What is a context match?
Thread poster: Dominique Pivard

Dominique Pivard  Identity Verified
Local time: 20:38
Finnish to French
Jan 9, 2013

First of all, sorry for the length of this post!

I made an experiment with two very similar sample documents containing nine segments. I translated the first document, committing all nine segments to an empty TM. Here is what the second document looked like (compared to the first one):



Segment nr 4 was not found in the first document. On the above picture, I'm highlighting it in yellow to show it was added, but in the actual document it just had normal formatting. The segment highlighted in red was found in the first document, but not in the second document. Again, I'm highlighting it in red to show it was deleted, but in the actual document it was missing altogether.

When I added the second document to the project and Studio pretranslated it with the TM created from the translation of the first document, Studio found six context matches, two 100% matches and one segment for which there was no match:



Based on the definition of a context match from the online Help, I was surprised to see that many context matches:

A match that takes into account the context within which the match occurs. Usually this means surrounding segments. A context match is even more likely to be correct than a 100% match of a segment. There are different kinds of context match, depending on which aspects of context are taken into account (for example, formatting).

As I see it, segment nr 3 is not a context match, because it is not followed by the same segment in document nr 2 as in document nr 1. Likewise, segments nr 7 and 9 are not context matches, because they were not followed or preceded by the same segment.

Here is what Heartsome Studio Ultimate 8.1 thought in the same situation:



and memoQ 6.2, which agrees with Heartsome:



They found four context matches, four 100% matches and one segment for which there was no match. I agree with them, except for segment nr 9, which was not preceded by the same segment in document nr 1.

And here is what Déjà Vu X2 Professional thinks:



It's the most conservative, finding only three context matches (it agrees with me that segment nr 9 is not a context match), five 100% match and one segment for which there was no match.

Btw, DV has the most configurable context match feature since it allows you to define the "context size", by which it means the number of segments that constitute the context: three, five etc. I used "3", ie. actual segment + preceding and following segment.

What do you think?

[Edited at 2013-01-09 22:41 GMT]


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 19:38
English
What do I think? Jan 9, 2013

Dominique Pivard wrote:

As I see it, segment nr 3 is not a context match, because it is not followed by the same segment in document nr 2 as in document nr 1. Likewise, segments nr 7 and 9 are not context matches, because they were not followed or preceded by the same segment.



Hi Dominique,

To determine a Context Match Studio looks at the preceding segment and that segment’s translation, so assigning the CM property on segment #3, #7 and #9 is correct, as the preceding segments and their translations are available and the translation is confirmed.

Perfect Match can look at both sides of the segment and considers blocks of text for alignment, but you didn't test that. I did


Regards

Paul


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 20:38
Finnish to French
TOPIC STARTER
Context matches vs. perfect matches Jan 10, 2013

SDL Support wrote:
To determine a Context Match Studio looks at the preceding segment and that segment’s translation, so assigning the CM property on segment #3, #7 and #9 is correct, as the preceding segments and their translations are available and the translation is confirmed.

Well, "looks at the preceding segment" (your description) isn't exactly the same as "surrounding segments" (description of the online help), is it?

SDL Support wrote:
Perfect Match can look at both sides of the segment and considers blocks of text for alignment, but you didn't test that. I did

Yes, now I remember we have had that discussion before. But can you confirm there is no way I could have obtained these perfect matches with my copy of Studio 2011 Freelance, ie. one needs Studio Professional in order to extract perfect matches from a previously translated document?
Is this equivalent/similar to memoQ's X-Translate feature, which I must admit I've never figured out how to use in practice?


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 19:38
English
Depends... Jan 10, 2013

... on where you looked. If I do a search for "Context Match" in the online help I get many results. The first one is a reference to Context match in the online Glosssary which is pretty vague because we have many products and the Glossary is a Language Technologies Glossary. So it actually says "... usually means surrounding segments...". Usually isn't very specific so I look at the second result and find this:


X-Translate... I can tell you that Perfect Match, or X-Translate as it was known in the Trados days since around Trados 5 (around 2001 I think), are the same thing.

I thought I'd take a look at the memoQ version, which is described in the memoQ help as a similar thing to Trados but I can't add a document to create a project at all (must be a glitch in my shiny new 6.2). I'll raise a ticket and maybe report back later if you still haven't figured it out.

