Which CAT calculates fuzzies within one batch of files?
Thread poster: Jan Sundström

Jan Sundström  Identity Verified
Sweden
Local time: 10:17
English to Swedish
+ ...
May 10, 2011

Hi all,

Scenario: a client produces a lot of very similar models (laptops, printers, mobile phones etc) where only some tech specs like CPU speed, RAM size etc differ. They will send out a batch of files with descriptions of various model configurations.
When a new range is launched, there will be a lot of descriptive text which are almost identical, the only difference being the model number and tech specs occuring within the sentences.

When doing a standard Trados analysis on the batch there will only be a handful of fuzzies against the existing TM, since the product range is new. But the client pointed out that there should be a lot of "internal fuzzies" since the descriptions within the batch are similar.

Is there any CAT tool that takes these "internal fuzzies" into account?
Or is there a feature/setting in Trados that enables this?

Thanks a lot for your input,

/Jan


Direct link Reply with quote
 

Sergei Leshchinsky  Identity Verified
Ukraine
Local time: 11:17
Member (2008)
English to Russian
+ ...
join May 10, 2011

Join the files into one package/project and then analyze the project.
Trados can do it through Glue (merging all the files into ONE TTX).

[Редактировалось 2011-05-10 08:51 GMT]


Direct link Reply with quote
 

Jabberwock  Identity Verified
Poland
Local time: 10:17
Member (2004)
English to Polish
MemoQ May 10, 2011

Sergei Leshchinsky wrote:
Join the files into one package/project and then analyze the project.
Trados can do it through Glue (merging all the files into ONE TTX).


This does not make any difference - if you analyze all the files at once or if you glue them together you get the same results. For old Trados the only solution would be to analyze the files one by one with the option to remember the previous analysis results, but of course this is impractical.

MemoQ does it exactly as you would want it - you just have to turn on the "Homegeneity" option in the Statistics window.

Of course, this will reduce the weighted wordcount of the project, so make sure you get to profit from that arrangement


Direct link Reply with quote
 

Hermann Bruns  Identity Verified
Local time: 10:17
English to German
MetaTexis May 10, 2011

Hello Jan,

in MetaTexis you can do such an analysis with the Analysis or with the Batch function. Just add the batch of files to be analyzed, define the TMs, and run the Analysis.

Kind regards
Hermann


Direct link Reply with quote
 

Philippe Etienne  Identity Verified
Spain
Local time: 10:17
Member
English to French
Saving money to the last drop May 10, 2011

SDLX 2005 has such an option, where you can calculate the fuzzies-to-be within a text.
Just like the Homogeneity option in MemoQ, Trados Studio 2009 features such option: Options>Language pairs>all language pairs>batch processing>file analysis>option "Display recycling level of internal analogies" (or similar)
To my knowledge, Trados 2007 (Workbench) didn't have such feature.

With this option enabled, agencies who order weighted wordcounts make certain that the translator gets NO financial/speed benefit whatsoever from using a CAT tool, bought most of the time without any subsidy from them.

However, one of the rare joys of using a CAT tool, while being paid on a weighted wordcount basis, is precisely to find out on the fly that a sentence you just translated is propagated as fuzzies all over the place. Every time it happens, it brightens my day. Why should I be deprived of that?

By experience, no agency tells you before you state your standard rate that they take into account internal fuzzies to work out the wordcount, until you realise that your analysis and theirs don't match (in the agency's favour, of course)...

Philippe


Direct link Reply with quote
 

Selcuk Akyuz  Identity Verified
Turkey
Local time: 11:17
Member (2006)
English to Turkish
+ ...
Homogeneity option in MemoQ May 10, 2011

Philippe Etienne wrote:

SDLX 2005 has such an option, where you can calculate the fuzzies-to-be within a text.
Just like the Homogeneity option in MemoQ, Trados Studio 2009 features such option: Options>Language pairs>all language pairs>batch processing>file analysis>option "Display recycling level of internal analogies" (or similar)
To my knowledge, Trados 2007 (Workbench) didn't have such feature.

With this option enabled, agencies who order weighted wordcounts make certain that the translator gets NO financial/speed benefit whatsoever from using a CAT tool, bought most of the time without any subsidy from them.

However, one of the rare joys of using a CAT tool, while being paid on a weighted wordcount basis, is precisely to find out on the fly that a sentence you just translated is propagated as fuzzies all over the place. Every time it happens, it brightens my day. Why should I be deprived of that?

