Pages in topic:   [1 2] >
Inconsistent term recognition from glossaries
Thread poster: Khadhé

Khadhé
Local time: 13:07
English to French
Feb 19, 2013

Hello colleagues,

I have been experiencing a recurring problem with term recognition with glossaries under WordFast Classic Version 6.03t running on MS Word 2007 (SP3) and Vista Business SP2 :

In order for a term to be recognized, I sometimes have to go up (closing the segment) then down again (re-opening the segment) using ATL-UP ARROW/ ALT- DOWN ARROW to force an other attempt at term recognition from WordFast to finally manage to pick up one or more terms from one or both of the active glossaries.

This behavior is not consistent, i.e. it may, or may not happen, and I haven't been able to figure what other parameters it may correlate with, except one: glossary sorting and reorganizing. If I do perform that operation on each of the two glossaries I am using, then term recognition will work fine on opening of the next segment. But I can't obviously perform this before each segment opening. It would defeat the purpose. After an organize/sort operation, the erratic behavior can start again as soon as the second segment is opened.

I am usually running 2 glossaries: a small, client specific one, and a global massive one (currently over 27,000 entries but way bellow the stated limit of 250,000 entries per single glossary). QA during translation and QA in companion are both unchecked in Set-up.

Any idea regarding what could be causing this behavior is welcome,

Cheers,
Pascal


 

Dominique Pivard  Identity Verified
Local time: 23:07
Finnish to French
Location of glossaries? Feb 20, 2013

Khadhé wrote:
I have been experiencing a recurring problem with term recognition with glossaries under WordFast Classic Version 6.03t running on MS Word 2007 (SP3) and Vista Business SP2 :

In order for a term to be recognized, I sometimes have to go up (closing the segment) then down again (re-opening the segment) using ATL-UP ARROW/ ALT- DOWN ARROW to force an other attempt at term recognition from WordFast to finally manage to pick up one or more terms from one or both of the active glossaries.

Are your glossaries located on a network drive or in a Dropbox folder? Are they shared with someone else?


 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
No, none of that, but... Feb 20, 2013

...now that you mention that, the only thing I can think of is that access to the drive they are on might be slow (the drive being "green", i.e. it does not spin all the time). So, maybe WF gives up if it can not access that drive and scan the glossary on the fly. If this is the case, I am surprised I am not getting an disk access error instead of a seemingly no glossary match behavior.

I will test by copying the glossary to the faster system drive.

Thanks for your answer.

P


 

Dominique Pivard  Identity Verified
Local time: 23:07
Finnish to French
Did it work? Feb 25, 2013

Khadhé wrote:
I will test by copying the glossary to the faster system drive.

Did relocating your glossaries to the local drive (C:) solve your problem?

In order for terminology recognition to work, Wordfast needs to have access to the glossary index when it opens a new segment. If it has to wait for a sleeping drive to wake up ("green" external drive), or for the index to be synchronized (Dropbox), it won't have access to that file when it needs it. If access is obtained a couple of seconds later, it's too late (for that segment).


 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
Kind of... Feb 26, 2013

I tried to relocate the big glossary : it got worse (WF had to check 2 HDs), then both glossaries (no difference).

It's only when I put my document, the TM and both glossaries on the system drive that WF starts behaving normally (consistent terminology recognition from both glossaries)

I usually don't keep any data files on the system drive (I don't use the typical MS generic Documents, Pictures, Video, etc. directories in the Windows user account, and my weekly schedule back up for system drives is not as frequent as for data drives (daily) so I am going to have to juggle around a bit.

Green drive is a nice idea but there are drawbacks... here is one.

thanks for your comment.

EDIT (Feb. 26, 2013) : In fact, I notice today, just one day after implementing the change described above, that even in this configuration, WF keeps missing term recognition, even in the smallest of the 2 glossaries. The missing terms are eventually recognized if WF is forced back to the previous segment, then forward again.





[Edited at 2013-02-26 01:03 GMT]

[Edited at 2013-02-26 18:09 GMT]

[Edited at 2013-02-28 03:56 GMT]


 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
Term recognition still malfunctioning Feb 27, 2013

Unfortunately, term recognition will quite often never happen, despite the term being in one or both glossaries and me trying to go back and forth between segment, as mentioned in my previous post.

This is a major inconvenience since I will then try to enter one or more terms, thinking they are not in any of the open glossaries. When I try to validate the "new" entry (pressing "OK" in the Edit entry Popup), I suppose WF properly checks its glossaries because I will occasionally get a warning that after all, the term IS actually already there, in one or both glossaries! So I have then to cancel one or both new entries. A major waste of time.

I wonder if anybody else is experiencing this behavior, i.e. if it's inherent to WF or it has anything to do with my set-up/harware.



[Edited at 2013-02-27 18:32 GMT]


 

Dominique Pivard  Identity Verified
Local time: 23:07
Finnish to French
Sample source segment? Feb 28, 2013

Khadhé wrote:
Unfortunately, term recognition will quite often never happen, despite the term being in one or both glossaries and me trying to go back and forth between segment, as mentioned in my previous post.

