Video on 101% matches, with memoQ & Olifant
Thread poster: Dominique Pivard

Dominique Pivard  Identity Verified
Local time: 15:10
Finnish to French
Dec 16, 2011

I just posted a short video on 101% matches (aka in-context matches, guaranteed matches, perfect matches etc.), using memoQ translator pro and Olifant:

http://youtu.be/WjBomTycIWk?hd=1

If you are aware of other CAT tools (in addition to those mentioned in the video) that support 101% matches, do let me know!


Direct link Reply with quote
 

Chunyi Chen
United States
Local time: 05:10
English to Chinese
Thank you for creating an excellent video demonstration. Dec 16, 2011

great resource indeed. I am bookmarking the link.
Thanks again!

Chun-yi


Direct link Reply with quote
 
xxxPzit
English to Italian
Thanks very useful! Dec 16, 2011

Thanks, that cleared some foggy notions I had about the 101% matches. I subscribed to your channel, looking for more memoQ stuff!

Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 14:10
Member (2006)
English to Afrikaans
+ ...
The next level would be structural exact matching Dec 16, 2011

Dominique Pivard wrote:
If you are aware of other CAT tools (in addition to those mentioned in the video) that support 101% matches, do let me know!


Idiom DesktopWorkbench has it too (called ICE matches), and so does Google Translate Toolkit (called in-context translations). My first Idiom job was in October 2007, and at that time it already had ICE matching.

OmegaT has it in a limited sense. In OmegaT, 101% matching is not a feature but a kludge. As you may know, when OmegaT creates target files, it performs a blind find/replace using the TM, which means that (under non-101% matching circumstances) non-unique segments would all be translated the same way, even if the translator attempts to provide multiple translations. OmegaT solves the problem of non-unique segments using the 101% match feature, which takes into account the one segment before and one segment after the matched segment. This does not solve the problem of non-unique triplets of segments, though...

Personally I think that triplets of segments do not offer sufficient context for 101% matching, and I would rather see quintlets (i.e. two segments before and two segments after). In all of the Idiom DesktopWorkbench jobs that I have done so far, 101% matches (from previous TMs) were non-paid, and from what I've seen, triplets of segments are not enough to ensure exact matching.

The next level that I would like to see in CAT tools is 101% matching that also takes into account the structural location of the match, e.g. whether the match is in a heading, in a list, in a paragraph, in a header/footer, etc.

==

Oh, and Wordfast Pro has something called "golden" matches (so indicated in the TXML file) but the user manual doesn't say what it is (it contains the word "gold[en]" only once, on the page about penalties, but does not define it in any way). Would that be similar? Or is WFP's "golden" matches simply their way of saying "pure 100% match" (i.e. same capitalisation, same formatting, same numbers, etc)?



[Edited at 2011-12-16 21:20 GMT]


Direct link Reply with quote
 

Tomás Cano Binder, BA, CT  Identity Verified
Spain
Local time: 14:10
Member (2005)
English to Spanish
+ ...
Yes... Dec 16, 2011

Samuel Murray wrote:
Personally I think that triplets of segments do not offer sufficient context for 101% matching, and I would rather see quintlets (i.e. two segments before and two segments after). In all of the Idiom DesktopWorkbench jobs that I have done so far, 101% matches (from previous TMs) were non-paid, and from what I've seen, triplets of segments are not enough to ensure exact matching.

The next level that I would like to see in CAT tools is 101% matching that also takes into account the structural location of the match, e.g. whether the match is in a heading, in a list, in a paragraph, in a header/footer, etc.

Indeed, I would pay for two features like that. Good point Samuel.


Direct link Reply with quote
 

Rodolfo Raya  Identity Verified
Local time: 09:10
English to Spanish
In-context exact matches Dec 16, 2011

Hi Dominique,

Swordfish supports in-context exact matches.

Regards,
Rodolfo


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 14:10
Member (2006)
English to Afrikaans
+ ...
Swordfish Dec 16, 2011

Rodolfo Raya wrote:
Swordfish supports in-context exact matches.


Since when, Rodolfo, and how much did the first version that supported it cost?

