Concordance search completely useless?
Thread poster: Heinrich Pesch

Heinrich Pesch  Identity Verified
Finland
Local time: 04:26
Member (2003)
Finnish to German
+ ...
Feb 14, 2013

I just noticed that in concordance search (f3) the software does not function properly. My search was for "Auslagern". The system presents fuzzy matches like "Auslösen" and "Auflagen".
This is completely useless. How can I change the settings so that only 100% matches are offered, like it was standard in Trados and Wordfast?


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 04:26
Finnish to French
Old Trados also used fuzziness for concordance Feb 14, 2013

Heinrich Pesch wrote:
I just noticed that in concordance search (f3) the software does not function properly. My search was for "Auslagern". The system presents fuzzy matches like "Auslösen" and "Auflagen".
This is completely useless. How can I change the settings so that only 100% matches are offered, like it was standard in Trados and Wordfast?

The "old" Trados also had problems with concordance search in languages like Finnish or German: it would fail to return useful matches or return useless matches, sometimes both. It relied on fuzziness, which you can adjust, but whatever level you set it at, you will still have the aforementioned problems. Concordance search in Wordfast and memoQ is much better for languages like Finnish and German (which both work in a similar way with compound words), and especially Finnish (because of declensions and other endings).


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 03:26
English
Settings Feb 14, 2013

Dear Heinrich,

Yes, you can change settings for this:



If you change this to 100% it will only find instances of "Auslagern".

Please be aware that there are multiple places to change this, for the current project to use them, change it in Project Settings.

Cheers,
Luis

[Edited at 2013-02-14 09:59 GMT]


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:26
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
How about Options? Feb 14, 2013

SDL Support wrote:

Dear Heinrich,

Yes, you can change settings for this:



If you change this to 100% it will only find instances of "Auslagern".

Please be aware that there are multiple places to change this, for the current project to use them, change it in Project Settings.

Cheers,
Luis

[Edited at 2013-02-14 09:59 GMT]


Thanks, Support! I would like to change this in the general settings. I cannot think of any project where I would need less than 100% for concordance search.


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 04:26
Finnish to French
Finnish? Feb 14, 2013

Heinrich Pesch wrote:
I cannot think of any project where I would need less than 100% for concordance search.

Don't you search for concordance search in Finnish texts?! How would you find anything if you set fuzziness to 100%?


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 03:26
English
yes Feb 14, 2013

Heinrich Pesch wrote:

SDL Support wrote:

Dear Heinrich,

Yes, you can change settings for this:



If you change this to 100% it will only find instances of "Auslagern".

Please be aware that there are multiple places to change this, for the current project to use them, change it in Project Settings.

Cheers,
Luis

[Edited at 2013-02-14 09:59 GMT]


Thanks, Support! I would like to change this in the general settings. I cannot think of any project where I would need less than 100% for concordance search.


Hi Heinrich,

Yes, you can also set it in Options, it will then apply for all future projects. Unless you use Project Templates, then you would also have to adjust the Project Template Settings through File->Setup->Project Templates

Cheers,
Luis


Direct link Reply with quote
 

Heinrich Pesch  Identity Verified
Finland
Local time: 04:26
Member (2003)
Finnish to German
+ ...
TOPIC STARTER
I select only parts of words or sentences Feb 14, 2013

Dominique Pivard wrote:

Heinrich Pesch wrote:
I cannot think of any project where I would need less than 100% for concordance search.

Don't you search for concordance search in Finnish texts?! How would you find anything if you set fuzziness to 100%?


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 04:26
Finnish to French
selecting parts of words with fuzziness set to 100% won't work Feb 14, 2013

Heinrich Pesch wrote:
I select only parts of words or sentences

... but then you won't find anything if you set fuzziness to 100%, which is what Paul suggested if you want to find exact matches only.


Direct link Reply with quote
 

Grzegorz Gryc  Identity Verified
Local time: 03:26
French to Polish
+ ...
Character based concordance search Feb 14, 2013

Heinrich Pesch wrote:

I select only parts of words or sentences


If you select only part of words, you'll get no hits at all for the default TM settings unless the selected part of word corresponds to an entire word already present in your TM.
You should enable character based concordance search when creating your TMs (it can't be done for existing TMs).
Nonetheless, it may slow down Studio for huge TMs.

The 100% setting should work if the character based concordance search is enabled.

