Assemble doesn't work - DVX2 professional
Thread poster: Lucia Messuti
Lucia Messuti  Identity Verified
Italy
Local time: 13:32
Member (2003)
English to Italian
+ ...
Oct 10, 2011

Hello
I'm quite new to DVX (latestst edition). Overall, I think it's a really one of the best CATs I've ever used.

However, I'm experiencing a problem, or at least I think it's a problem.

I checked the ASSEMBLE box to get terms from TM and Termbase popping into the target segment. In this case, as I've quite comprehensive TMs I should get several terms / sentences in the target segment.
But, that's what I did and got in the end:
- I did assemble (from pretranslate) and although I was sure I had the translated term in the TM, it was not inserted automatically into the target segment (I can view the terms in Autosearch Segments and through Scan with Ctrl +S, but not in the Autosearch - Portions).
The same occurs also when I don't do the pretranslate, and when translating (I've checked in GENERAL - ASSEMBLE - FIRST SCAN) I don't get the term I'm sure I've in the translation database.

I only get the terms in the termbase, but not the ones contained in the TM.

Then did anybody tried to use DVX2 with txml files, i.e. files from Wordfast Pro? I don't like working with txml files, but I couldn't successfully import/translate/export them through DVX. I imported an xml filter, as eventually they are xml files, but nothing.

Can anybody help me?

Thanks
Lucia


Direct link Reply with quote
 

Selcuk Akyuz  Identity Verified
Turkey
Local time: 15:32
Member (2006)
English to Turkish
+ ...
Settings Oct 11, 2011

Hello Lucia,

Lucia Messuti wrote:

I checked the ASSEMBLE box to get terms from TM and Termbase popping into the target segment. In this case, as I've quite comprehensive TMs I should get several terms / sentences in the target segment.
But, that's what I did and got in the end:
- I did assemble (from pretranslate) and although I was sure I had the translated term in the TM, it was not inserted automatically into the target segment (I can view the terms in Autosearch Segments and through Scan with Ctrl +S, but not in the Autosearch - Portions).
The same occurs also when I don't do the pretranslate, and when translating (I've checked in GENERAL - ASSEMBLE - FIRST SCAN) I don't get the term I'm sure I've in the translation database.

I only get the terms in the termbase, but not the ones contained in the TM.


Let's look at all settings once again.

Under Project>Properties >Databases
read & write both checked?

Under Tools>Options>General
Database Lookup minimum score: something between 60-75

Under Tools>Options>Environment
AutoSearch
Enable AutoSearch checked?

Checking other options may help, I use Display portions from Translation Memory and Display all subportions as well.

But I am a bit confused, you have segments in your Translation Memory and terms in your Term Base. So if you are expecting one word segments to appear in your AutoSearch window, the options Display portions from Translation Memory and Display all subportions should be checked. But then you will have a very long list of entries in your AutoSearch window. In general, Termbase or Lexicon are used to store one or two-word terms.


Then did anybody tried to use DVX2 with txml files, i.e. files from Wordfast Pro? I don't like working with txml files, but I couldn't successfully import/translate/export them through DVX. I imported an xml filter, as eventually they are xml files, but nothing.


There is a TXML filter at http://tech.groups.yahoo.com/group/dejavu-l/files/Wordfast%206/ created by David Turner. Well, Wordfast Pro was first launched as WordFast 6 and later renamed as Wordfast Pro. You can use this filter to translate Wordfast Pro files.

You need to copy all source segments to target in Wordfast Pro (Ctrl+Alt+Insert if I am not mistaken), and then import this file to DVX.

Hope all is clear now. Are you a member of http://tech.groups.yahoo.com/group/dejavu-l/ You will receive much more answers in that group page from more than 2000 users. And there is always the Atril support, just send a mail to support@atril.com

Selcuk


Direct link Reply with quote
 
Lucia Messuti  Identity Verified
Italy
Local time: 13:32
Member (2003)
English to Italian
+ ...
TOPIC STARTER
Thanks fo ryour reply Oct 12, 2011

Hello Selcuk,
thanks for your reply.
I've checked all settings you mentioned checked and enabled, but nothing happens.

