Pages in topic:   [1 2] >
Use a real percentage comparison to choose the right translator
Thread poster: Felipe Gútiez

Felipe Gútiez  Identity Verified
Germany
Local time: 07:43
German to Spanish
+ ...
May 30, 2012

I have an idea.
What about having a software online which makes possible following:

-You have a TM or several TMs with all the Translations you have done in your life
-You want to show that you are THE BEST for the Translation that someone needs
-You want to preserve the discrection and confidentiality regarding your old customers

How can you show that "really" you have experience in that area because you have "done" lots of translations, and at the same time, preserve confidentiality?

There should be a software which permits to show how good you are and the similarity of the translations that you have done with respect to the Texts someone wants you to translate and all this preserving the confidentiality of your customers.

The software should give a kind of log which tells you:
-Keywords similarity (for example very specialized terminology)
-general similarity


How to know that the TMs that you are introducing in the Software are really translations that you have done? Because they are signed from you and because you tell the software that they have been done by you. It is also a matter of trust. Maybe there is some way to even improve the way to avoid fake.

What do you think of this idea?
Do you think this could be a good business?
Do you think business will trust this software?

If people use online Banking, why not use TM comparing?

Hope you like the idea.

Maybe this exists already. Has someone knowledge of something of the art?


Direct link Reply with quote
 

Dr. Andrew Frankland  Identity Verified
Spain
Local time: 07:43
Member (2007)
English
+ ...
Shooting yourself in the foot May 30, 2012

The way I understand your proposal is that you want to tell the customer how many repetitions arise from your TMs in order to convince them that you're the best person for the job. Surely they would then ask for a discount based on those repetitions and you'd earn less. Or maybe I misunderstood.....

Andy


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 07:43
Member (2006)
English to Afrikaans
+ ...
Could work, if offline May 30, 2012

Felipe Gútiez wrote:
There should be a software which permits to show how good you are and the similarity of the translations that you have done with respect to the Texts someone wants you to translate and all this preserving the confidentiality of your customers.


The way I see this working is that there would be a program that you would install on your own computer. You'd tell that program where all your TMs are. It would then read your TMs and create a profile of the subject fields that you have worked in so far (this profile could be in the form of a customisable report, which may be useful for the translator to get an overview of his work). You can then supply this profile (in the export format of that program) to potential clients who are aware of the system. If the profile contains graphic graphs, you could add it to your web site, for example.

Alternatively, a client could feed his source text into his installation of the program, which produces a keyword profile for the source text. The client sends you the source text profile and then your program compares it to the TMs that you have on your computer, to determine a broad match percentage (based on subject-specific keywords and possibly short subject-specific phrases).

Such a system would be open to abuse, but if it is simple to use then I think most translators will not abuse it.

I don't think an online system would work because then translators would have to upload their TMs (and do it regularly, too). I would have no objection to using such a program if it runs on my own computer (though it would have to be able to process WFC TMs).


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 07:43
Member (2006)
English to Afrikaans
+ ...
Keyword matching, not segment matching May 30, 2012

Dr. Andrew Frankland wrote:
The way I understand your proposal is that you want to tell the customer how many repetitions arise from your TMs in order to convince them that you're the best person for the job. Surely they would then ask for a discount based on those repetitions and you'd earn less.


The way I understand this system is that it shows the client how experienced you are in a subject field, by telling the client what percentage of your TMs (globally) match the keywords in his source text. Since this is keyword matching (not segment matching), no discounts can be calculated from it.

The OP's "general similarity" may include segment matching, but even so I think the report generated by the program should not make discount calculation possible.

After all, the fact that a client's text contains many segments that have high fuzzy matches with many of my TMs does not mean that I will be using those TMs when translating the client's text (in fact, I don't mix TMs from different clients at all). The fact that I have many matching segments should tell the client that I have translated many similar segments before, but it would not affect the speed with which I'll translate his text.





[Edited at 2012-05-30 12:26 GMT]


Direct link Reply with quote
 

Katalin Horváth McClure  Identity Verified
United States
Local time: 01:43
Member (2002)
English to Hungarian
+ ...
Why would it prove anything about quality? May 30, 2012

Even if a piece software said that a translator had many segments with specific keywords in his TM or TMs, what would it mean in terms of the quality of the translation? Nothing, IMHO.
As far as I understand, the comparison would start with the new source text, and would compare it with source segments in the TM...


Direct link Reply with quote
 

Nicole Schnell  Identity Verified
United States
Local time: 22:43
English to German
+ ...
With Katalin May 31, 2012

If you are not capable of describing any machines or technologies that you are highly familiar with in your own words and without naming the manufacturer, then there is something wrong with the picture.

Direct link Reply with quote
 

Phil Hand  Identity Verified
China
Local time: 14:43
Chinese to English
Quality is always unknowable May 31, 2012

And experience is a good proxy for quality, so I can see the value in Felipe's suggestion.

You get an objective measure of experience without breaching confidentiality. And basing it on TMs makes it more likely (not a guarantee, but more likely) that the information translators submit is real, because TM formats aren't in wide circulation. The alternative would be directly submitting documents you've translated, but that way anyone could feed in any old documents.

My problem would be that I don't do all of my translation through CAT, so I'd lose out a bit. But if this was introduced as another metric which clients & outsourcers can look at among others to assess a translator, I don't see that it would be a bad thing. Probably some big outsourcers already have similar kinds of systems internally.


