Fluency Translation Suite Thread poster: Pablo Bouvier
|
|---|
Do you think that a forum about Fluency Translation Suite would be useful? I am not a fan of CAT tools, but I have seen some utilities in Fluency that in my humble opinion deserve to be monitored.
For example, automatic machine translation acceptance if Google Translator and Bing translations match. I guess we should not forget that both, GT as like Bing, base part of their translations in human translations and that may be this will be the seed of some kind of translation AI in ... See more Do you think that a forum about Fluency Translation Suite would be useful? I am not a fan of CAT tools, but I have seen some utilities in Fluency that in my humble opinion deserve to be monitored.
For example, automatic machine translation acceptance if Google Translator and Bing translations match. I guess we should not forget that both, GT as like Bing, base part of their translations in human translations and that may be this will be the seed of some kind of translation AI in the next future.
[Edited at 2010-10-23 01:10 GMT] ▲ Collapse
| | | |
Hi, all!
I downloaded a demo version of the software. The installation of the program on Windows 7 went with no problems and the main window appeared. After a quick overview of the functions and menus, I decided to take if for a spin and tried to translate a doc file with an average formatting difficulty. The import of the file went well, but then the first glitches started to appear. At first, my impression was that it is not a finished product, that is still a work in progress and it is ... See more Hi, all!
I downloaded a demo version of the software. The installation of the program on Windows 7 went with no problems and the main window appeared. After a quick overview of the functions and menus, I decided to take if for a spin and tried to translate a doc file with an average formatting difficulty. The import of the file went well, but then the first glitches started to appear. At first, my impression was that it is not a finished product, that is still a work in progress and it is only a half-baked product. The first impression was confirmed after a while, when the software tried to update, but instead did not respond for at least half an hour. After restarting the application, it maybe started a loop, because I was not able anymore the pass by the first windows after the update, so I could not use the program anymore. I tried to reinstall it, used the repair function of the installer, but to no avail, so I decided to uninstall it.
The short period of time I was able to use the program was enough to get an overall impression of the program. The author / authors praises the integration of terminology dictionaries in the software. While the initiative is certainly to be praised, in my case, for Romanian, the dictionary was of poor quality. In fact, many words lacked diacritics, some were wrongly spelled, some had illegible characters and I think the company (whose site was actually down the last time I checked) selling the dictionary takes some freely available dictionaries from the net without editing and just puts them on display. So what should have been a strong point proved in the end to be a frustrating experience. Here I would like to see a feature implemented as in Transit XV: when right clicking a word, one should have a context menu with the possibility to choose from words from the dictionary, let's say by entering a number. This could be a competitive advantage over other programs, so the programmers could take into account this proposal.
I like to use shortcuts a lot, so I missed the fact that the menus had no underlined letters to navigate easily within the menus with the keyboard.
The segmentation seemed to be ok for the file I imported, also there seems to be no support for SRX segmentation rules. I had no luck editing the initial automatic segmentation. I entered some words in the manual segmentation rules, for instance it should not end the segment for "nr.", but I could not test the results after reimporting the files because, as I mentioned before, the program crashed and I could not use it anymore.
In conclusion, there are some technical glitches which prevented me from using the software at its full potential and I honestly think the program still needs a lot of work. As stated before, the program seems to be still in a prolonged beta stage and, at least for me, not ready for productive work.
[Edited at 2011-06-04 05:02 GMT] ▲ Collapse
| | | |
I'm one of the programmers of Fluency and I wanted to take a chance to respond to Leon's complaints. Fluency has been released and has been used by translators all over the world for over 8 months now. The words "half-baked" and "not ready for productive work" would seem like quite a stretch to the many people who are currently using Fluency for production work. We do understand that as with all software, there will be "technical glitches," but Fluency is far removed from the beta stage. We a... See more I'm one of the programmers of Fluency and I wanted to take a chance to respond to Leon's complaints. Fluency has been released and has been used by translators all over the world for over 8 months now. The words "half-baked" and "not ready for productive work" would seem like quite a stretch to the many people who are currently using Fluency for production work. We do understand that as with all software, there will be "technical glitches," but Fluency is far removed from the beta stage. We are more than happy to work through any installation specific issues when we are contacted directly and go to any length to make sure Fluency is working correctly.
