| Pages in topic: < [1 2] | Correcting alignment in Trados 2014 Thread poster: Tiffany Hardy
|
|---|
| If you really need WinAlign, Studio 2014 includes it | Feb 24, 2014 |
For those who need WinAlign, you should find it in one of these two folders:
"C:\Program Files (x86)\Common Files\SDL\T2007\TT\Winalign.exe"
or
"C:\Program Files\Common Files\SDL\T2007\TT\Winalign.exe"
It is still installed by default with Studio.
Daniel
| | | | Worder Company
Russian Federation
Local time: 10:19
Russian to English
+ ...
| if you do not need to align it, at all | Feb 25, 2014 |
if you do not need to align the files at all, but the "product if for internal, inter-team use", try doing this procedure:
it's easy and efficient (untill you have a lot of hits in your tm, but still copypasting would always beat the alignment procedure -- if you have pdfs that have a ... See more if you do not need to align the files at all, but the "product if for internal, inter-team use", try doing this procedure:
it's easy and efficient (untill you have a lot of hits in your tm, but still copypasting would always beat the alignment procedure -- if you have pdfs that have a lot of graphics, etc.)
http://www.proz.com/forum/sdl_trados_support/264215-translation_memory_without_alignment.html ▲ Collapse
| | | |
I fail to see the usefulness of that procedure. All that gives you is an indication of which file contains a hit. Then if you want to find the target language text, you need to open the two files side by side, identify the hit in the SL file and find the same place in the TL file.
Compare this to alignment: you autoalign all your files and generate a TM without doing a manual review. This takes about the same amount of time as your procedure, or maybe less, as you don't need to sort files ... See more I fail to see the usefulness of that procedure. All that gives you is an indication of which file contains a hit. Then if you want to find the target language text, you need to open the two files side by side, identify the hit in the SL file and find the same place in the TL file.
Compare this to alignment: you autoalign all your files and generate a TM without doing a manual review. This takes about the same amount of time as your procedure, or maybe less, as you don't need to sort files into separate folders. Then, when you get a TM hit, there are two possibilities:
1) the alignment is correct, the correct TL sentence is inserted into your translation. WIN!
2) the alignment is off, and the TL sentence is not the one that corresponds to the SL sentence. In this case, you open an xls file containing the aligned documents side by side. You find the SL sentence, and in 90+ per cent of cases, the correct TL sentence will be right there on your screen, one or two rows up or down. No need to open a second file and start rummaging around in it.
The upshot is that even if you only do a raw autoalignment, you still get what you need quicker than with your procedure, even if the autoalignment is 100% faulty, which it never is. If the source texts are of reasonable quality, autoalignments tend to be at least 90% correct. ▲ Collapse
| | | | Worder Company
Russian Federation
Local time: 10:19
Russian to English
+ ...
| In Reply to Comments by Farkas | Feb 26, 2014 |
Hi, Farkas.
Thanks a lot for commenting on the procedure.
I have atually posted a request for any comments or alternative procedures in the post "Translation Memory Without Alighnment"
But, as I have commented on the post "please exclude any references to winalign-based procedures".
You have to think about the scale, not the quality in case 1 and the other way around in case 2.
Case 1:
1. You have 50 files (this is... See more Hi, Farkas.
Thanks a lot for commenting on the procedure.
I have atually posted a request for any comments or alternative procedures in the post "Translation Memory Without Alighnment"
But, as I have commented on the post "please exclude any references to winalign-based procedures".
You have to think about the scale, not the quality in case 1 and the other way around in case 2.
Case 1:
1. You have 50 files (this is oil and gas for me), these 50 files are in pdfs, they contain various columns, names of details with arrows pointing to components at a figure, tables -- something like a magazine (not the economist type, but you know, like Vogue or Cosmopolitan). You don't have any time to do something about aligning these documents, but you WANT to have them in your TM. The verb "want" clearly identifies that it's you who decides on how you would process with the task in question. You might do alignment or you might do "pseudo-memory". If it is a formal request from a customer or a translation company, folks down there would not want to receive some kinda "pseudo-memory" (whatever that means), but they need to get a TM!
