Mobile menu

Can I build a Trados memory using an untranslated rtf file?
Thread poster: Alvin Liu

Alvin Liu  Identity Verified
China
Local time: 00:34
Member (2007)
English to Chinese
+ ...
Jun 30, 2005

Assume I have 2 rtf files A and B, and
1) A is untranslated, i.e. not a bilingual file;
2) B is a an update to A
3) I want to build a memory with rtf A to see to what extent A is upgraded to B.

Any one can help?


Direct link Reply with quote
 

Ralf Lemster  Identity Verified
Germany
Local time: 18:34
English to German
+ ...
Comparative analysis Jun 30, 2005

Hi Alvin,
You cannot build a TM using a monolingual file (as a translation memory is, by definition, bilingual).

However, you don't need a TM to analyse the degree of changes between A and B:

- Analyse file A; keep the dialogue open.
- Check Use TM from previous analysis.
- Remove file A from the list box.
- Add file B and analyse it.

HTH, Ralf


Direct link Reply with quote
 

Caroline McLoughlin  Identity Verified
Ireland
Local time: 17:34
Member (2003)
German to English
+ ...
Yes you can do this Jun 30, 2005

Hi,

Create a blank translation memory, go to
File\New...etc

Go to Tools\Analyse
Add File A for analysis

When the analysis is complete do not close the Analysis window
Clear file A from the dialog box
Check "Use TM from previous analysis"
Add file B for analysis
Analyse

This compares file B with file A
Your results should show the number of 100% matches, new words, fuzzy matches etc in file B compared with file A.

Hope this is clear.
All the best,
Caro


Direct link Reply with quote
 

RWSTranslation
Germany
Local time: 18:34
Member (2007)
German to English
+ ...
Another possibility Jun 30, 2005

Ralf Lemster wrote:

Hi Alvin,
You cannot build a TM using a monolingual file (as a translation memory is, by definition, bilingual).

However, you don't need a TM to analyse the degree of changes between A and B:

- Analyse file A; keep the dialogue open.
- Check Use TM from previous analysis.
- Remove file A from the list box.
- Add file B and analyse it.

HTH, Ralf


Hello,

if you use Ralfs method, you will only analyse the second file against a smaller part of the TM with the additional input of the file A.

If you want a more precise analyse you can do the following:

Make an alignment of file a against file a.
Export the alignment with a user name which is not ALIGN.

Make a copy of your tm.
Import the alignment into the copy
Analyse file B against the copy of your TM.

With kind regards

Hans


Direct link Reply with quote
 

Alvin Liu  Identity Verified
China
Local time: 00:34
Member (2007)
English to Chinese
+ ...
TOPIC STARTER
different results Jun 30, 2005

Method B - build a memory in WinAlign using rtf file A as indicated by DSC, and then analyzed file B, the results as follows:

Match Types Segments Words Percent Placeables
XTranslated 0 0 0 0
Repetitions 323 1,088 12 0
100% 0 0 0 0
95% - 99% 148 880 10 0
85% - 94% 56 395 4 0
75% - 84% 35 288 3 0
50% - 74% 0 0 0 0
No Match 985 6,411 71 0
Total 1,547 9,062 100 0
-------------------------------------------

Method A - as indicated by Ralf and Caro, the results as follows:

First analyzed File A, then proceeded to File B with ¡°Use TM from previous analysis" box checked, the results as follows:

Match Types Segments Words Percent Placeables
XTranslated 0 0 0 0
Repetitions 202 806 9 0
100% 245 998 11 0
95% - 99% 57 378 4 0
85% - 94% 106 690 8 0
75% - 84% 74 486 5 0
50% - 74% 0 0 0 0
No Match 863 5,704 63 0
Total 1,547 9,062 100 0
----------------------------------------
so, what causes the difference?


Direct link Reply with quote
 

RWSTranslation
Germany
Local time: 18:34
Member (2007)
German to English
+ ...
Differences Jun 30, 2005

Hello,

did you inport the Alignment into an existing memory ?

If you didn't change the user of the Alignment units, you can set the punishment of Align Units to 0 %.

Hans

[Edited at 2005-06-30 11:15]


Direct link Reply with quote
 

Harry Bornemann  Identity Verified
Mexico
English to German
+ ...
The trade secret caused the difference Jun 30, 2005

Alvin Liu wrote:
...so, what causes the difference?

If the Alignment Penalty caused the difference (like Hans proposed), then Method B (WinAlign) should have lower values than Method A.
But the opposite happened.

I am afraid you will never know what really caused the differences, because the way the percentages are calculated is a trade secret of each CAT manufacturer, and the algorithms would fill several pages.