As far as our site is concerned, we released a marketing email from ProZ last week and had an overwhelming response, with over 500 people downloading the software at the same time which caused issues with our servers. These were fixed within 8 hours.
As far as terminology is concerned, we did not just take free dictionaries and make them available. The terminology is licensed from a third party provider who gathered the terminology over decades from LSPs and translators. We have not personally verified the millions of terms in every language and apologize for any errors. We have ongoing QA efforts in that area and welcome any specific feedback.
As far as the segmentation is concerned, we have considered adding SRX support, but our users have not requested such support. We do have regular expression based segmentation rules and exceptions that can be added to or edited. If SRX is a priority to any of our users, it will be a priority to us.
Despite some letters in the menus not being underlined (some are), you can still use the keyboard to navigate the menus. Although we have strived to make Fluency relatively simple, there is still a great deal of complexity to deal with, and one of the main things that we are trying to roll out in a conscientious way is keyboard shortcuts which may conflict with menu shortcuts. Once we have enough feedback and have determined the best way to move forward, we will make more of these features available.
We are happy to have constructive feedback from our users and offer 1st rate technical support. Leon, thank you for your feedback and please feel free to contact us at [email protected] if you'd like us to work through your installation issues with you.
[Edited at 2011-06-07 15:48 GMT] ▲ Collapse
| | | | Samuel Murray 
Netherlands
Local time: 05:18
Member (2006)
English to Afrikaans
+ ...
Leon Tiberiu wrote:
I ... tried to translate a doc file with an average formatting difficulty. ... then the first glitches started to appear. At first, my impression was that it is not a finished product, that is still a work in progress and it is only a half-baked product.
It seems from your description that your impression that it is a halfbaked product is based mostly on the fact that the update failed. What is your impression about the tool as a CAT tool itself?
I also downloaded and installed it just now. Downloading the 350 MB file took 3 hours on my 10 Mbps line, but installing it took less than a minute. I'm using Windows XP Pro SP2. Fluency did try to connect to the internet when I ran it the first time (but I denied it), so I can't comment on the update feature, unfortunately.
From what I've seen, Fluency is your average CAT tool. It imported a bilingual RTF file in lieu of a TM without a hitch, and was able to translate a file without any problems. I disliked the fact that the glossary pane showed a "match" for each word in the source text, and for words without matches it showed the source text as the target text in the glossary pane. I would have preferred it if the glossary pane showed only word that have actual matches, and not all words regardless of whether they have matches.
I also watched the video tutorial on Youtube. The TM viewer/editor looks nice and user-friendly. The aligner is elementary but seems quite useful and nicely integrated into the whole.
On another mailing list, a colleague said that he had heard that Fluency supports up to 20 different TMs, but from the video it became clear that Fluency's people use the word "translation memory" to mean "translation unit" or "fuzzy matched segment", where the "20" came from is simply that one can see up to 20 matches from the translation memory in the TM pane.
The author / authors praises the integration of terminology dictionaries in the software.
Fluency seems to make no distinction between glossaries and dictionaries. I wonder what glossary formats it supports for importing.
One thing I'm a little concerned about is the fact that it would appear (from comments in the video) that there is no one-to-many matching in Fluency, and that you can't translate non-unique segments in unique ways. Is this so?
| | |
|
|
|
Marco Ramón
Mexico
Local time: 21:18
Member (2007)
English to Spanish
+ ...
MODERATOR | Tried it with TM's import | Jun 8, 2011 |
It amazed me how intuitive was the interface to import TM's. And how fast. I imported half a million TU's of the EU Commission Corpora - EN-ES only. It did ti in few minutes. ( I can't help to mention that - Deja Vu X 2 crashed miserably every time I tried to import just one of the twelve files that Fluency managed beautifully). But, Fluency crashed, too.