I'll try to explain better.
I've an extensive TM (read and write) and a small termbase.
When I translate (CTRL+DOWN) and go to the next segment (assemble option enabled) I'm supposed to get all terms contained in the TM into the target segments. What's strange is that if I do a scan I find entire segments in the TM, and I'm supposed I should get them in the target segment (assemble function), but I don't get them. The only terms I get in the target segment (after CTRL-DOWN) (assembled, as there is the blue dash aside) are the ones contained in the database.

When you use the pretranslate function, what options do you enable? I've also tried machine translation but it doesn't seem a good idea. On the other side, where does DeepMining gets the terms? Only in the databases?

Except this issue, I'm very happy about this CAT. It seems much more usefull than Trados and Wordfast.

Thanks for your help and have a nice day
Lucia


Direct link Reply with quote
 

Victor Dewsbery  Identity Verified
Germany
Local time: 13:32
German to English
+ ...
TM and termbase Oct 12, 2011

Hi Lucia,

I'm not sure what "terms" you want to see from the TM. The basic concept is that the TM is used for segments (e.g. sentences), and the termbase is used for "terms". The terms may consist of one word or whole phrases, but the termbase is the place to put terms that you want to be found during Pretranslate or Assemble. And terms that are specific to a single project can be put in the lexicon.

In theory (for CAT tools in general), the TM finds whole segments which are the same as your current segment. It can also find similar segments - you can vary how similar the segments must be by changing the fuzzy percentage (Tool>Options>General>Database Lookups). If you set it to 90%, DVX will only find very similar segments, if you set it to 50% or 60% it will find much more. This setting is also used for the termbase, so if this setting is at 50% or 60%, you will see more terms (e.g. with different endings, depending on the rules in your source language). But the lexicon only finds 100% matches for the terms.

So "terms" which are in the TM may not be found during Pretranslate/Assemble. There are a couple of tricks to get round this:
1. Scan (highlight a source term and press "CTRL-S"). This shows you the segments in the TM which contain the term you have highlighted. This function has been in DVX for years (and even in the predecessor DV3). But the next two tricks are only in the new DVX2 version:
2. DeepMiner (if you have enabled it). This searches your TM for extra clues. You will need to experiment to decide whether it really makes a difference in the work you do. Performance and speed vary depending on various factors, including the size of your TM.
3. AutoWrite. This is similar in a way to the suggestions you get when you write an SMS text on some mobile phones. The difference is that AutoWrite takes its suggestions from your TM (and also the termbase and lexicon), and sometimes it gives you whole phrases which are taken from a sentence you did just a couple of segments further up. It takes some getting used to - when I first saw AutoWrite, I was horrified because I thought it was an immense distraction. But then I looked at what it actually suggested, and I found that it sometimes saves so much typing that it is definitely worthwhile.

[Edited at 2011-10-12 07:52 GMT]


Direct link Reply with quote
 
Lucia Messuti  Identity Verified
Italy
Local time: 13:32
Member (2003)
English to Italian
+ ...
TOPIC STARTER
Thanks for your reply Oct 12, 2011

Hi Victor
thanks for your reply.
So, the assemble function doesn't get words in the TM here and there, to assemble them in the target segment; it only gets them in the termbase. The TM is only implemented in the target segments according to fuzziness, right?

As to termbase, for instance if I have product = prodotto, it implements the term Prodotto to translate, for instance, Production. So, is there a system to keep exact terms? Capital letters or not, etc?

The Autowrite function is very useful. Me too, at the very beginning was quite upset, but now I've learnt to give a look at it while writing and I can spare a lot of time and "hand use".

I also tried machine translation, but I don't like it. While, I haven't tried the DeepMining function. Did you try it?

