Pages in topic:   [1 2] >
What is the correct number of repetitions?
Thread poster: Stefan Stefanov

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
Mar 20, 2011

Hello,

I am looking at the Analisys report of a large project, and I think there is something wrong with it or I am failing to understand it.

105,965 words total
repetitions: 43,775 words ---- 48%
100% = 0
95% - 99% = 0
85% - 94% = 0
75% - 84% = 0
0-74%: 62,190 words --- 52%

The percentage figure is incorrect. All types of repetitions show zero.

This analysis is made by WFP on an empty TM.

Considering that Wordfast does not automatically propagate matched words or pharses but only segments (unlike Trados), do you think it is correct to say that there are 43,775 repetitions and these will not be paid for?

I'd like to hear your opinions, please.


 

Samuel Murray  Identity Verified
Netherlands
Local time: 02:42
Member (2006)
English to Afrikaans
+ ...
@Stefan Mar 20, 2011

Stefan Stefanov wrote:
105,965 words total
repetitions: 43,775 words ---- 48%
100% = 0
95% - 99% = 0
85% - 94% = 0
75% - 84% = 0
0-74%: 62,190 words --- 52%


The above figures are typical of an empty TM. The "repetitions" are whole segments that occur in the text more than once (the first instance of a repetition isn't counted as a repetition).

The percentage figure is incorrect. All types of repetitions show zero.


No, the figures marked 100%, 95%-99% etc aren't repetitions. They are matches from the TM, which in this case is all zero (except for the bottom category) because the TM is empty. I think you're confusing the term "match" with "repetition". A match may be partial, but a repetitions is never partial.

Perhaps you should look at the figures like this, rather:

Total number of words: 105 965 (100%)
Number of repetions: 43 775 (48%)
Number of non-repetitions: 62 190 (52%)

Fuzzy matches from TM:
100% = 0
95% - 99% = 0
85% - 94% = 0
75% - 84% = 0


The figure 0-74% means that so many words have a match of less than 75% in the TM (and this would include segments that have no match at all).

Considering that Wordfast does not automatically propagate matched words or pharses but only segments (unlike Trados), do you think it is correct to say that there are 43,775 repetitions and these will not be paid for?


There are 43 775 repetitions. Whether or not you'll get paid for them depends on your arrangement with the client. If the client wants you to leave them alone (not touch them, but simply let the CAT tool automatically fill them in), then not paying for them may be seen as fair.

Are you sure Trados does subsegment matching and subsegment match *insertion*?




[Edited at 2011-03-20 10:00 GMT]


 

Magdalena Szewciów  Identity Verified
Poland
Local time: 02:42
Member (2008)
English to Polish
+ ...
huh Mar 20, 2011

I think Stefan perhaps meant internal repetitions (available in any Trados 2009 Studio analysis). But WF doesn't offer the feature.

 

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
TOPIC STARTER
Trados Mar 20, 2011

Magdalena Szewciów wrote:

I think Stefan perhaps meant internal repetitions (available in any Trados 2009 Studio analysis). But WF doesn't offer the feature.


Hello Magdalena and Samuel,

Thanks for answering.

I see now, yes, I was confusing repetitions with TM matches. Anyway, the whole situation is problematic and I'll wait for the agency's reply, but I feel misled. I had made a 25% discount for volume, and then they try to get another 33% off for repetitions, which actually are of no help in Wordfast.

Working with Trados 2009, I noticed exactly this automatic partial propagation - the source was copied to the target, but where a word or a phrase from the source was found to have a match in the TM, the target segment appeared half-translated.

I am not sure what the required settings for this are, but Wordfast clearly cannot do this, and it would have been great if it could.

The closest thing WF can offer is supply the whole translated segment from the TM -- and after decreasing the fuzzy match threshold from 75% to 60% I have been able to get a bit more "matches" and thus, make the process less laborous.


 

Lori Cirefice  Identity Verified
France
Local time: 02:42
French to English
so why don't you use Trados for this project? Mar 20, 2011

You should be charging your full rate for 62190 words, and negotiate the repetition discount for 43775 words. Wordfast will recognize those 43775 words and automatically insert them, they are full segments as the WF analysis only counts full segments as repetitions. WF does not have any subsegment recognition feature, other than fuzzy matching which is really only effective for very short segments, especially those containing numbers.

It sounds like in addition to the full segment repetitions, you also have a lot of subsegment repetitions (I am not really familiar with this Trados feature, have only vaguely heard about it!). You can leverage those subsegment repetitions by using the glossary feature as I mentioned in the other thread (I suppose it must be possible to get similar results in WFP as in WFC).

