How can the word-count analysis differ several hundred thousand words between Memsource and Studio? 论题张贴者: Fredrik Pettersson
|
How can the word-count analysis differ several hundred thousand words between Memsource and Studio? I applied the two TMs I received from the customer when running the analysis in SDL Trados Studio 2021, but I still get a total of 411,000 words whereas the customer gets a total of 284,752 words in his Memsource analysis. He gets new words 63,611 in Memsource, I get new words 121,838 in SDL Trados Studio. | | | Samuel Murray 荷兰 Local time: 00:28 正式会员 (自2006) English英语译成Afrikaans南非语 + ...
Fredrik Pettersson wrote: I applied the two TMs I received from the customer when running the analysis in SDL Trados Studio 2021, but I still get a total of 411,000 words whereas the customer gets a total of 284,752 words in his Memsource analysis. He gets new words 63,611 in Memsource, I get new words 121,838 in SDL Trados Studio. Did he send you a complete analysis report or just a short summary of one? Perhaps your client can shed light on this -- did you ask him? It may be that he applies certain settings in Memsource that results in this discrepancy. Are you sure the two TMs that he sent you are the only TMs that he applies? | | | Memsource excludes the confirmed segments | Sep 20, 2021 |
I received the explanation that Memsource excludes the confirmed segments. But there is still such a discrepancy between the different bands of the fuzzy matches. Can also that be explained by the exclusion by Memsource of the confirmed segments? It might be that the customer applies a third TM, which would be their server TM, I will ask him. | | | Memsource wordcount | Nov 19, 2021 |
Hi Frederik, I discovered that, if you download the .doc file, you'll see that you have a table with the original AND the pre-translation from TM. I believe this explains the difference. On the other hand, if you select all the files at the portal and double-click over them, you will open all the files joined and will get the same wordcount sent by the client. But this does not happen if you use de Desktop Editor. ANYWAY, whenever I translate using MemSour... See more Hi Frederik, I discovered that, if you download the .doc file, you'll see that you have a table with the original AND the pre-translation from TM. I believe this explains the difference. On the other hand, if you select all the files at the portal and double-click over them, you will open all the files joined and will get the same wordcount sent by the client. But this does not happen if you use de Desktop Editor. ANYWAY, whenever I translate using MemSource, I feel that I am working twice as much as with any other software. I always feel the wordcount is much higher. And MemSource does not allow you to run your own analysis. Unless, of course, you buy the software, which costs an arm and a leg! Best, Heloisa
[Edited at 2021-11-19 19:53 GMT] ▲ Collapse | |
|
|
Had a somewhat similar problem | Nov 24, 2021 |
I've had a somewhat similar problem with MemSource. See my topics: "MemSource XLIFF analysis gives 3x as many repetitions vs. SDL Analyse on an exported DOCX" in MemSource Support "MXLIFF word count, including repetitions—alternatives to MemSource?" in CAT Tools Technical Help (For some reasons, this post doesn't work correctly with links.) "ANYWAY, whenever I translate using MemSource, I feel that I am working twice as much as with any other... See more I've had a somewhat similar problem with MemSource. See my topics: "MemSource XLIFF analysis gives 3x as many repetitions vs. SDL Analyse on an exported DOCX" in MemSource Support "MXLIFF word count, including repetitions—alternatives to MemSource?" in CAT Tools Technical Help (For some reasons, this post doesn't work correctly with links.) "ANYWAY, whenever I translate using MemSource, I feel that I am working twice as much as with any other software." My sentiments exactly. Well, they can argue that there's nothing wrong with the software/service per se, and it's probably true technically, but I'm afraid that much of MemSource success is due to the powerful ways it offers to screw the translator.
