.
Grunden till de kundspecifika ordlistorna skapar jag ofta från material i mina gamla projekt för en viss kund, och sedan fortsätter jag med att lägga till och redigera termer direkt i Studiomiljön. Det är faktiskt lika enkelt, eller kanske lite enklare än att skriva tvåställiga tabeller i Word. Du markerar orden i båda språken och trycker på Alt F11.

För att termverifieringen ska fungera bra är det inte heller vettigt att använda en omfattande ordlista med 20 000 poster, utan bara specialtermerna for den aktuella kunden. Annars får du bara en väldigt lång fellista på avvikande termer, som ofta inte är relevant i sammanhanget.

Studio SP2 is now launced (Swedish review)

2010-03-03T20:26:00.000+01:00

The review of Studio SP 2 is in Swedish only!
Why not try Google Translate:
http://translate.google.com/

Recension av SDL Studio SP2
Nu har Studio SP2 lanserats! Eftersom jag haft förmånen att ta en förhandstitt på den, så kommer här en kort recension. I versionen SP2 har en del buggar har rättats till och ett antal nya funktioner har tillkommit. Det finns säkert mycket kvar att förbättra, men redan nu är Studio klart överlägset Workbench och TagEditor vad gäller funktionalitet och arbetssätt. Efter att ha arbetat med Studio några månader så skulle det kännas svårt att återgå till det gamla sättet att arbeta.

Några nyheter i SP2:
- Automatisk översättning med Google Translate har tillkommit, och är faktiskt ett utmärkt hjälpmedel.
- Det går nu att stänga av lokaliseringen av siffror

Här en lista över de tre största fördelarna och nackdelarna med SDL Studio:

Tre fördelar med Studio
- Minskat skrivmaskinsjobb, ökat sammanställande. Till detta bidrar Google Translate, en nyhet i SP2, på ett positivt sätt. Tillsammans med en omfattande AutoSuggest-ordlista som jag kompilerat ur mina gamla TM sedan 1997 skapar Studio en effektiv och trivsam arbetsmiljö.
- Side-by-Side- editorn. Som gammal Trados-fantast trodde jag aldrig att jag skulle kunna acceptera att se original/översättning sida vid sida. Men det är faktiskt ett effektivt sätt att hantera översättningar, som man snabbt vänjer sig vid. Filtrering är en ny trevlig egenskap; att kunna ta fram enbart det som är nytt i översättningen, visa alla dubbletter eller att visa enbart de rader som innehåller siffror är mycket bra.
- Flexibiliteten i kommunikationen mellan MultiTerm och Studio. Det kan ju hända att det beror på att jag äntligen lärt mig hantera MultiTerm ordentligt, men jag använder nu kundspecifika termordlistor och redigerar dem on-the-fly i en mycket större omfattning än tidigare. Det känns som om jag nu har ett betydligt bättre grepp om samordningen av terminologi än jag haft tidigare och att jag bättre kan utnyttja mina gamla ordlistor.

Tre nackdelar med Studio
- Det känns fortfarande som om Studio är en oslipad diamant. Det finns en del buggar och felmeddelanden som dyker upp i olika lägen, och det finns vissa saker som är omständliga att göra. Med SP 2 så är dock förhoppningsvis de flesta barnsjukdomarna avhjälpta.
- Tar tid att lära sig. Studio är ett kraftfullt system med många nya finurliga egenskaper, och det tar tid att lära sig allt det nya, ifall man verkligen vill utnyttja systemets fulla kapacitet. Jag har alltid hittills lärt mig ny program genom att läsa handboken och experimentera, men denna gång valde jag att gå igenom kurserna noggrant. Det finns tyvärr ingen bra handbok till Studio, och hjälpfilerna är alltför ofullständiga och kortfattade. Det kan ha förbättrats i SP 2, SDL har tagit till sig av kritiken.
- Dålig kompatibilitet med gamla TRADOS, vilket gör att det är ofrånkomligt att köra båda versionerna parallellt när det gäller vissa filer. Det går inte att bearbeta "Uncleaned word" direkt i Studio, vilket ju är vanligen förekommande i den "gamla" världen. Det uppstår också en del förluster av 100 % hits eftersom segmenteringsprinciperna är olika.

Studio vs DVX: The winner is ...

