Что такое context match trados

Context Match vs. PerfectMatch

RevleX™ translation memory (TM) engine
A ground-breaking TM engine which delivers a host of time-saving features such as Context Match, improved concordance searching, plus many more.

Context Match: take accuracy to new heights
Provides «beyond 100%» matches by recognizing location and context to deliver the best translation. No complicated set-up or configuration required.

SDL PerfectMatch™
Leverage previously translated bilingual files to create PerfectMatch content. A quick, simple way to reduce review time and ensure consistency.

Context Match: take accuracy to new heights
Provides «beyond 100%» matches by recognizing location and context to deliver the best translation. No complicated set-up or configuration required!

My question is: what’s the difference between Context Match and PerfectMatch? Is it that with the former, the context is stored in the TM, while with the latter, it is stored in documents (SDLXLIFF)?

And what is the «ground-breaking» RevleX™ TM engine? How does it differ from the previous TM engine? Do I understand correctly that context matches did not exist in Studio 2009, which didn’t use the «ground-breaking» RevleX™ TM engine? &#x25B2 Collapse

Jerzy Czopik
Германия
Local time: 08:26
This person is a ProZ.com Certified PRO in немецкий => польский

Context matching is TM matching plus additional information.
This additional information is document structure level and preceding sentence (segment).
Perfect matching is document matching. A TM is used only for segmentation purposes, but will NOT be used for any translation. The process compares the current document with its previous version, which must be available as bilingual file (TTX, SDLXLIFF) and inserts all parts of the translation, where the current document and the preceding document are identical. This does not only include the preceding and following sentence, but also the whole paragraph and document structure.

Context Match has been first introduced in Studio 2009, where the new TM format made that possible.
Perfect Match is a successor of a technology called X-Translate and later ContextTM.

[Edited at 2012-01-12 18:06 GMT] &#x25B2 Collapse

Thanks, Jerzy, for clarifying the difference! I have another question related to creating PerfectMatches (only possible with Studio Professional) vs. merely using them (possible with Studio Freelance): «creating PerfectMatches» means running a batch task that inserts the translation and the «PM» status into the new SDLXLIFF, right? What does «using PerfectMatches» mean in practice for a translator who has Studio Freelance 2011? That he will be able to see the «PM» status for PerfectMatches inserted by Studio Professional? Or is there anything else he can do to use these special matches?

Jerzy Czopik
Германия
Local time: 08:26
This person is a ProZ.com Certified PRO in немецкий => польский

Thanks! I suppose using PM matches also works in Studio 2009 Freelance (since the SDLXLIFF format is the same as in 2011), although one needs to have Studio 2011 Professional in order to insert PerfectMatches? Or won’t Studio 2009 be able to display the PM status?

[Edited at 2012-01-13 13:15 GMT]

I made a small video about 101% matches in various CAT tools (including Studio) a few weeks ago and there was a confusion between Context Matches and PerfectMatches. I edited it based on Jerzy’s answer, adding an annotation at 01:18:
http://youtu.be/WjBomTycIWk?t=1m14s

Jerzy Czopik
Германия
Local time: 08:26
This person is a ProZ.com Certified PRO in немецкий => польский

Thanks! If you no longer have Studio 2009 installed, just send me a sample file (TTX, SDLXLIFF) that contains PM matches. I can then open them in Studio 2009 (haven’t upgraded yet) and let you know whether or not they are displayed with the PM status.

Jerzy Czopik
Германия
Local time: 08:26
This person is a ProZ.com Certified PRO in немецкий => польский

This was always possible in 2009 because any files from legacy files that contained Perfect Matches were handled as such.

The video from Dominique was very interesting, but I think it’s wrong to compare Perfect Match with 101% matches. I think Context Match is the equivalent.

This was always possible in 2009 because any files from legacy files that contained Perfect Matches were handled as such.

The video from Dominique was very interesting, but I think it’s wrong to compare Perfect Match with 101% matches. I think Context Match is the equivalent.

One interesting difference around the way we handle matches is that we only store one TU for one multiple 100% translation because we merge the context information into the same TU. This saves a lot of space over time and keeps the TM cleaner. We also don’t store the full text (again requiring space) but instead we calculate a hash code based on the previous segment and its translation.

Interesting stuff though and worth more explanation I think. something to work on.

Источник

What is a context match?