Anyway, I think it is a great CAT, much better than Trados and trados Studio and Wordfast.
Much more close to the translators' needs.


Direct link Reply with quote
 
Lucia Messuti  Identity Verified
Italy
Local time: 13:32
Member (2003)
English to Italian
+ ...
TOPIC STARTER
TXML filter Oct 12, 2011

[quote]Selcuk Akyuz wrote:

There is a TXML filter at http://tech.groups.yahoo.com/group/dejavu-l/files/Wordfast%206/ created by David Turner. Well, Wordfast Pro was first launched as WordFast 6 and later renamed as Wordfast Pro. You can use this filter to translate Wordfast Pro files.


Selcuk,
would it be possible to provide here in Proz forum the filter for Wordfast Pro, as well?, so also people looking at proz forum will benefit of it?
Thanks
Lucia


Direct link Reply with quote
 

Victor Dewsbery  Identity Verified
Germany
Local time: 13:32
German to English
+ ...
TM, termbase, GT, DeepMiner Oct 12, 2011

So, the assemble function doesn't get words in the TM here and there, to assemble them in the target segment; it only gets them in the termbase. The TM is only implemented in the target segments according to fuzziness, right?


Yes.

As to termbase, for instance if I have product = prodotto, it implements the term Prodotto to translate, for instance, Production. So, is there a system to keep exact terms? Capital letters or not, etc?


Lexicon

I also tried machine translation, but I don't like it. While, I haven't tried the DeepMining function. Did you try it?


Machine translation: It depends. Google Translate probably varies in quality depending on the language pair. In my pair, it sometimes comes up with helpful suggestions, and sometimes with absolute rubbish. It always needs careful editing. But for me there is a bigger problem involved in using GT: confidentiality. I often have contracts to translate. Such texts are always confidential, and sometimes they are very sensitive indeed. I do not wish to let Google get its hands on such material.

DeepMiner: I have tried it, but I am still trying to work out how helpful it is. There is a training video which shows the principle using very simple examples at http://www.youtube.com/user/AtrilVideo


Direct link Reply with quote
 
Lucia Messuti  Identity Verified
Italy
Local time: 13:32
Member (2003)
English to Italian
+ ...
TOPIC STARTER
Thanks to everybody Oct 12, 2011

Now everything is much much clearer

Direct link Reply with quote
 

Lbenito_atril
Local time: 13:32
Assemble and TXML Oct 14, 2011

The answers offered by Selcuk Ajyuz and Victor Dewsberry to the questions about Assemble in DVX are very good. For best performance it is better to rely on the TM for whole sentences and on the TB to contain individual terms or phrases.

Certain options in DVX (such as "Allow portions containing commas" for Assemble) can be used to make DVX attempt to match parts of sentences found in the TM to parts of segments in the current document, thus bridging the gap between TM and TB, but a well maintained TB is important.

The filter definition file available in the Yahoo Groups works well. It is enough to select SGML/XML as the filter type used, and then specify the filter definition file, in order to import TXML files correctly. My own tests just now confirm this.


Direct link Reply with quote
 
Lucia Messuti  Identity Verified
Italy
Local time: 13:32
Member (2003)
English to Italian
+ ...
TOPIC STARTER
DVX is confimed to be the best in CATs Oct 17, 2011

Thanks for your explanation, now things are much clearer.
And I'm using DVX to translate many documents of different formats and it works fine.
Lucia


Direct link Reply with quote
 


To report site rules violations or get help, contact a site moderator:

Moderator(s) of this forum
Pavel Tsvetkov[Call to this topic]

You can also contact site staff by submitting a support request »

Assemble doesn't work - DVX2 professional

Advanced search






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 »
CafeTran Espresso
You've never met a CAT tool this clever!

Translate faster & easier, using a sophisticated CAT tool built by a translator / developer. Accept jobs from clients who use SDL Trados, MemoQ, Wordfast & major CAT tools. Download and start using CafeTran Espresso -- for free

More info »



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