2009-12-28T13:12:00.000+01:00

After trying out both Studio and DVX, I have now decided about which tool to choose: It will be Studio.

And I will now also sign a maintenance contract for Studio; it will probably not be able to run without constant support. I will also take a webinar for Studio upgraders on 8 January, and take the certification exam.
The main reasons for my choice:
Studio +:
+ Handling of "rouge codes" does not cause any extra work
+ AutoSuggest, that is so superior to DVX AutoAssemble
+ SDL Automatic Translation Server, which actually boost productivity
+ Dual editor has a better user interface
+ TW_users list is a good free source for support
Studio -:
- Multiple translations problem
- Instability and error messages
- Bad concordance search compared with Workbench
- Numbers and codes "localization" cannot be turned off
DVX +:
+ Intuitive and fast on the fly editing in all databases
+ The AutoAssemble and Lexicon function is very good in some cases
DVX -:
- Multiple translations problem
- Tags handling can become overwhelming in some Word documents
- Déjà vu users lists is only for DVX fanatics
It is difficult to compare the two systems function by function, as they are working so differently. Actually, you could have use of both AutoSuggest/AutoAssemble in the same system, and that also goes for the AutoSuggest dictionary/Lexicon as they could complement each other.

Multiple translations and duplicates

2009-12-22T09:46:00.000+01:00

I am going to explain this problem and also describe how it is handled in three main CAT tools. First, some definitions, as these terms are used differently, and it is easy to misunderstand the issues:
Multiple translations: Same source segment translated differently in TM/MDB
Duplicate: Same source segment in the document
File duplicate: Same source segment in the same file
Project duplicate: Same source segment in the same project
General duplicate: Same source segment in files from different projects
Some historical references
If I remember things correctly (which I sometimes do not), it was like this:
The first versions of TRADOS did not produce multiple translations if you edited a segment, for which there was a 100 % hit in the memory. As some translators wanted this to happen, this was changed in one of the updates. There were a lot of complaints on this change, as many translators (including me) wants to avoid multiple translations in most cases. If you work with highly repetitive material from the same sources year after year, you would appreciate not to have a pollution of the memory by segments that are edited or corrected.
Then Trados made the possibility to choose approach, by setting Multiple translations: No and Options: Do not create new translation units if only text fields differ. I then reorganized my databases and removed index fields, and only used a text field for the project number, then it was possible to stop the memory pollution. This is a Workbench setting, so it works both in TagEditor and when working with Word + Workbench.
The upgrade to Studio showed that SDL/Trados almost completely had forgotten about this debate, and the Atril people never had it.
Studio:
Creation of multiple translations
If you come to a line in the translation that has a different translation in memory, there is a little yellow triangle showing below the translation results window. Clicking on that triangle gives you a warning message. But if you do not observe that, the default is that a new memory entry is produced.
Checking multiple translations
If you already made the mistake, nothing shows that the same source is translated differently. The verify function does not check this. There is a possibility to filter on Duplicate segments, and thus eliminate multiple translations for file duplicates, but not for project or general duplicates.

Atril DVX
Creation of multiple translations
There is a protection against file duplicates, and also a claim that project duplicates are protected. But if you add a new file in an old DVX project, then that protection is overridden.
If you come to a line in the translation that has a different translation in memory, there is no warning message and a new memory entry is produced.
Checking multiple translations
If you already made the mistake, nothing shows that the same source is translated differently. The QA function does not check this. There is a possibility to filter on Duplicate segments and also for duplicate segments with different translations, and thus eliminate multiple translations for file duplicates, and probably for project duplicates if you run all files in the project in the editor, but not for general duplicates.
Old Trados
Creation of multiple translations
It is possible to choose setting for this, and multiple translations can easily be avoided as the default is that memory is changed.
Checking multiple translations
Translate until fuzzy resolves all kinds of differences between memory and document. This kind of function that opens unit by until, compares document to memory and reports differences unfortunately is missing in both Studio and DVX
Autopropagate
The remedy for keeping the translation consistent in Studio and DVX seems to be a function called Autopropagate. I have tried it in both program, but it is complicated to set up and it is limited to file duplicates (in DVX also project duplicates) but cannot handle general duplicates. And in the present version of Studio the Autopropagate is useless because of the numbers/codes corruption. If you have Autopropagate turned on, and there is codes like 2.3.4, they are changed in all segments to 2,3.4 (Swedish language). This will probably be fixed in next service pack.
There are practical problems with the Autopropagate. If you choose prompting for changes, and have 20 occurrences of same headers, the prompt will turn up 20X20 times and you will have 400 extra clicks which is not a productivity enhancement. If you have automatic update, and then there actually *is* a few units that must have different translations in the same file, then you get into trouble.
My translation model
The way I mainly work with translation may be somewhat old-fashioned, but I actually prefer to work without pretranslation and propagation, I want to start with the first line, then take the next, and get the hits from memory, edit all hits, when needed, and if I edit an 100% hit I will have the edition in memory.
After translation I would like to have a consistency check to see that all duplicates are translated the same and that all sentences are consistent with the memory.
If I make terminology changes with search and replace,, I need a simple method of browsing through the document that lets me edit and resave changed TU:S
This very simple and straight forward MO was possible in Trados, but unfortunately is not in Studio or DVX. The main problems are the multiple translation production, and the lack of methods to compare document to memory.
Both function seems to be easy to implement at least as an option, as there is possibilites in both programs for this, which are not utilized.