This is a major inconvenience since I will then try to enter one or more terms, thinking they are not in any of the open glossaries. When I try to validate the "new" entry (pressing "OK" in the Edit entry Popup), I suppose WF properly checks its glossaries because I will occasionally get a warning that after all, the term IS actually already there, in one or both glossaries! So I have then to cancel one or both new entries. A major waste of time.

Can you post a sample source segment in which this is happening, mentioning what source terms are in the glossaries and what new terms you are adding?


 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
I might be the only one experiencing this problem Feb 28, 2013

Dominique,

Thanks for wanting to help or investigate. Terms that are not recognized (despite being in one or both glossaries) are random. Going backward then forward will sometimes "force" the recognition. That's the main symptom. I am not sure what you would be able to infer by knowing a particular term that was not recognized properly despite being in one or both glossaries.

All I am asking at this stage is whether somebody out there has experienced the same type of problem. If nobody ever experienced that with WF classic, then I will experiment more on my side to see if I can improve the situation.


 

lestertrad
Local time: 22:07
French to English
+ ...
I too am experieincing the problem of failure to recognize glossary terms. Mar 6, 2013

Khadhé wrote:

All I am asking at this stage is whether somebody out there has experienced the same type of problem. If nobody ever experienced that with WF classic, then I will experiment more on my side to see if I can improve the situation.


I have definitely been experiencing this problem of random failure to recognize glossary terms. When the term is preceded or followed by punctuation, its chances of being recognized seem to drop to almost nil.

This thread has been instructive. I have disabled spin-down on the drive where the glossaries are. I will report on whether that helps.

Thanks to both of you for the information.



[Edited at 2013-03-06 13:50 GMT]


 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
Let's try to figure out what we have in common! Mar 6, 2013

Thanks for chiming in lestertrad

If we can figure out what is common between our two configs, we might be able to pinpoint where the problem is coming from.

I have described my system details in my first post in this thread.

At this stage, since moving the two active glossaries onto the fast spinning system drive did not change anything, my primary suspect is the glossaries size.

Do you mind posting:
1) your system details? (WF type classic or pro and version #; MS Word version; and OS.)
2) Number and size (nb of entries) of active glossaries?

Mines are

1)WordFast Classic Version 6.03t running on MS Word 2007 (SP3) and Vista Business SP2

2)
Glossary #1 : 2027 entries
Glossary #2 : not active
Glossary #3 : 28174 entries


 

Salvador Scofano and Gry Midttun
Norway
Local time: 22:07
Member (2006)
English to Portuguese
+ ...
save the glossary as Unicode Mar 8, 2013

Have you tried to save the glossary as Unicode? Then click in the Terminology tab and reorganize it. Good luck.

 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
Already UTF, and reorganizing does not solve the problem Mar 8, 2013

Thanks for the suggestion.

The character format for both glossaries is already the Unicode character set (UTF-8). I switched to that format long ago.

I often reorganize the glossaries but unfortunately, that does not make any difference on the general behavior.
Reorganizing actually just triggers the same thing as going backward then forward: it will force another recognition for the present segment.


 

Milan Condak  Identity Verified
Local time: 22:07
English to Czech
Unicode vs. UTF-8 Mar 8, 2013

Unicode is Unicode or Unicode LE16
UTF-8 is UTF-8

http://www.condak.net/tmx/wfconverter/cs/00.html

Milan

Khadhé wrote:

Thanks for the suggestion.

The character format for both glossaries is already the Unicode character set (UTF-8).



 

Khadhé
Local time: 13:07
English to French
TOPIC STARTER
Correct, Unicode LE16 is used but it does not solve the problem Mar 8, 2013

Yes, you're correct. WF probably uses the LE16 to encode unicode. UFT-8 is actually also a unicode version but is designed for the Linux platform.

Anyway, I used the WF "rewrite glossary as unicode" tool for both glossaries. So WF should be able to deal with its own unicode encoding.


 

Milan Condak  Identity Verified
Local time: 22:07
English to Czech
WFC is using Unicode Mar 8, 2013

1. WF uses the Unicode (LE16)
2. In WFC is no tool to convert UTF-8 to Unicode
3. WFC can "rewrite glossary as unicode" from ANSI (your national Windows code)
4. UTF-8 is for all platform, basicly for HTML
5. Open your file in NotePad, try Save as and you can see your current encoding. Change encoding to Unicode and save.

Milan

Khadhé wrote:

Yes, you're correct. WF probably uses the LE16 to encode unicode. UFT-8 is actually also a unicode version but is designed for the Linux platform.

Anyway, I used the WF "rewrite glossary as unicode" tool for both glossaries. So WF should be able to deal with its own unicode encoding.



[Upraveno: 2013-03-08 19:32 GMT]

[Upraveno: 2013-03-08 19:32 GMT]


 
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 »

Inconsistent term recognition from glossaries

Advanced search


Translation news related to Wordfast





TM-Town
Manage your TMs and Terms ... and boost your translation business

Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.

More info »
PerfectIt consistency checker
Faster Checking, Greater Accuracy

PerfectIt helps deliver error-free documents. It improves consistency, ensures quality and helps to enforce style guides. It’s a powerful tool for pro users, and comes with the assurance of a 30-day money back guarantee.

More info »



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