It crashed when trying to find duplicates. And then it started giving weird "error" messages.
Sad.
I have no ... See more It amazed me how intuitive was the interface to import TM's. And how fast. I imported half a million TU's of the EU Commission Corpora - EN-ES only. It did ti in few minutes. ( I can't help to mention that - Deja Vu X 2 crashed miserably every time I tried to import just one of the twelve files that Fluency managed beautifully). But, Fluency crashed, too.
It crashed when trying to find duplicates. And then it started giving weird "error" messages.
Sad.
I have no quarrel with Fluency, since I have not purchased the product. But it really, really, looks half-baked, I agree. ▲ Collapse
| | | |
Thank you for your informed response. In answer to your questions:
I would have preferred it if the glossary pane showed only word that have actual matches, and not all words regardless of whether they have matches.
If you turn off Fluency terminology (this option is available in the Terminology preferences menu) you will only see words in the glossary section that have hits on your client/personal terminology. We will make the change so that if you l... See more Thank you for your informed response. In answer to your questions:
I would have preferred it if the glossary pane showed only word that have actual matches, and not all words regardless of whether they have matches.
If you turn off Fluency terminology (this option is available in the Terminology preferences menu) you will only see words in the glossary section that have hits on your client/personal terminology. We will make the change so that if you leave Fluency terminology on, words that don't have a hit will not be copied over to the target side of the glossary pane.
On another mailing list, a colleague said that he had heard that Fluency supports up to 20 different TMs, but from the video it became clear that Fluency's people use the word "translation memory" to mean "translation unit" or "fuzzy matched segment", where the "20" came from is simply that one can see up to 20 matches from the translation memory in the TM pane.
I think you're correct with the misuse of terms in the video. We will have a new video out in the near future. It is the case, however, that Fluency uses a TM "super container" that imports as many TM files as you want. TMs are differentiated in the super container by client, project, domain, and locale.
Fluency seems to make no distinction between glossaries and dictionaries. I wonder what glossary formats it supports for importing.
We are trying to better clarify our use of these terms. Fluency currently comes with bilingual glossaries/terminology built in, and monolingual dictionaries. For personal terminology/glossary imports, we accept SDL Multi-term, Excel, csv, tab delimited, and TBX files.
One thing I'm a little concerned about is the fact that it would appear (from comments in the video) that there is no one-to-many matching in Fluency, and that you can't translate non-unique segments in unique ways. Is this so?
One-to-many is possible when the "many" are from different clients or different locales. One TU can only have one translation for a given client and locale. This is done for consistency's sake. If this is not acceptable for your needs, please let us know how we could make it better.
Thanks for your interest and feedback. Please let us know if there is more that we can do to support you.
[Edited at 2011-06-08 20:19 GMT]
[Edited at 2011-06-08 20:22 GMT] ▲ Collapse
| | | |
I'm sorry for the issues that you experienced with duplicates. If you want us to work through these issues for you, please contact us at [email protected]
It would seem that by the definition being applied to "half-baked" in this thread, all software is "half-baked," meaning all software has bugs/glitches. That is the reality of software. Our users that use our tool daily encounter problems now-and-again despite th... See more I'm sorry for the issues that you experienced with duplicates. If you want us to work through these issues for you, please contact us at [email protected]
It would seem that by the definition being applied to "half-baked" in this thread, all software is "half-baked," meaning all software has bugs/glitches. That is the reality of software. Our users that use our tool daily encounter problems now-and-again despite the testing that we do in-house. We certainly are not exempt from that, but do take pride in being responsive to our users when they contact us.
