Pages in topic:   [1 2] >
Percentage deducted for Matching
Thread poster: Peter Simon

Peter Simon  Identity Verified
Netherlands
Local time: 20:56
Member (2013)
English to Hungarian
+ ...
Mar 14, 2014

Dear Colleagues,

I am quite new to this brave new world with CAT tools, and I've always found the idea of using them a bit weird. Now that I use one, I am mostly displeased, especially when I think about the cost of buying it. I'm also pestered by the idea of "Term agreement" and similar things when checking existing machine-translated texts, or proofreading work of others. How do we and the tools define agreement, based on which one gets a lower % of his/her rate if this "agreement" is high? To my mind, non-related languages can never be in agreement, but on word level, there can be apparent agreement. However, even though the words themselves may be the same, the target text may have lost its meaning due to endings, suffixes and the like among certain language pairs. This may be less apparent among more related languages, like among Dutch, German and English.

I've just wrote a post in my blog about this, using examples between English and Hungarian. If you are interested, please read http://learnenglishinnetherlands.com/2014/03/14/the-extent-translation-is-correct/. Thank you for your kind replies.


Direct link Reply with quote
 

Erik Freitag  Identity Verified
Germany
Local time: 20:56
Member (2006)
Dutch to German
+ ...
MT vs. CAT Mar 14, 2014

Dear Peter,

I've only had a brief glance at your post and your blog entry. You seem to be seriously confusing MT (Machine Translation) and CAT (Computer Assisted Translation). These are two completely different things (although both can be used together).

Kind regards,
Erik


Direct link Reply with quote
 

Peter Simon  Identity Verified
Netherlands
Local time: 20:56
Member (2013)
English to Hungarian
+ ...
TOPIC STARTER
correct Mar 14, 2014

Erik, thank you for pointing this out, you're right I think, but that's because I'm new to this non-human side of the industry. Also, I don't essentially see much difference between them, and from the point of view of usability, they don't make a difference in my major pair for me either. Both are both a bit helpful but mess up my work as well. So I would say my opinion and the question is also about both because they are both about non-human translation. My question, however, also concerns proofreading fees as they may also be decided based on perceived 'agreement' percentages, but I may be wrong.

Direct link Reply with quote
 

Erik Freitag  Identity Verified
Germany
Local time: 20:56
Member (2006)
Dutch to German
+ ...
Misunderstanding: CAT vs. MT Mar 14, 2014

Peter,

If this interests you, I really suggest you do some more research.

Peter Simon wrote:
I don't essentially see much difference between them, and from the point of view of usability, they don't make a difference in my major pair for me either


That's because you don't know yet what a CAT tool does. You could start by reading a Wikipedia article about CAT tools.

Peter Simon wrote:
… they are both about non-human translation.


No, they aren't.

Kind regards,
Erik


Direct link Reply with quote
 

Peter Simon  Identity Verified
Netherlands
Local time: 20:56
Member (2013)
English to Hungarian
+ ...
TOPIC STARTER
Not really Mar 14, 2014

Erik, you also misunderstand my intentions. CAT tools are not my interests (in my general view, they've given a lot of work and profit for those developing them while their use costs us a lot of money and yields us less and less income), and I've already done research earlier, thank you, but while I'm using a CAT tool (Wikip: "Some advanced computer-assisted translation solutions include controlled machine translation (MT), and mine is higher-priced), I find it quite of low value in practice. That's my problem.

But my original Q seems to have been overlooked: if I use the tool, which is capable of rendering % to my solutions, what is my loss if the tool (incorrectly, in lot of practical cases) declares that there's a high % of correlation with the one suggested by the tool, based on somebody's earlier solutions (mine, among them) or MT solutions that are incorporated in the tool.


Direct link Reply with quote
 

Sheila Wilson  Identity Verified
Spain
Local time: 19:56
Member (2007)
English
+ ...
Is the question about CAT usage, or CAT discounts? Mar 14, 2014

Peter Simon wrote:
But my original Q seems to have been overlooked: if I use the tool, which is capable of rendering % to my solutions, what is my loss if the tool (incorrectly, in lot of practical cases) declares that there's a high % of correlation with the one suggested by the tool, based on somebody's earlier solutions (mine, among them) or MT solutions that are incorporated in the tool.

