Fuzzies in MemoQ Thread poster: Christin Lee
|
Hi there, Could you please explain to me that how MemoQ generates a word count analysis with Fuzzy matches when starting with an empty new TM? For example, I tried to analyze over 10 files in MemoQ with an empty TM, and the result shows there're fuzzies. I mean if it's a new project, empty TM, there could only be 100% and reps, correct? Could you please explain why? Also it seems MemoQ automatically deletes or ignores variables/tags during anal... See more Hi there, Could you please explain to me that how MemoQ generates a word count analysis with Fuzzy matches when starting with an empty new TM? For example, I tried to analyze over 10 files in MemoQ with an empty TM, and the result shows there're fuzzies. I mean if it's a new project, empty TM, there could only be 100% and reps, correct? Could you please explain why? Also it seems MemoQ automatically deletes or ignores variables/tags during analyzing, unlike Trados Studio. Is it true? Please let me know. Thanks ▲ Collapse | | |
Homogeneity checkbox is selected | May 19, 2014 |
When the Homogeneity checkbox is selected in Statistics options, MemoQ calculates internal fuzzy matches. I'm sure there is a Help topic in the online help about this. | | |
Thanks a lot for your quick response, Mikhail. | | |
One more question | May 19, 2014 |
Also it seems MemoQ automatically deletes or ignores variables/tags during analyzing, unlike Trados Studio. That generates discrepancies of the word count analysis between MemoQ and Trados, correct? | |
|
|
I have no idea about that. What makes you think it does? | | |
Because I did analyze files by using both softwares | May 19, 2014 |
Since I did analyze this big project by using these two softwares, it turns out there's some difference, even though the total is pretty close, in Trados, there're more New (0%~74%), in MemoQ, Fuzzies(75%~99%) are much more than in Trados....... | | |
Each program yields their own analysis | May 20, 2014 |
Christin Lee wrote: Since I did analyze this big project by using these two softwares, it turns out there's some difference, even though the total is pretty close, in Trados, there're more New (0%~74%), in MemoQ, Fuzzies(75%~99%) are much more than in Trados....... Do not expect all CAT programs to give the exact same analysis. Algorithms used to compute the results are likely to be trade secrets, and each software maker probably has their own recipes and parameters. And then there are user-defined options (autojoin/split, segment end properties, TM/tag/placeable penalties...). For translators, the Homogeneity (MemoQ)/Internal match (Trados) trick mainly lowers translation prices : the difference between analyses with and without the option can be significant, depending on discount grids and text types. So much less for translators when you think how difficult it is for them to increase their base rate by eg. 20% to compensate the use of the option. Until v2007 included, "old Trados" (an ubiquitous CAT tool at the time) didn't have this feature, while SDLX had. With old Trados, this hidden and unmeasured leverage was all for us translators. And we were getting insanely wealthy because of this unchartered goldmine. Now, this option is the norm for most agencies who work with weighted wordcounts. Agencies who were Trados users during the transition from Trados to SDL Trados, and many of them were, have automatically benefited from a substantial discount over all their translation orders just checking a box. Smooth, clean and nice. And many of them don't even seem to be aware of this nice rebate brought to them by those "next-gen" CAT tools. The MemoQ Help file explains the benefits of this feature: This is one of memoQ's power features. [...] You are also able to give a much better estimation of your resources to be spent on translation than without homogeneity. If you use the analysis to give a quotation, always look for the aggregate results as they reflect the real productivity gain through using memoQ. So this option is useful to figure out a more accurate leadtime. Sure. Measuring internal leverage may have been designed for planning purposes. But in real life, the option transfers the productivity gain from internal matches to agencies, not to those who bought and own, maintain, troubleshoot and optimise the tool. Translators don't work less hours (internal matches exist even when they are not measured), but they are paid less. Christin Lee wrote: I mean if it's a new project, empty TM, there could only be 100% and reps, correct? 0% (no match) and repeats only, rather. Using your own discount grid, you can check the weighted wordcount discrepancy (ie cost difference) for yourself with and without the Homogeneity option. Philippe | | |
Well put, Philippe | May 20, 2014 |
I second every word. | |
|
|
Thanks a lot | May 24, 2014 |
Thank you so much, guys~ | | |