Discrepancy betw. export and analysis
Thread poster: Heinrich Pesch

Heinrich Pesch  Identity Verified
Finland
Local time: 04:14
Member (2003)
Finnish to German
+ ...
Sep 25, 2006

Workbench (SDL Trados 2006) with the TM sent by the customer analyses the files and finds about 250 words in matches below 99% and no match together.
But when I export the less than 99% matches to rtf and remove all tags I get 500 words to translate.
I have been suspicious already for long about these jobs, which show rather little work according to Trados analysis but take much more time than expected. The customer of course wants to pay according to analysis.

Can anybody explain why WB does not show all these words in analysis it exports?
The client's analysis result is the same as mine, we use the same settings.

Regards
Heinrich

[Bearbeitet am 2006-09-26 02:01]


Direct link Reply with quote
 
kimjasper  Identity Verified
Denmark
Local time: 03:14
Member (2006)
English to Danish
+ ...
Have a look at the segmentation rules Sep 25, 2006

When I have seen differences in Trados logs, the reason has been different Trados segmentation rules. If your Trados setup has different segmentation rules, Trados will split the text differently - it will either give you smaller or larger translations units. Since you are using different translation units, you will have less 100 % matches than your client. I would ask the client to send a screenshot of the segmentation rules Translator's workbench > files > setup > segmentation rules).

Hope this helps
Kim


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:14
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
We use the same settings Sep 26, 2006

The analysis result is the same as in the analysis file the client has sent me. I should have mentioned this in my original posting.

So I wonder why WB first analysis the files, announces 250 words in all less thän 100% matches and exports in the next moment a file with 500 words.
And this is not the first time, I believe. I should double my rates in order to make up for the actual work.

Regards
Heinrich


Direct link Reply with quote
 
bmann
Local time: 20:14
English
Does the document have a lot of numbers? Sep 26, 2006

Trados does not count numbers in the analysis but Word does.

Direct link Reply with quote
 

Vito Smolej
Germany
Local time: 03:14
Member (2004)
English to Slovenian
+ ...
500 or 250 ... Sep 26, 2006

Heinrich Pesch wrote:And this is not the first time, I believe. I should double my rates in order to make up for the actual work.


This kind of work should not be charged by word count but by actual time spent on it. Of course the issue gets serious at 5000 vs 2500...

If Trados gives 250 counts, where do you get 500 from?

Regards

Vito


Direct link Reply with quote
 

Antoní­n Otáhal
Local time: 03:14
Member (2005)
English to Czech
+ ...
a few unclear points Sep 26, 2006

- Is it a Word file of ttx, or anything else?

- What does Trados analysis show when applied on your "exported rtf" file?

- Does the TM contain tags? If yes, the difference may be caused by your deleting the tags (why do you delete them, by the way)?

Antonin


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:14
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
Its a ttx Sep 26, 2006

After your suggestion, Antonin, I analysed the rtf which WB had exportet. There are a total of 480 words and of course no 100% matches.
I removed the tags because with tags Words counts them as words and finds more than 520 words.
There are not much numbers either in the export.


Direct link Reply with quote
 

tectranslate ITS GmbH
Local time: 03:14
German
+ ...
QED Sep 26, 2006

Heinrich Pesch wrote:

Its a TTX

That explains why the results are different - when you export these contents to Word (and change/delete the tags), the segmentation and thus the number of matches WILL differ.


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:14
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
Why? Sep 26, 2006

tectranslate wrote:

Heinrich Pesch wrote:

Its a TTX

That explains why the results are different - when you export these contents to Word (and change/delete the tags), the segmentation and thus the number of matches WILL differ.


Segmentation has nothing to do with it. WB analyses the files and exports the unknown (less than 100% known) segments. So these exported segments should show up in the analysis, they are the stuff I have to translate, no matter if they are in a ttx or in whatever.
The tags are not deleted when I analyse the export in WB. They are deleted prior to wordcount in Word. Deleting tags cannot create more text.