Or, maybe you should switch to Trados for this job, since you seem to be familiar with it and you think that you can gain some advantage with the subsegment recognition?

Question - did your client base their pricing on a Trados or WF analysis? That may make a difference in your negotiation approach.


 

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
TOPIC STARTER
WF analysis Mar 20, 2011

Lori Cirefice wrote:

You should be charging your full rate for 62190 words, and negotiate the repetition discount for 43775 words. Wordfast will recognize those 43775 words and automatically insert them, they are full segments as the WF analysis only counts full segments as repetitions. WF does not have any subsegment recognition feature, other than fuzzy matching which is really only effective for very short segments, especially those containing numbers.

It sounds like in addition to the full segment repetitions, you also have a lot of subsegment repetitions (I am not really familiar with this Trados feature, have only vaguely heard about it!). You can leverage those subsegment repetitions by using the glossary feature as I mentioned in the other thread (I suppose it must be possible to get similar results in WFP as in WFC).

Or, maybe you should switch to Trados for this job, since you seem to be familiar with it and you think that you can gain some advantage with the subsegment recognition?

Question - did your client base their pricing on a Trados or WF analysis? That may make a difference in your negotiation approach.


Hi Lori,

It is a WF analysis and I was able to reproduce it exactly. Unfortunately, I can't use Trados, because the client is supplying (to the agency, and the agency to me) Wordfast Pro TXML files.

At first I paid little attention to the huge number of repetitions claimed because I had only had a brief look at the file to be translated - and that's a HUGE file - and I expected that there are many repeated segments. I realized that was not the case when I started working - indeed many phrases and words are repeated but in differing contexts and WF does not automatically supply the translation, which is the reason why I came to this forum to seek advice.

And, yes, I have begun adding some phrases to a subject-specific glossary I am building now.

[Edited at 2011-03-20 11:15 GMT]


 

Lori Cirefice  Identity Verified
France
Local time: 02:42
French to English
Ah Mar 20, 2011

Stefan Stefanov wrote:


At first I paid little attention to the huge number of repetitions claimed because I had only had a brief look at the file to be translated - and that's a HUGE file - and I expected that there are many repeated segments. I realized that was not the case when I started working - indeed many phrases and words are repeated but in differing contexts and WF does not automatically supply the translation, which is the reason why I came to this forum to seek advice.



If you are getting the same analysis results as your client, that is reassuring. Those repetitions *must* be in your file and you should be able to leverage them. Maybe you just haven't gotten to them yet?


 

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
TOPIC STARTER
Repetitions... Mar 20, 2011

Lori Cirefice wrote:

Stefan Stefanov wrote:


At first I paid little attention to the huge number of repetitions claimed because I had only had a brief look at the file to be translated - and that's a HUGE file - and I expected that there are many repeated segments. I realized that was not the case when I started working - indeed many phrases and words are repeated but in differing contexts and WF does not automatically supply the translation, which is the reason why I came to this forum to seek advice.



If you are getting the same analysis results as your client, that is reassuring. Those repetitions *must* be in your file and you should be able to leverage them. Maybe you just haven't gotten to them yet?


icon_smile.gif I can't tell where these repetitions are, if they are there in fact. I mean, there are many repeated words, but in varying contexts, and since WF doesn't supply them, I still think this number of repeated words means nothing...


 

Lori Cirefice  Identity Verified
France
Local time: 02:42
French to English
tricky wording Mar 20, 2011

Although the WF analysis says "words", really it means "words that are in full segments". I really think that you just haven't gotten to the right place yet, hopefully soon you will start getting some 100% matches.

You could try analysing files with another tool in order to reassure yourself, but I think you are just misunderstanding what the analysis report means.


 

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
TOPIC STARTER
Hope you're right Mar 20, 2011

Lori Cirefice wrote:

Although the WF analysis says "words", really it means "words that are in full segments". I really think that you just haven't gotten to the right place yet, hopefully soon you will start getting some 100% matches.

You could try analysing files with another tool in order to reassure yourself, but I think you are just misunderstanding what the analysis report means.


Thanks. I hope you are right, because right now I am feeling very scr***d.


 

B D Finch  Identity Verified
France
Local time: 02:42
Member (2006)
French to English
+ ...
Repetitions should never be free Mar 20, 2011

Even segemnts that are repetitions or 100% matches need checking and processing. Sometimes they turn out not to be 100% matches because the TM translation is bad. I always charge one third of my rate for the translation in question for repetitions or 100% matches. For that, I review each occurrence and amend or suggest amendments as appropriate. The way a CAT tool segments the source document can lead to false positives for repetitions.