From the description:
http://www.maxprograms.com/products/swordfish.html#incontext
it would appear that Swordfish's TM format does not support in-context matching but that Swordfish performs in-context matching based on imported legacy XLF files. Do I understand correctly?


Direct link Reply with quote
 

Rodolfo Raya  Identity Verified
Local time: 09:10
English to Spanish
In context exact matches Dec 16, 2011

Samuel Murray wrote:

Rodolfo Raya wrote:
Swordfish supports in-context exact matches.


Since when, Rodolfo, and how much did the first version that supported it cost?


Since version 1.0-2, released on May 2008. The price at that time was €240 if I remember correctly.


From the description:
http://www.maxprograms.com/products/swordfish.html#incontext
it would appear that Swordfish's TM format does not support in-context matching but that Swordfish performs in-context matching based on imported legacy XLF files. Do I understand correctly?


Swordfish does not rely on a TMX file or TM database. A TMX file does not have enough information about the structure of a document.

Swordfish compares two XLIFF files, a previous translation that you did and the XLIFF file generated from a modified version of the same document. As Swordfish knows how the XLIFF files were created, it knows the document structure and can compare segment location, content and segment ID.

Comparing two XLIFF files is like using tracked changes to compare two versions of a document.

Regards,
Rodolfo

[Edited at 2011-12-16 21:41 GMT]


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 15:10
Finnish to French
TOPIC STARTER
Thanks for the info about Swordish! Dec 17, 2011

Rodolfo Raya wrote:
Since version 1.0-2, released on May 2008. The price at that time was €240 if I remember correctly.

Thanks for the info, Rodolfo! I have added an annotation about it in the video at 1:20s:

http://youtu.be/WjBomTycIWk?hd=1&t=1m18s

So it looks like Swordfish was 2nd in introducing 101% matches in an affordable package, one year after memoQ, but one year before Trados (and two years before Déjà Vu).

Rodolfo Raya wrote:
Swordfish does not rely on a TMX file or TM database. A TMX file does not have enough information about the structure of a document.

Swordfish compares two XLIFF files, a previous translation that you did and the XLIFF file generated from a modified version of the same document. As Swordfish knows how the XLIFF files were created, it knows the document structure and can compare segment location, content and segment ID.

Comparing two XLIFF files is like using tracked changes to compare two versions of a document.

So this means there is no way to export information about 101% matches via TMX? Not that I'm convinced the context information included by memoQ in its TMX can be used by other tools (in fact, it's something I may have a look at). Of course, a bunch of XLIFF files translated in Swordfish could be added to a memoQ LiveDocs corpus, in which case memoQ should be able to serve 101% matches from the corpus. The same should work with XLIFF produced by other tools (eg. Studio 2011) or with other bilingual files (eg. TXML or uncleaned DOC). Actually, the ability to keep context information (something I have to check) makes LiveDocs a very attractive solution compared to TMX.


Direct link Reply with quote
 

Dominique Pivard  Identity Verified
Local time: 15:10
Finnish to French
TOPIC STARTER
memoQ playlist on my YouTube channel Dec 18, 2011

Stefano Armanini wrote:
Thanks, that cleared some foggy notions I had about the 101% matches. I subscribed to your channel, looking for more memoQ stuff!

My channel doesn't feature only memoQ stuff, but I've created a memoQ playlist, so all memoQ content can now be accessed from a single place: http://www.youtube.com/playlist?list=PL459FF807E628EE1C


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 »

Video on 101% matches, with memoQ & Olifant

Advanced search






SDL Trados Studio 2017 Freelance
The leading translation software used by over 250,000 translators.

SDL Trados Studio 2017 helps translators increase translation productivity whilst ensuring quality. Combining translation memory, terminology management and machine translation in one simple and easy-to-use environment.

More info »
WordFinder
The words you want Anywhere, Anytime

WordFinder is the market's fastest and easiest way of finding the right word, term, translation or synonym in one or more dictionaries. In our assortment you can choose among more than 120 dictionaries in 15 languages from leading publishers.

More info »



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