By experience, no agency tells you before you state your standard rate that they take into account internal fuzzies to work out the wordcount, until you realise that your analysis and theirs don't match (in the agency's favour, of course)...

Philippe


That is exactly what happened in my first MemoQ project, of course I asked the Agency to send a new PO, which was 30% higher than the first one! Caution, translating MemoQ projects can result in lower income.


[Edited at 2011-05-10 15:10 GMT]


Direct link Reply with quote
 

Jan Sundström  Identity Verified
Sweden
Local time: 10:17
English to Swedish
+ ...
TOPIC STARTER
Thanks all! May 10, 2011

Philippe Etienne wrote:
With this option enabled, agencies who order weighted wordcounts make certain that the translator gets NO financial/speed benefit whatsoever from using a CAT tool...


Thanks all, this was a very interesting discussion!
Just trying to be the devil's advocate here. With the increasing awareness of the end client, agencies are getting squeezed too. Big corporate clients are finding out about this feature, and demand that the word logs take internal fuzzies into account.

Of course, the agency should be fair when using Homogeneity, there's no reason to be sneaky - I'm sure it's something that we will see much more of in the future. With the added awareness among freelancers, it's only reasonable to raise your rates accordingly too.

[Edited at 2011-05-11 06:32 GMT]


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 10:17
English
Old Trados and this feature May 11, 2011

Jabberwock wrote:
For old Trados the only solution would be to analyze the files one by one with the option to remember the previous analysis results, but of course this is impractical.

Philippe Etienne wrote:
To my knowledge, Trados 2007 (Workbench) didn't have such feature.


Just to add to this discussion about something which is not really new... and this is just for interest really. In SDLX it is was always there and always easy to use. But a less well known fact is that in old Trados it was there too, although less easy to access and use.

It was called the use previous TM feature in Workbench and was used to simulate the availability of a TM, obtained through alignment or partial translation of a project. It was quite an obscure feature and whilst it is theoretically possible to use it to compute the project-internal leverage it wasn't documented well.

The idea behind it (which in some ways is an extension of the homogeneity idea (if you like this terminology) already in SDLX back then) is that you’d “simulate” a full or partial translation of a project by running the next analysis against the temporary pseudo-TM created by the last run. So by extension I mean that this provided for the ability to split up your files and use “from previous analysis” TMs to optimize translation order, i.e. maximize internal leverage even if the project was split and no shared TM would be used during project execution.

In Studio it's easier to access and use if you want this type of analysis.

Regards

Paul

[Edited at 2011-05-11 08:30 GMT]


Direct link Reply with quote
 

Philippe Etienne  Identity Verified
Spain
Local time: 10:17
Member
English to French
file-level leverage vs. segment-level leverage Sep 6, 2011

SDL Support wrote:
...It was called the use previous TM feature in Workbench and was used to simulate the availability of a TM, obtained through alignment or partial translation of a project. It was quite an obscure feature and whilst it is theoretically possible to use it to compute the project-internal leverage it wasn't documented well.
...

Hi Paul,

As far as I know, the "use TM from previous analysis" option in old Trados didn't work out intra-file matches like the options available in various CAT tools today: in a one-file project, there was no way to anticipate the internal leverage, and any internal match within the file was then to the translator's benefit. The "Use TM from previous analysis" option worked at file level, not at segment level. This is why I wrote that Trados didn't have the internal match counting feature.

Now in Jan's case the "use previous TM" option could be of some limited use, provided you select first the right files that give the most leverage. Still it would require far more guesswork than with the old SDLX and newer CAT tools, for a far less accurate view.

Internal fuzzy matches are now added up segment after segment within a same file, just like it used to be done in old SDLX.

Regards,
Philippe


Direct link Reply with quote
 

sentrodil
Turkey
Local time: 11:17
just what I thought! Sep 7, 2011

Philippe Etienne wrote:

With this option enabled, agencies who order weighted wordcounts make certain that the translator gets NO financial/speed benefit whatsoever from using a CAT tool, bought most of the time without any subsidy from them.

However, one of the rare joys of using a CAT tool, while being paid on a weighted wordcount basis, is precisely to find out on the fly that a sentence you just translated is propagated as fuzzies all over the place. Every time it happens, it brightens my day. Why should I be deprived of that?

By experience, no agency tells you before you state your standard rate that they take into account internal fuzzies to work out the wordcount, until you realise that your analysis and theirs don't match (in the agency's favour, of course)...

Philippe


what is your response to such agencies?


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 »

Which CAT calculates fuzzies within one batch of files?

Advanced search







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 »
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

More info »



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