I have approximately 300 files (pdfs, Cosmopolitan-style in my pseudo-memory for a particular segment of oil and gas) that I use to identify any "large-scale" hits -- I mean 20-30 pages of text or when I desperately need to find a translation of a term (for some reasons, difficult term, first-time term, argument as to how to translate the term within the team).
If I send you 300 files today and ask you to prepare the TM in 1 week (with no money paid and consequently no time allocated by you as you are a freelancer), you might not want to do this... I, on the other hand, would be able to do this FOR FREE, as it's going take me only 2 hours to create the tmx files in memo, then put the second notebook nearby, and then just import a next tmx into my TM after the previous one has been completed. It's virtually going to take me 2 hours to create tmx files and click every 30 minutes afterwards.
Now, your procedure about winaligning and xls as a back-up. Have you ever done this with 50 files? If you do, how??? Because I have just opened a file with a 50-page four-column table with no dots at the end of sentences, numericals and letters-numericals all the place around and lots of schematics. How on eath would you be able to take A SINGLE file into xl, put the the source and target near each other? How would you be able to align the file, if, (such cases are reasonably expected, as a specialists might and does require translation of a particular section of a file), if you have a translation for only 23 out of 72 pages in file 25, the rest of the file has not been translated for some reason. And such a file has translation of pages 1-6, 20-32, 64-72? Just imagine, opening the win-align results and doing something about the aligned results!
Now, you can 1) automatically align a file and then add to the memory (you have 50 files, 60 pages each in source language), 2) ??? I really don't know what you are going to do next. Probably, you would say "bummer" and close the winalign. Because if you import the winalign results in your TM, you would get mislaligned text all the way down (read- downtown), so you would reasonably decide not to do anything at all. Now, you can import the winalign result into your TM and PLACE A MARKER FOR THE SEGMENTS FOR THIS PARTICULAR FILE. Then, you would have to put the source and target in a folder, name it "5". Then, when you translate you go to that file and copypaste the term/text. (no xl is feasible due to complexity of the file; anyway, even if it is feabile to import the files in xl, what you gonna do when you 15 files in your xl, create sheet number 2, number 3, -- I mean, once again, just place files in the folders; in my case, translation from English into Russian, you would have (after 5 files) English text at line 1000 and Russian text at line 1234, just because Russian is longer and you do not have any identification as to when one sentence is finished in the Russian translation as you don't have dots, no nothing, nada).
Case 2:
You never work with large-scale projects, TMs and "files for fun" (i.e. nothing required to be done with them, just in case you might be bored and read, like a couple of ks of pages with a cup of tea before going to sleep). You receive a 30-page file, you mis-align it, you place it in xl, you open xl and find the needed stuff in xl. OR you can actually get it done correctly, as you ARE REQUIRED to prepare a TM, so that you or other translators may use the TM and won't have to "rammage" across the your files register system.
So, in brief, been there done that. I mean, if you need a solution for a small-scale project -- like 100 pages in each of 3 files and that's it. Then winalign them, add them to your TM, open your 3 pds and ctrl-f the terms in those one by one (as any import into xl would not be feasible). OR you can just create three folders, name them 1, 2, 3, create pseudo-memory. When you have a hit, see to your right that it's file No. 2, open the source and target, find what you need. Even when we speak about such a small-scale project, this folder system beats the procedure you describe at any time (once again, import in xl is not feasible; because the moment it becomes feasible, you just import the source and target into xl and winalign the xl or prep the tmx in memo).
P.S.
Just take a Cosmopolitan and try creating any usable TM on the basis of this file (if you have tachnical documentation and specifications, or marketing materials, it's just the same in terms of formating in a file; it's not a 50 contract in word with separate passages in separate lines of word table).
http://www.snobessentials.com/2012/11/bag-snob-tinas-column-in-china-cosmopolitan.html
try winalign and xl back-up a page in the link.
I mean, c'mon, know-it-all approach never works when you have peer-review-style forums like this one. I really think that if an interpreter takes on this approach, they might benefit from it immensely (as I have been doing for quite some time), but if an interpreter reads some post "NO" like this one, they would think -- it's a bad procedure, but the fact is -- you show me a better one, I would send you 100 bucks!