[Edited at 2011-06-08 20:33 GMT]
[Edited at 2011-06-08 20:34 GMT] ▲ Collapse
| | | | Samuel Murray
Netherlands
Local time: 05:18
Member (2006)
English to Afrikaans
+ ...
| @Fluency support | Jun 9, 2011 |
rtregaskis wrote:
I think you're correct with the misuse of terms in the video. We will have a new video out in the near future. It is the case, however, that Fluency uses a TM "super container" that imports as many TM files as you want. TMs are differentiated in the super container by client, project, domain, and locale.
The terms I referred to also occur on Fluency's GUI, so the video itself isn't to blame.
Ah, so Fluency has only "one" TM, but that is not a limitation, since whenever you import or export TMs, you have to (or can?) specify a "domain" (an attribute that is added to each TU from that TM) so that you can filter TM results based on that domain. One can also filter on domain in the TM editor, if I remember correctly.
We are trying to better clarify our use of these terms. Fluency currently comes with bilingual glossaries/terminology built in, and monolingual dictionaries.
I suppose the words have different meanings for different people. Some people use the word "glossary" for any dictionary that is bilingual and/or that does not have definitions. When I used the word I meant a specific prescriptive list of terminology.
I'm under the impression that a user using Fluency will see matches from the general glossary (or bilingual dictionary) mixed with matches from his own specific glossary (prescriptive terminology list) and that there is no way to separate the two types of matches, is that so? I.e. whe I see glossary match in Fluency, there is no way of knowing whether the match comes from the large general dictionary or from my own, small terminology list (and therefore no way of telling whether the term is optional).
[quiote]One-to-many is possible when the "many" are from different clients or different locales. One TU can only have one translation for a given client and locale.[/quote]
This is generally regarded as a deficiency (though developers or activists would often try to paint it as an advantage, or play down the disadvantages). Let me see if I understand correctly, with an example:
You can:
* Brush your hair
* Clean your teeth
* Wash your face
You can do the following:
* Brush your hair
* Clean your teeth
* Wash your face
In English, the three items in the two lists are identical, but in my language, the word order of the list items depends on the sentence that precedes it. In the above case, the correct translation of the first three items would be different from the second three items. If these two lists occur in a single text, can a translator translate them differently in Fluency? And if so, which translation would be added to the TM -- the first or the second?
| | |
|
|
|
The general Fluency terminology is differentiated both in the gloss pane (any cell containing a gloss that comes from your terminology is shaded blue) and in the extended glossary section (each of your terminology entries is proceeded by ***) that displays client/project/notes/etc term info.
The TM/TU distinction is a little fuzzy in Fluency because of the "one" TM super-container. We do try to give a pseudo "TM" count by counting the unique number of projects in our super-container... See more The general Fluency terminology is differentiated both in the gloss pane (any cell containing a gloss that comes from your terminology is shaded blue) and in the extended glossary section (each of your terminology entries is proceeded by ***) that displays client/project/notes/etc term info.
The TM/TU distinction is a little fuzzy in Fluency because of the "one" TM super-container. We do try to give a pseudo "TM" count by counting the unique number of projects in our super-container, assuming each project has its own "TM". You can, of course, export your TUs from Fluency in a lot of different ways, not just based on project name, so the actual number of TMs is completely subjective. Each TU is labeled with the client/project/domain/locale and can be exported/managed by filtering on any of these fields. Domain, project, and client are optional fields though strongly recommended, especially if you need to export a TM to return to the client (Fluency will automatically insert a project name for you unless told otherwise).
As far as the one-to-many question, you can translate them uniquely but only the most recent translation will be stored in the TM (for a given client). The hackish work-around would be to create a larger unique TU, but we will certainly take your feedback into consideration, thank you. ▲ Collapse
| | | | There is no moderator assigned specifically to this forum. To report site rules violations or get help, please contact site staff » Fluency Translation Suite | Trados Studio 2022 Freelance | The leading translation software used by over 270,000 translators.
Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop
and cloud solution, empowering you to work in the most efficient and cost-effective way.
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... | | | | | |
Login to reply/comment 