Direct link Reply with quote
 

Antoní­n Otáhal
Local time: 03:14
Member (2005)
English to Czech
+ ...
Yes, I agree with tectranslate Sep 26, 2006

A TM of ttx files does contain tags, so that is a likely reason for a lower rate of "similarity" after you delete them

Having said that, I would also add that you may be artificially increasing your workload here - from my viewpoint, exporting unknown sgements to a rtf (which I used myself some time ago) was good for "emulating" TagEditor approach by Word (AFAIK justified before TagEditor could transform Word into ttx directly), while directly translating the ttx in TagEditor would better exploit the TM provided to you by your client and save your work.

In other words, you feel like charging them more because you have more work, but if the more work is caused by your nonstandard processing of the files, such attitude would be a bit unethical and they would be entitled to refuse such an increase of pay, I am afraid.

Of course, I am just guessing without actually having seen it, so I may be utterly wrong here.

Antonin


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:14
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
I did not delete the tags Sep 26, 2006

Antoní­n Otáhal wrote:

A TM of ttx files does contain tags, so that is a likely reason for a lower rate of "similarity" after you delete them

Having said that, I would also add that you may be artificially increasing your workload here - from my viewpoint, exporting unknown sgements to a rtf (which I used myself some time ago) was good for "emulating" TagEditor approach by Word (AFAIK justified before TagEditor could transform Word into ttx directly), while directly translating the ttx in TagEditor would better exploit the TM provided to you by your client and save your work.

In other words, you feel like charging them more because you have more work, but if the more work is caused by your nonstandard processing of the files, such attitude would be a bit unethical and they would be entitled to refuse such an increase of pay, I am afraid.

Of course, I am just guessing without actually having seen it, so I may be utterly wrong here.

Antonin



I have to export the unknown segments, because they are only a small part of the whole file. If I work directly in ttx, I have no clue what was actually translated anew and what not, and the proofreader would have to read the whole large file, and I would have to pay him more than I get from the client.

And I have not deleted any tags! WB analyses, exports and analyses the rtf without me doing anything in between.
Of course with tags Word counts more words than without. But the file I feed into WB is untouched.

But I'll leave at that. Anyway I'm going to change my pricing back to lines. And without any discount for partial matches.

Thanks to you all.

Regards
Heinrich


Direct link Reply with quote
 

tectranslate ITS GmbH
Local time: 03:14
German
+ ...
Hm Sep 26, 2006

Heinrich Pesch wrote:

The tags are not deleted when I analyse the export in WB.
Hm, strange. I guess without seeing the actual file, I cannot come up with a logical explanation.

I just performed a test with an HTML file (this forum page saved on my HDD) and the analyses of the original and exported files with an empty TM came up with the precise same number of words.

Sorry - the process is usually very reliable. There must be something about your specific file(s) or the analysis process you're using that leads to these strange results. I understand your frustration, though.

Regards,
Benjamin


Direct link Reply with quote
 

RWSTranslation
Germany
Local time: 03:14
Member (2007)
German to English
+ ...
SDL Support Sep 27, 2006

Hello send the problem to the SDL support.

It would be interesting to see there answer.

Regards

Hans


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:14
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
SDL support Sep 28, 2006

Thanks for the advice, Hans! I did post a question regarding this matter and sent the files (analysis and export) to SDL. I'll keep you updated about any answers from SDL.

For further readers of this thread, who might wonder why I export to rtf: I rather work in Word because I spot typos immediately and Word has better functions for regrouping text. TE can only remove and replace, in Word I can drag. And my proofreader does not have Trados, so if I work in TE I have to save the target to rtf, send it to the proofreader, compare the corrected file to the original and do the corrections manually in TE, which is rather cumbersome. When translating in Word, I can send the rtf uncleaned to the proofreader and update the TM automatically, after which translating the ttx-files is only a matter of minutes (TE checkes the tags at the same time).
And because the uncleaned rtf is compatible to Wordfast, I can update also my Wordfast TMs, if I want to, and do not need to go through WB.
In fact my procedure is the most rapid in my situation.

Regards
Heinrich


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 »

Discrepancy betw. export and analysis

Advanced search







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 »
SDL Trados Studio 2017 Freelance
The leading translation software used by over 250,000 translators.

SDL Trados Studio 2017 helps translators increase translation productivity whilst ensuring quality. Combining translation memory, terminology management and machine translation in one simple and easy-to-use environment.

More info »



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