Direct link Reply with quote
 

Philippe Etienne  Identity Verified
Spain
Local time: 07:43
Member
English to French
Calls for tenders from agencies May 31, 2012

Some public RFPs request to specify translation experience in a certain field by the number of pages translated in that field.

Felipe's idea may stem from such enquiries we sometimes get from agencies.

Personally, I haven't the slightest idea of how many pages (or words or whatever) I have translated in a specific field. I keep only the end client name to categorise jobs (or TMs)

Philippe


Direct link Reply with quote
 

Miguel Carmona  Identity Verified
United States
Local time: 22:43
English to Spanish
What about TMs not created entirely by the translator? May 31, 2012

How would the system account for all the TMs that you actually use, but were not built entirely by you?

Sometimes you receive from a client a huge TM containing previous translations, and only from that moment on you start contributing to the TM.

Just because you have the TM (to which you only contributed partially) would you consider it in its entirety "part of your experience"?

The system would need to have that in consideration.


Direct link Reply with quote
 

Vikki Pendleton  Identity Verified
United Kingdom
Local time: 06:43
Member
German to English
+ ...
What about new translators? Jun 1, 2012

That sort of system would effectively rule out new translators from ever getting any jobs. It's hard enough as it is.

Direct link Reply with quote
 

Felipe Gútiez  Identity Verified
Germany
Local time: 07:43
German to Spanish
+ ...
TOPIC STARTER
Sort out by author Jun 6, 2012

Miguel Carmona wrote:

How would the system account for all the TMs that you actually use, but were not built entirely by you?

Sometimes you receive from a client a huge TM containing previous translations, and only from that moment on you start contributing to the TM.

Just because you have the TM (to which you only contributed partially) would you consider it in its entirety "part of your experience"?

The system would need to have that in consideration.


Hi Miguel,

I think you can always filter by your name and sort your own translations out. At least with most usual software like Trados and co.


Direct link Reply with quote
 

Felipe Gútiez  Identity Verified
Germany
Local time: 07:43
German to Spanish
+ ...
TOPIC STARTER
That is absolutely true Jun 6, 2012

Vikki Pendleton wrote:

That sort of system would effectively rule out new translators from ever getting any jobs. It's hard enough as it is.


To be honest, I would not recommend my children to become translators.
It is most probable that in a few years people will use only machine translation. That´s it.
But, absolutely, with such a system a new translator would have more difficulties.


Direct link Reply with quote
 

Christine Andersen  Identity Verified
Denmark
Local time: 07:43
Member (2003)
Danish to English
+ ...
Comparing translators with MT Jun 6, 2012

I agree with Katalin and Nicole.

I think this is a dangerous way to go. Even in technical translation, you can't assume that the same segment should be translated identically every time.
Language is not that simple, or they would have cracked the MT challenge half a century ago.

I came across quite a funny error in a website yesterday - caused by machine translation.

The English sentence was obviously 'Symptoms may disappear after a few hours or last years.'

The back translation from the Danish would have been something like:
'Symptoms may disappear after a few hours or the recent years.'

It was otherwise quite a fluent-sounding text. But once I caught that sentence, I began to wonder about several of the statements on the website, and although it was natural and correct Danish, it was not entirely coherent - some of the statements simply did not fit together.

Your TM-tool would perhaps give a high score for a lot of matches, but I would not call that website a quality translation. I did not have the English source, and some of the possibly mistranslated sentences were not as transparent as that one.

I have just translated a similar site sentence by sentence, and actually altered a lot of segments that came up from the TM, because in the new context they did not fit precisely.

The machines still cannot be trusted with the critical differences.

I would alomost regard a LOW level of identical matches as an indication of quality... But in fact I think this sort of statistical test could be very misleading.
________________________

I think the future for translators lies in showing that there may be occasions when MT is adequate, but that there will always be a need for real human intelligence that can sort the homonyms and find the tiny nuances. There will always be some critical differences that cannot be reduced to algorithms and formulae.

Happy translating!


[Edited at 2012-06-06 08:15 GMT]


Direct link Reply with quote
 

Rolf Keller
Germany
Local time: 07:43
English to German
Translation is not about words ... Jun 7, 2012

Felipe Gútiez wrote:

How can you show that "really" you have experience in that area because you have "done" lots of translations (...)?


Experience is neither a suffient nor a necessary prerequisite for quality. It's not even a good indicator - just look at KudoZ ...

Regarding terms:
Translation is not about words. It's about what the words are about.


Direct link Reply with quote
 

Rolf Keller
Germany
Local time: 07:43
English to German
Just ask your neighbour's teenage son about data manipulation ;-) Jun 7, 2012

Felipe Gútiez wrote:

I think you can always filter by your name and sort your own translations out. At least with most usual software like Trados and co.


You can do the opposite as well, i. e. change all the names to your name. Automatically.


Direct link Reply with quote
 
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 »

Use a real percentage comparison to choose the right translator

Advanced search






WordFinder
The words you want Anywhere, Anytime

WordFinder is the market's fastest and easiest way of finding the right word, term, translation or synonym in one or more dictionaries. In our assortment you can choose among more than 120 dictionaries in 15 languages from leading publishers.

More info »
SDL MultiTerm 2017
Guarantee a unified, consistent and high-quality translation with terminology software by the industry leaders.

SDL MultiTerm 2017 allows translators to create one central location to store and manage multilingual terminology, and with SDL MultiTerm Extract 2017 you can automatically create term lists from your existing documentation to save time.

More info »



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