A question on weighted words Thread poster: Filip Sibl
|
|---|
Hi everyone,
As a new Trados user, I have a question concerning weighted words. How does a client check for the number of complete or fuzzy matched words in a project? I read somewhere that we should never send our translation memories along with finished translations (them being gold was mentioned). Does that mean that the client has to trust our calculation, or is there a better method?
| | | | | A matter of trust | Nov 29, 2014 |
Filip Sibl wrote:
Does that mean that the client has to trust our calculation...
If my fairness/honesty were challenged that way, I'd be tempted to reply that if they don't trust my wordcount, then I would apply full rate for all words, the count of which is usually easily confirmed in a word processor, using a wordcount utility or counting them manually.
But this shouldn't be an issue:
1) The agency provides you their TM and both them and you can run the analysis on the files (with the same TM options enabled: penalties, internal concordance, tag handling, etc.) to check each other's results
2) They don't provide a TM and you both run the analysis on an empty TM. You then get only reps and no-matches. And fuzzies if you enable this further rate-lowering option of homogeneity/internal concordance/whatever it's called
3) You handle TM management on behalf of the customer and if they trust you for that, they should trust your analyses
Anyway, trust should always prevail in a sound customer/provider relation.
Philippe
| | | | Neil Coffey 
United Kingdom
Local time: 09:52
French to English
+ ...
| If the client isn't supplying the TM... | Nov 29, 2014 |
... then what's the rationale for them wanting a discount based on a TM that they didn't supply? (Other than they're trying to scam you, of course.)
[Edited at 2014-11-29 14:29 GMT]
| | | | Filip Sibl
Local time: 10:52
English to Croatian
TOPIC STARTER | Thanks a lot | Nov 29, 2014 |
for clearing that up guys. It makes sense that client base would be used, yet in my noobishness that never crossed my mind. So much to learn about CATs...
| | |
|
|
|
Gerard de Noord
France
Local time: 10:52
Member (2003)
English to Dutch
+ ...
| Discount for expected leverage | Nov 30, 2014 |
Hi Filip,
So you did understand Philippe? Generally speaking, there is no TM and trust doesn't come into play:
1. PMs analyze the project against an empty TM and weigh the 100% matches and repetitions to obtain a discount.
2. PMs push an extra button to calculate internal fuzzy matches too and weigh the 100% matches and repetitions as well as the virtual fuzzy matches to obtain an even larger discount. This ... See more Hi Filip,
So you did understand Philippe? Generally speaking, there is no TM and trust doesn't come into play:
1. PMs analyze the project against an empty TM and weigh the 100% matches and repetitions to obtain a discount.
2. PMs push an extra button to calculate internal fuzzy matches too and weigh the 100% matches and repetitions as well as the virtual fuzzy matches to obtain an even larger discount. This is explained very well in this video:
http://www.youtube.com/watch?v=dKgmJb2Eupw
The good thing is that you can do the same analysis in e.g. Studio 2014 too, to check the PMs' word count.
Cheers,
Gerard
[Edited at 2014-11-30 17:40 GMT] ▲ Collapse
| | | | Filip Sibl
Local time: 10:52
English to Croatian
TOPIC STARTER | Think i got it | Nov 30, 2014 |
Thank you very much for the explanation Gerard. Now the process makes sense.
Since I didn't know about the filters needed to display internal fuzzies, my test project reports always found zero matches. That led me to believe discounts depended on stored TM bases and not just on repetitions inside a specific project.
| | | | Samuel Murray
Netherlands
Local time: 10:52
Member (2006)
English to Afrikaans
+ ...
| I agree with most others here | Dec 1, 2014 |
Philippe Etienne wrote:
1) The agency provides you their TM and both them and you can run the analysis on the files (with the same TM options enabled: penalties, internal concordance, tag handling, etc.) to check each other's results
2) They don't provide a TM and you both run the analysis on an empty TM. You then get only reps and no-matches. And fuzzies if you enable this further rate-lowering option of homogeneity/internal concordance/whatever it's called
3) You handle TM management on behalf of the customer and if they trust you for that, they should trust your analyses
4) You create your own TM for that subject matter, from other sources than the client provide, and you run an analysis against your own TM for your own benefit, without telling the client. This is not dishonest -- the client simply can't expect a discount from materials that he didn't provide. Giving a discount for matches from a translator's personal TM would be akin to giving a discount for word matches from the translator's own dictionary.
| | | | | It is normal to return a TM of the translated files | Dec 1, 2014 |
but ONLY the translated files.
In Trados Studio it is quite easy to filter off segments used after a particular date and export them from your TM. That way you can filter off only the segments used or changed for that particular translation and send them to the client.
You can send them the XLIFF files. In the earlier versions it was very normal to send the bilingual files.
The assumption is that if the translator does not send a TM extract or bilingual fil... See more but ONLY the translated files.
In Trados Studio it is quite easy to filter off segments used after a particular date and export them from your TM. That way you can filter off only the segments used or changed for that particular translation and send them to the client.
You can send them the XLIFF files. In the earlier versions it was very normal to send the bilingual files.
The assumption is that if the translator does not send a TM extract or bilingual files, the client can align the source and target and obtain the files they want anyway.
But never send your own main TMs. Here you are right, you are handing over gold - and very possibly other clients' confidential information.
When you make a TM extract, you can see how many segments are included.
The number should correspond to the segments in your text or be lower (repeats will not be counted in the TM, and segments that are numbers only will not be included).
You can also create a TM for that text or for that client only, and send that to the client.
If you do, then updated versions of work you have translated earlier will come up largely as repeats. This is where you really have to compare the amount of work you actually do, and what the policy is on matches and near matches.
If the cilent will not pay you to check that they still fit in the new context, then you are in principle entitled to ignore them, but you may do so at your peril!
I often find it necessary to revise usage of proper nouns and pronouns.
Danish loves to plaster the name of the company or some other main topic at the beginning of the sentence.
SuperDan is your best choice of widgets.
SuperDan will provide the best service.
SuperDan tailors everything to your needs.
SuperDan are also cheapest ...
... might be more or less acceptable in Danish, but in English you would run a couple of sentences together, or perhaps use 'we' or some other pronoun to shift the focus from the company's name to what they are actually selling!
SuperDan is your best choice and will provide the best service.
Our widgets are tailored to your needs, so they are often cheapest too ...
The original segments might be completely reshuffled in the source, and even when they come up as 100% matches, you have to decide whether to make the same changes as last time or whether you need to do something different.
And how you are going to be paid...
Insisting on the empty TM for word counts is often your best policy.
[Edited at 2014-12-01 17:45 GMT] ▲ Collapse
| | |
|
|
|
Filip Sibl
Local time: 10:52
English to Croatian
TOPIC STARTER
Let me just thank you all again for the info. I am amazed at how friendly and helpful ProZ forum community is. Hope I will have a chance to contribute as well and help answer other members questions. (this comes from a habitual lurker on many other forums, that usually types down a post and then promptly deletes it)
| | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » A question on weighted words | 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
Buy now! » |
| | 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 » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news