Deja vu multiplex

2009-12-08T11:11:00.000+01:00

To see how Déjà vu handles multiple translations is unfortunately a Déjà vu experience. I tried this for my almost finished Studio project; export the memory and retranslate/proofread in Déjà vu. There is of course some loss of hits, as segmentation differs, but that was in the beginning compensated by the very nice Auto assembly/Lexicon function that picked up parts of the segments from memory and in many cases recreated the missing TU:s very nicely. But then I run into a serious problem and according to list mates on déjà-yahoo that is the way it is supposed to work and thus unsolvable.
The problem is that when I make an edit during proof to a 100 % match, Déjà vu for some reason produces a multiple translation in the memory! I think it was acceptable to correct 10 % of the sentences during proof, but to manually remove a multiple translation for every edit change in a 40 000 word proof would simply not be possible. Neither can I accept to have a database where a lot of the units are incorrect.

After making comprehensive tests with both Studio and DVX, I do not understand the logic behind the solutions that CAT programmers make. They are in many cases contra productive and an obstacle for the translator, instead of being a support in the translation process.

Tool change losses

2009-12-06T18:01:00.000+01:00

I will give you an example of how much segmenting differences to expect when migrating between different tools.
This is a project with 5 word files and one Excel. It is 100 % translated in Studio.
The below shows, that of the 41144 words fully translated in Studio you have
TE: Total 40430 words, 100 % 34383, 95/99 3198, no match 1065
DVX: Total 38638, 100% 28986, 95/99 4711, no match 850
The conclusion is that you will probably not lose more of your hits when migrating from Trados to Studio then you get when changing to Déjà vu. The loss seems to be around 10-15 %.
(Sorry for the lack of table format, unfortunately a little difficult to set up in this blog!)
When moving the project you get this analyse:
TagEditor
Analyse Total (6 files):
Match Types Segments Words Percent Placeables
Context TM 0 0 0 0
Repetitions 90 641 2 108
100% 3,958 34,383 85 124
95% - 99% 333 3,189 8 405
85% - 94% 44 592 1 71
75% - 84% 29 319 1 25
50% - 74% 11 241 1 6
No Match 63 1,065 2 26
Total 4,528 40,430 100 765
Chars/Word 5.19
Chars Total 210,093
Déjà vu
Swedish
Lines Words Chars Codes
Guaranteed 0 0 0 0
Exact 3335 28968 181710 581
Duplicate 489 973 7680 119
99% 149 2609 16339 132
98% 72 1210 7704 123
97% 33 431 2692 55
96% 16 233 1510 24
95% 15 228 1429 29
90%-94% 56 751 4569 81
80%-89% 76 727 4626 110
70%-79% 80 664 4346 96
60%-69% 63 454 3198 94
50%-59% 34 222 1614 58
40%-49% 20 95 668 27
30%-39% 15 129 855 13
20%-29% 13 131 844 11
10%-19% 2 8 57 2
0%-9% 0 0 0 0
No Match 216 850 6202 119
-------------------------------- ---------- ---------- ---------- ----------
Total: 4684 38683 246043 1674
-------------------------------- ---------- ---------- ---------- ----------
Internal Repetition: 33%