If you fail to see the usefulness of this procedure, it means nothing actually. I mean I know several translators who see no usefulness in trados or cats as such, or I know a guy who sees no usefullness in using shortcuts in trados, what's you gonna about that? But, anyhow, these folks would not not take time to write a peer review on somebody's procedure? ▲ Collapse
| | |
|
|
|
I will read your full post later and come back with detailed comments if needed, but here's a quick response.
Perhaps you misunderstood me. I never ever considered using WinAlign for anything. I was referring to using an aligner capable of autoalignment: my own LF Aligner or a commercial aligner that has an autoalignment feature. "Autoalignment" means that the software has a built-in algorithm for determining which sentence goes with which - this allows you to a) auto-generate TMs w... See more I will read your full post later and come back with detailed comments if needed, but here's a quick response.
Perhaps you misunderstood me. I never ever considered using WinAlign for anything. I was referring to using an aligner capable of autoalignment: my own LF Aligner or a commercial aligner that has an autoalignment feature. "Autoalignment" means that the software has a built-in algorithm for determining which sentence goes with which - this allows you to a) auto-generate TMs without manually pairing up sentences or b) generate TMs MUCH quicker and much easier if you decide to do a manual review - 10x less corrections to do means 10x less time. Sure, in some (rare) cases very messy files may no be worth aligning. This occasionally happens with pdf files due to terrible authoring, or in file pairs where there is a drastic mismatch between the two texts (SL text is twice as long as the TL etc.)
Trust me when I say I know what it takes to align many files. In the past year, I aligned approximately 2 million files to create a large multilingual corpus of about 200 million sentences. I'm working on processing another 100,000 or so files right now. (I don't think I'm allowed to post links here but you can find it if you google around.)
[Edited at 2014-02-26 18:30 GMT] ▲ Collapse
| | | |
Well, I read your whole post now and tried to make sense of it. Let's just say you haven't convinced me.
You seem to go round and round in circles describing this situation:
"in some (rare) cases very messy files may no be worth aligning. This occasionally happens with pdf files due to terrible authoring, or in file pairs where there is a drastic mismatch between the two texts (SL text is twice as long as the TL etc.)"
If the files are that bad, you are often better off ... See more Well, I read your whole post now and tried to make sense of it. Let's just say you haven't convinced me.
You seem to go round and round in circles describing this situation:
"in some (rare) cases very messy files may no be worth aligning. This occasionally happens with pdf files due to terrible authoring, or in file pairs where there is a drastic mismatch between the two texts (SL text is twice as long as the TL etc.)"
If the files are that bad, you are often better off discarding them, because whenever you get a hit, you may spend several minutes looking for the corresponding sentence in the TL file and you may not find it. ▲ Collapse
| | | | Worder Company
Russian Federation
Local time: 10:19
Russian to English
+ ...
| not worth aligning? | Feb 27, 2014 |
It's always been about "bad files", that would not be easy (or event possible) to align. And there hasn't been question -- to incorporate or to discard. If I wanted to discard them, I surely would. But, the question is that you want to incorporate the files, but you don't want to incoporate the files into your other TMs + you don't have tools to do this. If files are at least "not bad", I would quickly align them somewhere, like in memo, and incorporate them in one of the "real" TMs.
... See more It's always been about "bad files", that would not be easy (or event possible) to align. And there hasn't been question -- to incorporate or to discard. If I wanted to discard them, I surely would. But, the question is that you want to incorporate the files, but you don't want to incoporate the files into your other TMs + you don't have tools to do this. If files are at least "not bad", I would quickly align them somewhere, like in memo, and incorporate them in one of the "real" TMs.
About the time concern you have:
So, in terms of time: you have to translate 100 pages and you get 50 pdfs (50 pages each). You spend two-three hours to prep them and then just click every 30 minutes to put in a new tmx into your TM. (You might expect high probability of large-scale hits, like 10 pages from a file in your TM, or 5 pages from file 1 and 5 pages from file 2.)
My baseline speed is 2 pages per hour.
the costs to create the TM: 4 hours
Costs/Benefits Analysis: you would spend 5 hours to translate 10 pages, but you spend ca 1 hour (15 minutes to copy-paste + 1 hour to proof-read), thus your savings is 4 hours. So, your costs to prep the TM is 0 hours. If you locate 2 more pages, you would get the profit of 1 hour.