Regards

Paul


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 20:38
Finnish to French
TOPIC STARTER
Perfect Match vs. X-Translate Jan 10, 2013

SDL Support wrote:


Thanks, Paul! However, the last segment in my example did not have the same preceding segment in the first and in the second document, yet it was considered a CM.

SDL Support wrote:
X-Translate... I can tell you that Perfect Match, or X-Translate as it was known in the Trados days since around Trados 5 (around 2001 I think), are the same thing.

OK, but you can only extract Perfect Matches with Studio Professional, not Studio Freelance, right?

SDL Support wrote:
I thought I'd take a look at the memoQ version, which is described in the memoQ help as a similar thing to Trados but I can't add a document to create a project at all (must be a glitch in my shiny new 6.2). I'll raise a ticket and maybe report back later if you still haven't figured it out.

I have no problem adding documents in memoQ 6.2, but I can't figure out how to use X-Translate to translate the second document. I asked about it on the memoQ list.


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 19:38
English
Why not? Jan 10, 2013

Dominique Pivard wrote:

Thanks, Paul! However, the last segment in my example did not have the same preceding segment in the first and in the second document, yet it was considered a CM.



Hi Dominique,

Segment #8 is the same in both documents so why would segment #9 not be a CM? Segment #8 had a different preceding segment so this correctly gets a 100% match instead, as did segment #5.

Maybe you're looking at too many different tools, all with their own ways of interpreting this and are beginning to get a little confused?

Regards

Paul


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 20:38
Finnish to French
TOPIC STARTER
You are correct! Jan 10, 2013

SDL Support wrote:
Segment #8 is the same in both documents so why would segment #9 not be a CM? Segment #8 had a different preceding segment so this correctly gets a 100% match instead, as did segment #5.

Maybe you're looking at too many different tools, all with their own ways of interpreting this and are beginning to get a little confused?

Yes, you are right: segment #9 is indeed preceding by the same segment in both documents. I meant segment #7, which is followed by a different segment, but this is in accordance with the definition you provided, since it's the preceding segment that matters to Studio. Other tools have stricter definitions of context (preceding + following).

And I did manage to get X-Translate to work in memoQ, obtaining the same results as you with Perfect Match. One last question: can you confirm Perfect Matches can only be created in Studio Professional, not Freelance?


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 19:38
English
Sort of... I think Jan 10, 2013

Dominique Pivard wrote:

.... since it's the preceding segment that matters to Studio. Other tools have stricter definitions of context (preceding + following).



Studio compares source and target in the preceding segment... whereas I think other tools only compare the source (happy to be corrected on this as I have not tested thoroughly). It of course makes no sense to compare the succeeding segment on this basis because you won't have translated it yet! It would be ok for values coming from documents previously translated in your TM but not from interactive translation on new material and this is what allows Studio to provide Context Matches as you work.

And yes, Perfect Match is currently only available with the Professional edition of Studio. It can be seen and used in SDLXLIFF files in any edition, but only put there in the first place with the Professional edition.

Regards

Paul


Direct link Reply with quote
 

Łukasz Rejter
Poland
Local time: 19:38
English to Polish
+ ...
X-translate in memoQ Jan 10, 2013

In order to use X-translate in memoQ you need to use "record version history feature".

This is how it works:

1. translate the first document, eg. document.v.1.doc
2. use Reimport function to import its new version eg. document.v.2.doc
3. go to Operations - X-translate, and tell memoQ to use the translations from the previous major version (if there are more versions, you can pick any one of them).

You should have no problems setting the remaining features right.

Hope it helps.

Regards,
Łukasz


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 20:38
Finnish to French
TOPIC STARTER
X-Translate Jan 11, 2013

Łukasz Rejter wrote:
In order to use X-translate in memoQ you need to use "record version history feature".

This is how it works:

1. translate the first document, eg. document.v.1.doc
2. use Reimport function to import its new version eg. document.v.2.doc
3. go to Operations - X-translate, and tell memoQ to use the translations from the previous major version (if there are more versions, you can pick any one of them).

Thanks, Łukasz! I had got the same instructions from István on the memoQ list and it does indeed work fine. Here is what I got, it looks exactly as Perfect Matches with Studio Professional:



What I failed to do was using the Reimport function.


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 »

What is a context match?

Advanced search







Protemos translation business management system
Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!

The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.

More info »
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 »



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