The Deja vu experience

2009-12-06T11:58:00.000+01:00

CATS comes in different shades, and it is not always easy for a human being to understand the soul and mind of a CAT.
So for the next project I think I will be trying out Déjà vu.
The first thing was to move my Studio TM, 450,000 TU:s to Déjà vu.
Studio can only export TM in TMX, not in Trados txt. And when trying to import the TMX, DVX just froze. So I had to import the TMX to Trados Workbench, then export again as txt, then import to DVX. The import into DVX took 12 hours.
But, otoh, the conversion of my RH-term that I maintain in Excel, and used to import in MultiTerm was very quick. The 20 000 entries 4 language database used to take half an hour to transfer in Trados, but the import into Déjà vu took 1 minute! And the import function was very easy to understand, and worked perfectly! So much easier to work with than MultiTerm.
As to AutoSuggest, the function I liked the most in Studio: Déjà vu has it, and it is actually more powerful.
You are not forced into pretranslate in Déjà vu, and you do not have to use Autopropagate.
The problem with number conversions is not there, as you can opt it out.
The learning effort is less; I have spent this Sunday with the conversions, and it seems to be very easy to get into the way Déjà vu works, although it is of course somewhat different than Workbench/Studio.
But, I would say that if you are to convert from Trados Workbench/TagEditor, it is probably as easy (or even easier) to convert to Déjà Vu.
This is of course only my first enthusiastic reaction to something new, but I have a strong feeling I will change to Déjà vu as my main tool and keep the Trados for conversion purposes.
I have not investigated the duplicate handling in Déjà vu yet, but there is a very good function that is missing in Studio; Déjà vu tells you without having to use Autopropagation, when you hit a repetition, which Studio does not. And this solves my proofreading problem of the mess I got in Studio in my old 40 000 word project. So I will actually run it again in Déjà vu. As the segmenting differs, it will take a day extra, but that will be worth it as I do not have to worry about changing codes and I can easily check the repetitions and eliminate the false duplicates created by Studio.

Addendum: That is what I first thought, but after running DVX a while I realized that DVX also creates false multiple translations. It is easier to detect and remove them in DVX, but it is still an unnecessary manual task. And the multiples should not have been created at all. The prevention against multiple translations in DVX works only if the hits are created within a project, not if you have imported them from another source. And I do have 450,000 TU:s that I am not willing to throw away or edit manually, one by one.

13. SDL Trados studio vs TagEditor, tagging

2009-11-11T12:49:00.000+01:00

Tagging rules

Everyone who has tried to move a project from and to Word+TagEditor has experienced that the tagging is different. If you have translated a project 100 % in TW+Word, and then run it in TagEditor, you will not get all 100 % matches.
The differences in the tagging in Studio are even more different. So, when "upgrading" to Studio, you think that when an old customer returns with a manual revision you have done in old Trados, you get all your hits? Not true. I discovered this during a job I tried to do in Studio, and reverted back to TagEditor.
I have now finished the project in TagEditor, analyze:
Analyze Total (4 files):

Match Types Segments Words Percent Placeables
Context TM 0 0 0 0
Repetitions 1 1 0 1
100% 2,356 10,881 100 1,243
95% - 99% 2 3 0 2
85% - 94% 0 0 0 0
75% - 84% 0 0 0 0
50% - 74% 0 0 0 0
No Match 0 0 0 0
Total 2,359 10,885 100 1,246
Then I converted the database back to Studio, opened the source files in Studio, and made an analyze:
Repetitions 2349
100% 4799
95/99 1418
85/94 2143
75/84 903
50/74 128
No match 110
Total 11 895
There is a workaround for this, see chapter 15.