[Edited at 2021-11-24 21:56 GMT] ▲ Collapse | | | Roy Oestensen 丹麦 Local time: 00:28 正式会员 (自2010) English英语译成Norwegian(Bokmal)挪威语(波克默尔语) + ... | Working twice as much with Memsource, I agree | Nov 26, 2021 |
I have another customer that, for some reason, have a preliminary word-count stated when the job offer is sent out to freelance translators. This preliminary word-count is then adjusted afterwards when the translator has delivered the completed translation. And it's not minor adjustments we are talking about, I think I've experienced around 40 percent less payment in post-adjustments. This made me react of course, and I asked for the reasons to this. I don't remember the details in the agency's ... See more I have another customer that, for some reason, have a preliminary word-count stated when the job offer is sent out to freelance translators. This preliminary word-count is then adjusted afterwards when the translator has delivered the completed translation. And it's not minor adjustments we are talking about, I think I've experienced around 40 percent less payment in post-adjustments. This made me react of course, and I asked for the reasons to this. I don't remember the details in the agency's reply, but no matter the explanations, this adds up to what I hear from you other fellow translators about Memsource word-counts: That it's not a reliable way of counting words and it's a way of counting words which renders outputs that differ considerably from most other CAT tools. Furthermore, Memsource doesn't fall into the top category of advanced CAT tools but is a rather simple CAT tool with limited functionality. Wouldn't that limited capacity include limitations as to detailed and accurate word-counts as well? There really should be some universal word-count tool to resort to in cases like this when a less advanced CAT tool is used. ▲ Collapse | | | Easy to compare | Nov 27, 2021 |
You can easily check how Memsource counts. Memsource skips bullets and automatic numbering. You don’t translate bullets, do you? The longer your text is, the more such elements it contains. I tried to compare a “clean” text without bullets and other noise, and Memsource gave me exactly the same word count. Here are some images proving that: <... See more You can easily check how Memsource counts. Memsource skips bullets and automatic numbering. You don’t translate bullets, do you? The longer your text is, the more such elements it contains. I tried to compare a “clean” text without bullets and other noise, and Memsource gave me exactly the same word count. Here are some images proving that: https://www.proz.com/forum/memsource_support/348913-unrealistic_analysis.html In terms of work effort, Memsource shows you a realistic picture net of noise like bullets, autonumbering, and nontranlatables.
[Edited at 2021-11-27 10:13 GMT] ▲ Collapse | |
|
|
Roy Oestensen 丹麦 Local time: 00:28 正式会员 (自2010) English英语译成Norwegian(Bokmal)挪威语(波克默尔语) + ... Different word count before and after the translation??? | Nov 29, 2021 |
Fredrik Pettersson wrote: I have another customer that, for some reason, have a preliminary word-count stated when the job offer is sent out to freelance translators. This preliminary word-count is then adjusted afterwards when the translator has delivered the completed translation. And it's not minor adjustments we are talking about, I think I've experienced around 40 percent less payment in post-adjustments. This made me react of course, and I asked for the reasons to this. I don't remember the details in the agency's reply, but no matter the explanations, this adds up to what I hear from you other fellow translators about Memsource word-counts: That it's not a reliable way of counting words and it's a way of counting words which renders outputs that differ considerably from most other CAT tools. Furthermore, Memsource doesn't fall into the top category of advanced CAT tools but is a rather simple CAT tool with limited functionality. Wouldn't that limited capacity include limitations as to detailed and accurate word-counts as well? There really should be some universal word-count tool to resort to in cases like this when a less advanced CAT tool is used. I can see that the weighted word count would be different after a translation than before, since then all the segments would be 100% matches. But post-adjustment? Perhaps there are more than one translator working on a project, and then they take into account the contribution from each when counting the translation for the others. I would say I would not accept this. When I start a translation, I expect to have an analysis I will be paid according to, and the agency would not adjust it downwards afterwards. If they did, I would say farewell to them, since I would not find them trustworthy or willing to pay what a translation is worth. Roy | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » How can the word-count analysis differ several hundred thousand words between Memsource and Studio? 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 » |
| Wordfast Pro | Translation Memory Software for Any Platform
Exclusive discount for ProZ.com users!
Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value
Buy now! » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |