Penalty settings in trados file analysis.
Thread poster: Alexander Chisholm
I am having trouble replicating a file analysis sent to me by an agency.
The total number of words is the same, but the distriution of words among the various match bands is different, with my version having a lot more words in the "no match" band (and so presumably costing more).
When I asked them, they suggested setting all penalties to zero and repeating the analysis (this is after having received new source file and new updated TM from the client) however, the result was still closer to my analysis than the log extract they sent me.
I'm not so worried about this, but more concerned with the practice of setting penalties prior to performing an analysis for quotation and pricing.
Why does Trados have default values in the penalty panel? I asume, basd on my knowledge of other software in different sectors, that the default values are suggested by professionals who perform the task frequently and give the most representative results in the majority of cases.
Why then does the client choose to set different settings here, or is it purely a way of obtaining a resulkt that is favourable to them in terms of billing?
Has anyone had similar experiences and what do you think.
What is the best reply to give the agency about any future analyses used fo quotation and billing?
Many thanks for any suggestions.
| || || |
| Many thanks for the reply. || Jun 10, 2007 |
They have definitely set all penalties to zero, and they asked me to do the same to verify their analysis, although the results were still not the same, so I will have to ask them about version numbr and segmnentation settings.
What do you think about my question regarding their reasons for setting penalties to zero. Is this purely to achieve a lower number of "no matches" and so reduce the billable cost? And is this "honest" since I'm assuming the default settings are not set to zero for a reason?
| | Jerzy Czopik
Local time: 09:20
Polish to German
| I cannot give you a precise answer without knowing the files and TM... || Jun 10, 2007 |
but I can give you some ideas.
The penalties are set to certain default values not by accident.
If you take for example the penalty for missig TAG (it is called placeable) - it is obvious, that each missing TAG must (at least should) be inserted. If it is missing, you have an extra work with the match. The same is with penalty for formatting. When some formatting is missing in your sentece, you will have to restore it - means additional work. (Formatting corresponds to TAG - in Word you have formatting, elswhere TAGs)
Then there is a penalty for multiple 100% matches - very important for example for Polish. We often need different translation of the same phrase, depending on context. Whe you set this to zero, you will not see, that there are other possible translations in TM. And having multiple translations is additional work, so penalty is of course necessary.
You also have penalty for alignment - it is obvious, why, I think.
Then attributes and text fields - applies for a TM for different customers, sorted with attribute fields. Also very important, but this may be set to zero, if TM is only for one customer and one machine. One customer with different machines may require this penalty, if they use different meanings of words for different machines (happens sometimes).
Put in a nutshell: setting penalties to zero may cause you earn less, if you are applying rebates to different match types.
Ask the customer, why do they want to set penalites to zero. Ask them, if they have had set the proper language when analysing. Ask also, if they did change any segmentation rules - those might have influenced the analysis results even more, than the penalties.
| || || |
| Many thanks - very useful suggestions and tips. || Jun 11, 2007 |
| | vjuanatey
Local time: 08:20
English to Spanish
| Trados penalty settings query || Jun 15, 2007 |
I was reading your emails about the penalty settings in Trados, and I was wondering if you could help me.
We have to update a 25.000 words file. The new words are less than 6 thousand. We are sending the file for translation to one person and for ED to another one.
We want the editor to edit ONLY the new translated text, thus ignoring 100%. So he will just focus on Fuzzys. The problem we have is related to 98% matches. Most of these appear as fuzzys even if they are exact matches because of formatting issues.
I am aware there is a way to make these 98% show as 100% matches by using the Penalties settings.
Do you have any idea of how to do this?
Thanks very much for your time.