So, in terms of locating the hits of sentences in your TM, you are not correct about c/b analysis. But, again, I am a translator and I translate 5,000 words per day, 365 days a year and I don't align gazzilions of pages per year or decades, bacause I don't get paid to do that + the majority of "grown-up" customers come in or, more correctly, I start up on projects for such customers, who have 1-gigabite TMs with lots of termsbases for various fields. If you can read and brush up on the terminology for your customer's particular projects by reading 10 pages that have already been translated, it's a great benefit, imho.
In terms of locating terms in TM: I can spend up to 10 minutes to decide on a translaton of a term with a frequency of 1 term per 3 pages. // You use concord and you find this term in the file. You open the source file, see the number of the clause, open the target file, ctrl+F the clause number, find you term and you're done. I am not a professional professional translator, so right now for me the best option is to use a legacy term versus making up my own translations.
Two months ago I got an order for 100 pages (gym equipment, manuals, etc.), I pre-translated the files in trados with no cross-file hits found. I received 50 previously translated docs. I made the "pseudo-TM" in no time and found heaps of cross-file hits. The point is that the legacy files contained the translations for some older models, while this pack of docs had the new models with low inter-file correlation. Anyway, the customer would naturally not know about leveraging their legacy files (however, they might be sure that there might have been a lot of hits in the translated files and the legacy files).
So, I spent about 4 hours to prep the tmx files and then was clicking for 2 days every 30 minutes. The amount of hits was about 50 pages. The selling price was 100 pages * 10 bucks = 1 000 bucks. It generally takes you 1 hour to copy-paste + proof read 10 pages, so 50 pages / 10 pages per hour = 5 hours - (50 pages / 2 pages per hour = 25 hours), so savings = 15 hours * 2 pages * 10 bucks = 300 bucks.
+ I quickly created a glossary of the terms for the other 50 pages and the translation would have reasonably been more difficult with no leveraging legacy files.
If I didn't know about this procedure, I would not do the alignment, because I would never be able to calculate the time costs you would need to work thru (50 legacy files * 20 pages each) / ca 20 pages per hour = 50 hours. And if I started up this process, I would want to come back with a fully functional TM to bring back to the customer, because the next time they send it over to the next translator, that translator would tell them that the TM is misaligned, which they would pass on to me. Consequently, I would need to align it 100%. In addition, aligning files is extremely boring. I usually ask somebody to align the files and set off the costs versus the translation performed...
(I've never aligned gazzilions of pages.) I aligned once 1,000 pages of your best possible words. It took me ages. ▲ Collapse
| | | | | Has anything been done? | Nov 5, 2014 |
I am still looking for a way to edit alignment files, the way we could do it in Trados 2007.
Having to do it by hand was a harsh process, but not being able to do it at all is a major set-back!
| | |
|
|
|
MollyRose 
United States
Local time: 02:19
English to Spanish
+ ...
| If you still have Trados 2011 | Apr 22, 2015 |
you can still use WinAlign in the 2011 version and fix the errors and then import your alignment into your TM. I did this (I'm glad the lightbulb came on, finally!) and it worked just like when I was using Studio 2011.
| | | | Walter Blaser
Switzerland
Local time: 09:19
French to German
+ ...
| You can still use WinAlign (even with Studio 2014) .... | Apr 22, 2015 |
... if you prefer this to the new alignment editor integrated in Studio 2014.
You can find the old WinAlign in the following folder:
"C:\Program Files (x86)\Common Files\SDL\T2007\TT\Winalign.exe"
Locate this file in Windows Explorer, then create a shortcut for it, on your desktop or in the Start menu, so you can easily launch it when you need to.
Walter
| | | | | in SDL Trados 2014 SP1 | Sep 2, 2015 |
Please
could I delete in alignment windows all segments? Because by mistake I put wrong text. I can t find some button and need restart all Trados
Thank you
| | | | | Pages in topic: < [1 2] | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Correcting alignment in Trados 2014 | Protemos translation business management system | Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!
The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.
More info » |
| | TM-Town | Manage your TMs and Terms ... and boost your translation business
Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news related to SDL Trados