 

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
TOPIC STARTER
Yes Mar 20, 2011

B D Finch wrote:

Even segemnts that are repetitions or 100% matches need checking and processing. Sometimes they turn out not to be 100% matches because the TM translation is bad. I always charge one third of my rate for the translation in question for repetitions or 100% matches. For that, I review each occurrence and amend or suggest amendments as appropriate. The way a CAT tool segments the source document can lead to false positives for repetitions.



Yes, you're right. Even 100% TM matches need to be reviewed.

The problem is that the agency and I had made an agreement - I gave them 25% discount for volume and we agreed on a price and I confirmed taking the order.

Only when they sent me the PO, they said that there were repetitions and they aren't paying for them (I think you'll agree that's wrong - they should have said this before I accepted the order). That's why I feel scr***d. And I have contacted them, saying I am dissatisfied with this - their answer will probably come tomorrow.

[Edited at 2011-03-20 12:24 GMT]

[Edited at 2011-03-20 12:24 GMT]


 

René Stranz-Nikitin  Identity Verified
Czech Republic
Local time: 02:42
Czech to German
+ ...
You are right to feel scr***d. Mar 20, 2011

Stefan Stefanov wrote:

B D Finch wrote:

Even segemnts that are repetitions or 100% matches need checking and processing. Sometimes they turn out not to be 100% matches because the TM translation is bad. I always charge one third of my rate for the translation in question for repetitions or 100% matches. For that, I review each occurrence and amend or suggest amendments as appropriate. The way a CAT tool segments the source document can lead to false positives for repetitions.



Yes, you're right. Even 100% TM matches need to be reviewed.

The problem is that the agency and I had made an agreement - I gave them 25% discount for volume and we agreed on a price and I confirmed taking the order.

Only when they sent me the PO, they said that there were repetitions and they aren't paying for them (I think you'll agree that's wrong - they should have said this before I accepted the order). That's why I feel scr***d. And I have contacted them, saying I am dissatisfied with this - their answer will probably come tomorrow.

[Edited at 2011-03-20 12:24 GMT]

[Edited at 2011-03-20 12:24 GMT]


I really sympathize with you. It seems absolutely illogical to apply two discounts on the same job. You gave a flat discount of 25% for the whole job and after that they try to squeeze out another discount. They have to decide one way or the other: Either they use the flat discount or they apply a discount for the repetitions. They can't get both! If they insist I would turn down the job.

But anyway, from me they would never get the repeated segments for less than 30% of the full per word rate for the reasons mentioned by B D Finch.

This agency seems to be a cheapskate.

Good luck in this fight!

René Stranz-Nikitin


 

Samuel Murray  Identity Verified
Netherlands
Local time: 02:42
Member (2006)
English to Afrikaans
+ ...
So, what is your volume discount then? Mar 20, 2011

Stefan Stefanov wrote:
The problem is that the agency and I had made an agreement - I gave them 25% discount for volume and we agreed on a price and I confirmed taking the order. ... Only when they sent me the PO, they said that there were repetitions and they aren't paying for them...


So you gave 25% discount for volume of 105 000 words. Would you have given a volume discount for 62 000 words? If so, how much? Then decide whether you want to risk losing the client if you keep firm on this, and if so, then: tell the client that your 25% discount was based 105 000 words, not on 62 000 words, and tell him what your discount for 62 000 words would be (if any).

Whether the client wants to pay for repetitions, and whether you would be willing to accept no-pay for repetitions, is a separate issue. Your name won't appear next to the translation.


 

Stefan Stefanov  Identity Verified
Bulgaria
Local time: 03:42
English to Bulgarian
+ ...
TOPIC STARTER
Discount Mar 20, 2011

Hi Samuel and Rene,

You're both right and that's my opinion too.

I've already made a discount that's big enough, and I'm not making a second one.

I can afford to turn down this job if the agency insist to get both.

Thanks for your answers. It's good to know others think the same way as youicon_smile.gif


 
Pages in topic:   [1 2] >


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


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

What is the correct number of repetitions?

Advanced search


Translation news related to Wordfast





Protemos translation business management system
Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!

The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.

More info »
Déjà Vu X3
Try it, Love it

Find out why Déjà Vu is today the most flexible, customizable and user-friendly tool on the market. See the brand new features in action: *Completely redesigned user interface *Live Preview *Inline spell checking *Inline

More info »



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