BTW, I think the (character based) concordance algorithms are far worse in Studio than in Trados i.e. they return far more junk.

Cheers
GG

[Edited at 2013-02-14 10:44 GMT]


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 04:26
Finnish to French
character-based concordance search Feb 14, 2013

Grzegorz Gryc wrote:
The 100% setting should work if the character based concordance search is enabled.

BTW, I think the (character based) concordance algorithms are far worse in Studio than in Trados i.e. they return far more junk.

Thanks for bringing my attention to character-based concordance search, which I had overlooked. Explanations about it can be found here:

http://producthelp.sdl.com/SDL_Trados_Studio_2011/client_en/Edit_View/Overview__Concordance_Window.htm

It's the default, so chances are Heinrich's existing TM uses it. Nevertheless, even with TM's using it, you will get lots of false positives when searching for compound words in languages like Finnish (and probably German too), and then Studio won't find words that are in the TM, although they only differ by 1 or 2 characters (talking about words with at least 10-15 characters here) from the form used to search them. I haven't found Workbench much better than Studio in that respect.


Direct link Reply with quote
 

Grzegorz Gryc  Identity Verified
Local time: 03:26
French to Polish
+ ...
Settings and algorithm changes... Feb 14, 2013

Dominique Pivard wrote:

Grzegorz Gryc wrote:
The 100% setting should work if the character based concordance search is enabled.

BTW, I think the (character based) concordance algorithms are far worse in Studio than in Trados i.e. they return far more junk.

Thanks for bringing my attention to character-based concordance search, which I had overlooked. Explanations about it can be found here:

http://producthelp.sdl.com/SDL_Trados_Studio_2011/client_en/Edit_View/Overview__Concordance_Window.htm

It's the default, so chances are Heinrich's existing TM uses it.

Ooops, me too, I get corrected
It's indeed the default in the current Studio 2011 builds but it was not in Studio 2009.
This setting was stupid for most languages I'm aware, it was suitable probably only for English.

Nevertheless, even with TM's using it, you will get lots of false positives when searching for compound words in languages like Finnish (and probably German too), and then Studio won't find words that are in the TM, although they only differ by 1 or 2 characters (talking about words with at least 10-15 characters here) from the form used to search them. I haven't found Workbench much better than Studio in that respect.

It was, IMO.
I have a very poor idea of Finnish/Hungarian etc. but in these languages the internal inflection is not very frequent unlike e.g. in Polish and other Slavic languages (in some extent, also in Germanic ones, especially in German).
For Polish, Studio throws completely useless concordance matches the old Trados never did.
Just like in the Heinrichs's examples.

As I understand, the old Trados concordance algorithm was n-gram (trigram) based, the current one is not, only a basic string distance calculation is performed.

Cheers
GG

[Edited at 2013-02-14 12:03 GMT]


Direct link Reply with quote
 

Tomi Dobaj
Slovenia
Local time: 03:26
Member (2007)
English to Slovenian
+ ...
Concordance search returns old results in a large TM Apr 5, 2013

Hello everyone!

Have anyone encountered this issue?
It is an annoying one too, it mostly brings up ancient TM entries for simpler searches.

Does Studio have limits on how large a TM may be to function properly?
I did re-import those two TMs with "character based" option, to no avail.


Direct link Reply with quote
 

Artem Vakhitov  Identity Verified
Estonia
English to Russian
+ ...
Ctrl+F3 Apr 5, 2013

Try using Ctrl+F3 instead of just F3. For some reason this keystroke seems to work more consistently for me.

Direct link Reply with quote
 

Tomi Dobaj
Slovenia
Local time: 03:26
Member (2007)
English to Slovenian
+ ...
ctrl+f3... Apr 5, 2013

...doesn't seem to make any difference for me. I still get old results

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 »

Concordance search completely useless?

Advanced search







LSP.expert
You’re a freelance translator? LSP.expert helps you manage your daily translation jobs. It’s easy, fast and secure.

How about you start tracking translation jobs and sending invoices in minutes? You can also manage your clients and generate reports about your business activities. So you always keep a clear view on your planning, AND you get a free 30 day trial period!

More info »
Anycount & Translation Office 3000
Translation Office 3000

Translation Office 3000 is an advanced accounting tool for freelance translators and small agencies. TO3000 easily and seamlessly integrates with the business life of professional freelance translators.

More info »



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