Direct link Reply with quote
 

Alvin Liu  Identity Verified
China
Local time: 00:34
Member (2007)
English to Chinese
+ ...
TOPIC STARTER
differences still exist after the alignment penalty has been set to 0% Jul 1, 2005

DSC wrote:
did you inport the Alignment into an existing memory ?

In both methods, I created empty memory.

If you didn't change the user of the Alignment units, you can set the punishment of Align Units to 0 %.

I did not do this last time, but now differences still exist after the alignment penalty has been set to 0%. Only repetitions and No Match remain, as follows:

Match Types Segments Words Percent Placeables
XTranslated 0 0 0 0
Repetitions 323 1,088 12 0
100% 0 0 0 0
95% - 99% 0 0 0 0
85% - 94% 0 0 0 0
75% - 84% 0 0 0 0
50% - 74% 0 0 0 0
No Match 1,224 7,974 88 0
Total 1,547 9,062 100 0


Direct link Reply with quote
 

Alvin Liu  Identity Verified
China
Local time: 00:34
Member (2007)
English to Chinese
+ ...
TOPIC STARTER
The Use TM from Previous Analysis option (Analysis dialog)-additional information from Trados manual Jul 1, 2005

This option allows you to use a translation memory created during a previous analysis.
This option is particularly useful if you want to compare different translation projects with each other. For example, assume you want to analyse two different releases or versions of a manual. The manual files are called manual1.doc and manual2.doc and you have a translation memory called manuals.tmw. To find out the differences between the two, do the following:

1. Analyse the first version of the manual (manual1.doc), running it against your translation memory manuals.tmw. When analysing files, Translator's Workbench builds up a temporary translation memory containing all source segments from the analysed files. In this example, after analysis Translator's Workbench has created a temporary translation memory containing all segments of the first manual, manual1.doc.

2. Clear the file manual1.doc from the Files to Analyse list and add the second file, manual2.doc, to it instead. Now, if you check the Use TM from Previous Analysis option, Translator's Workbench does not run the second file against the translation memory manuals.tmw, but against the temporary translation memory created in step 1, containing all source segments from
manual1.doc. The results of this second analysis shows the differences between the two manuals, enabling you to see exactly to what extent they are different and to what extent they are identical or roughly the same.

Taken From: Translator's Workbench Help, (c) 2004 TRADOS Incorporated.


Direct link Reply with quote
 

Harry Bornemann  Identity Verified
Mexico
English to German
+ ...
Empty TM Jul 1, 2005

Alvin Liu wrote:
...
Match Types Segments Words Percent Placeables
XTranslated 0 0 0 0
Repetitions 323 1,088 12 0
100% 0 0 0 0
95% - 99% 0 0 0 0
85% - 94% 0 0 0 0
75% - 84% 0 0 0 0
50% - 74% 0 0 0 0
No Match 1,224 7,974 88 0
Total 1,547 9,062 100 0

This means that this time you either used an empty TM or a wrong language specification, or applied wrong filter settings.

[Edited at 2005-07-01 07:46]


Direct link Reply with quote
 

Alvin Liu  Identity Verified
China
Local time: 00:34
Member (2007)
English to Chinese
+ ...
TOPIC STARTER
the differences still exist but seemingly narrowed Jul 1, 2005

Thank Harry for your correction! This time, the differences still exist but seemingly narrowed:

Method A ("Use TM from previous analysis" box checked, and the memory is empty)
Match Types Segments Words Percent Placeables
XTranslated 0 0 0 0
Repetitions 204 807 9 0
100% 305 2,466 27 0
95% - 99% 60 789 9 0
85% - 94% 128 1,079 12 0
75% - 84% 65 472 5 0
50% - 74% 0 0 0 0
No Match 516 3,505 38 0
Total 1,278 9,118 100 0
-----------------
Method B (Alignment Penality set to 0%, and the memory is built in WinAlign using the untranslated rtf file A)

Match Types Segments Words Percent Placeables
XTranslated 0 0 0 0
Repetitions 282 944 10 0
100% 201 1,959 21 0
95% - 99% 70 788 9 0
85% - 94% 63 652 7 0
75% - 84% 36 266 3 0
50% - 74% 0 0 0 0
No Match 626 4,509 50 0
Total 1,278 9,118 100 0


Direct link Reply with quote
 


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

Can I build a Trados memory using an untranslated rtf file?

Advanced search


Translation news related to SDL Trados





TM-Town
Manage your TMs and Terms ... and boost your translation business

Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.

More info »
Wordfast Pro
Translation Memory Software for Any Platform

Exclusive discount for ProZ.com users! Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value

More info »



All of ProZ.com
  • All of ProZ.com
  • Term search
  • Jobs