9. Trados Studio File Structure – Moving files

2009-11-03T22:35:00.000+01:00

In many old Trados programs there have been problems, when files are moved. The old "log file bug" in Trados 2007 seems to be there still; if you move or delete the log file, then you get a totally incomprehensive error message. There was a question on TW_users about this just a few days ago. And this error has always been there, since the first versions of Trados in 1997.
Not to mention the MultiTerm files, that are not possible to move.
And, the file handling problem seems to resurface in Studio too. Today I by mistake registered a new project and the files unfortunately went into a sub-folder in another project, as I was careless when I registered the project. So I thought that this would be easy to fix: I shut down Studio, moved the project folder, with all the files, to the location where it should have been. Then I opened Studio, and the project was not on the project list. And, here it would be nice with an error message saying:
"The project xxxx.xxx was not found. Delete the project (yes) or find it in the file system (yes)"
When clicking yes, there would be a window where you can pick the new location for the project.
As it is now, you can delete a project folder (by mistake) and there is no trace of it when opening Studio, not even an indication that a project has been lost.
Then I tried to open the project with File/Open/Project, and found the project file named xxxx.sdlproj in the project folder, but when trying to open the file I get this error message:
"The selected project could not be opened"
I also tried to move another project, with the same result. Obviously Studio is sensitive for the exact location, so you cannot move a project folder to another location and open it. That is a very hard restriction. Sometimes you want to move projects. I have a habit of moving last year's projects to another hard disk. And sometimes you have to buy a new computer. That would mean that all old projects would be impossible to open, unless you manage to create exactly the same file structure the project was created in.
Suggestion from SDL: Buy the Professional version and work with "Packages". Seems to be a complicated and expensive solution to a simple problem

3. AutoSuggest and Automatic Translation

2009-10-27T14:47:00.000+01:00

AutoText and AutoSuggest
In Word, and also in later versions of TagEditor, there is a function called AutoText. The AutoText entry in Word can be accessed in two ways; a) typing the first four letters of the word makes a suggestion displayed below the segment b) entering the abbreviation you have chosen and then push F3. You can enter new AutoText entries in the glossary by highlighting the word and push Alt+F3. The glossary can also be edited separately.
In the Studio AutoText abbreviations are not used, so you always have to type the four first letters. And it is not possible to add AutoText entries on the fly, you have to use Tools/Options/AutoText, and enter the words one by one. I have not yet found any method of separately editing an AutoText file. If that is possible, it would be nice to compile a list or even import the AutoText entries already prepared in Word. There is an import/export function in Studio's AutoText, so it is probably possible, if you can find software that is able to edit the file type.autoTxt.
AutoSuggest
One of the new functions in Studio is AutoSuggest, but it is not included in the Freelance version, you have to buy a separate license. When typing, the AutoSuggest functions like AutoText as above. The difference is that the AutoSuggest glossary entries are context dependent, so only words that AutoSuggest believes you will use in the segment are displayed. And that is exactly what makes the function so useful!
The AutoSuggest glossary is generated by processing a translation memory that has to hold at least 25 000 TU: My memory for English into Swedish was a little larger than 400 000, the maximum recommended value, and the result was very good. There is also a possibility to download AutoSuggest glossaries from SDL, but not in my language combinations (Eng>>Swe. Fin>>Swe and Ger>>Swe).
The AutoSuggest is actually a very useful function, especially in Swedish, as we often have very long words. And, the AutoSuggest creation includes not only commonly used single words from the TM, but also short often used phrases. It is also case sensitive, which is nice as I sometimes get the first part of standard phrases from the suggestion.
And, you of course also get the terms from your MultiTerm glossary in the same drop down box below the segment. As I have saved all special terminology and customers glossaries from all projects during the years in my MultiTerm, this is also a powerful resource.
The AutoSuggest gets terms from three sources; the AutoText glossary, your AutoSuggest compiled from translation memory and from MultiTerm.
Automatic Translation
In old Trados, you get suggestions for translations in the target window for exact and fuzzy hits. That is of course the same in studio, but you can also add suggestions from the SDL Automated translation server. There is a version for English into Swedish, but not for my other language combinations. The connection to the server is fast and stable, there seems to be very few delays, save the opening of the first segment, which takes a few seconds. Most of the suggested translations are of course useless, and would take more time to edit than to rewrite. But in some cases, about 1 of 10 sentences I choose the suggestion and edit. In a very few cases I get a complete sentence that only have to be slightly corrected. And, the database quality could be improved; there are even misspelled words in the suggested sentences.
In rare cases you get a hit like this in the Automatic Translation:
Source:
If a solution with necessary concentration is available and can be transported without problems to the plant, it can be used.
Automatic target:
Om en lösning med nödvändig koncentration finnaer till hands och kan transporteras utan problem till plantan, kan användas det.

Ok, there are errors, but definitely so few that it is worth editing. And I see a possibility to development here: I think it would be relatively easy to combine the SDL Automated translation server with a technical vocabulary that explains the fact that plant mostly is anläggning. The "finnaer" seems to be a misspelled word in the database that could easily be corrected.

But there definitely is a future for this kind of translation support, and it is actually useful at least for the type of text I work with. So I will definitely have it up and running.
I believe that the combination of the new support functions in Studio already is boosting my productivity, and will do even more, as I get used to this new way of combined typing/selecting/editing.

Trados Studio upgrade 2 - Editing

2009-10-25T09:21:00.000+01:00

I would say that the new Trados Studio is not an upgrade; it is a completely new product with very few similarities to vintage Trados.
The side-by-side editor is more like all the other CAT programs than old Trados. The WYSIWYG approach in TW + Word that many Trados fans loved is completely abandoned. But it was not so easy to maintain WYSIWYG in the old Trados for Word files either.
For most freelancers I believe that Word still is the mainstream document format. In many cases even large 100 page manuals are written in Word. In the old Trados, there were very often peculiar changes in the formatting of Word documents, and in some cases Word documents refused to open if they were heavily formatted. According to Trados it was never a Trados problem, it was a Microsoft problem. That is a highly philosophical question of who is to blame when two programs lack in compatibility.
Project approach
Then there is the project approach, which from a freelancer point of view often seems to be unnecessary complicated and not needed. If you get a one page translation that has to be delivered the same day, you do not want to waste time with complicated project settings. There is also a single file option in the Studio, Open document, but it is actually almost the same settings as for a project. So my choice is to enter everything as projects, be it a huge multifile manual or a single press release. The project settings are comprehensive (read: complicated) but they can be saved as a template, and reused. I have made a template for each of my language combinations, and when starting a new job the correct TM, language combination in MultiTerm and other preferences are automatically set.
Editing
Most of us vintage Trados fans found it difficult to start working with TagEditor, but it was necessary, as it was required for all file formats other than Word, and in fact it became the first choice even for large formatted Word documents. The first versions of TagEditor were unstable and crashed at the drop of a hat. Later on the TagEditor became more stable, but despite the wailing and moaning amongst the Trados freaks on TW_users and other fora, the editing was still far from Word, and was not possible to adapt to the individual user, as keyboard shortcuts could not be defined and there was no AutoText, at first.
Side-by-Side editing: The feeling is something in-between Word and TagEditor, maybe the best of both. In Studio, you have the same editing functionality as in Word, as it comes to dragging around words, keyboard combinations for highlighting etc. The AutoText functionality whit the possibility to add new words on the fly is unfortunately not there, but otoh this is compensated with the new function AutoSuggest, which I will write about later.
The font formatting in the side-by-side editor has also borrowed something from both TagEditor and Word; there is a tagging system that is like TagEditor, but changes in the visible formatting also affects what will be seen in the target Word document. Underlining a word in the target field will be seen in the target document. It is also possible to copy formatting from the source field to the target without using tags.

Trados Studio upgrade 1 - Introduction

2009-10-24T23:30:00.001+02:00

Hi all Trados lovers!
I have now upgraded to SDL Trados Studio 2009 Freelance Plus version 9.1.1107.0 from Trados 2007. After spending some hours upgrading, the general impression is that it is a very nice package. All the new functions seem to improve the workflow. I am going to jot down some notes about the upgrade process, support, problems, and also pros and cons with the new system. Feel free to comment!

The installation of Trados Studio was relatively easy, and the new programs could easily convert both TM and MultiTerm databases. We installed it on two computers, as it is allowed for the Plus licence. There were some initial problems of course, there always is; I got a "deserialization" error message a few times, and there was some problems converting word files. I will tell you all about it later on.
The SP1 upgrade arrived. At that point I decided to do a complete reset of my computer, as it had become somewhat overloaded with all kinds of stuff, and was working slower than normal. Of course I forgot to return the license before doing this! So late in the evening, I found that I had no working version of Trados at all. But the Trados support was fast, and it did not take many hours the next day before I had a new licence number. It took a few telephone discussions with the sales person and the licence department to fix this, and I do not even have any service contract. I also got a license for the AutoSuggest add-on to my Freelance version.
So: Remember to return the license before uninstalling!