I see you're using Wordfast Pro, and you work in subject areas such as poetry & lit., culinary (i.e. probably menus), tourism, etc. So, what are you expecting of Wordfast in these areas? Are you offering your clients discounts based on matches in the TMs they send you or, even worse, on MT propositions? If so, you're bound to be upset.

What have the vendors of Wordfast promised? I'm sure they've promised nothing other than (a) reminding you of the words you've used before in similar phrases, and (b) QC checks to make sure you leave nothing untranslated etc.

Discounts are (for example) applicable for technical manuals, where:
- there's a whole raft of instructions and warnings that are repeated on every single EU instruction leaflet (which nobody ever reads); or
- a manual for version 125.02 of some product may be ever so similar to the one for version 125.01 - again, apply a discount of less than 100% and you're laughing.

But for creative translation of the type you do, there can rarely be any justification for the grid discounts that you often see requested. I work in marketing and tourism and I never apply them, although I get great benefits from Wordfast and it does sometimes enable me to give a discount (of my choosing) on a job so that the client and I both benefit from the savings. Mainly though, I use a CAT tool for the concordance searches, the glossaries, the QC functions, etc. Fuzzy matches (and to some extent even repetitions and 100% matches) really aren't significant in the more creative areas of translation.

Don't let yourself be bullied into accepting discounts simply because some clients demand them. Or, if you do, please don't blame the CAT tool.


Direct link Reply with quote
 

Peter Simon  Identity Verified
Netherlands
Local time: 20:56
Member (2013)
English to Hungarian
+ ...
TOPIC STARTER
Thanks ... Mar 14, 2014

Sheila, for these ideas. You're right, except or WordFast ... I don't want to use this forum to bash any software, but I use a more expensive one, and, plainly, don't get what is promised. I hope that, over time, I get more out of it when I manage to put together my autosuggest dictionary and to create term-bases. Perhaps I wanted returns too fast.

Direct link Reply with quote
 

Sheila Wilson  Identity Verified
Spain
Local time: 19:56
Member (2007)
English
+ ...
Wordfast, Trados... Mar 14, 2014

Sorry Peter - I saw you'd attended a training course in Wordfast, but I see you use Trados, too. I don't use Trados, and I certainly don't want to discuss the pros and cons of various CAT tools either.

I just wonder who promised you what from your CAT tool investment. I suspect that you are indeed expecting too much. You may not save very much in the early days in terms of time, so you can't afford to give high-percentage discounts, but you will be increasing your consistency of terminology (where that's important), and you'll be able to use fuzzy matches and/or concordance searches as brainstorming fodder: I often find myself stuck for the perfect adjective (for example), but my CAT tool comes up with just the right one. It isn't magic - it's just reminding me of the flash of inspiration I had a week/month/year before.


Direct link Reply with quote
 

Erik Freitag  Identity Verified
Germany
Local time: 20:56
Member (2006)
Dutch to German
+ ...
Difficult to answer Mar 14, 2014

Peter Simon wrote:

But my original Q seems to have been overlooked: if I use the tool, which is capable of rendering % to my solutions, what is my loss if the tool (incorrectly, in lot of practical cases) declares that there's a high % of correlation with the one suggested by the tool, based on somebody's earlier solutions (mine, among them) or MT solutions that are incorporated in the tool.


I see what you mean. Your question ("what is my loss") is difficult to answer. It depends on too many factors.

Sheila has given good advice, and I think she's right that you probably have unrealistic expectations. As a general hint: I'm using a CAT tool for all my work, and I find it very helpful. Still, I don't offer any discounts for fuzzy matches. Indeed, the CAT tool we both seem to use calculates match values that imho bear next to no significance (except for very high values).


Direct link Reply with quote
 
Maxi Schwarz
Local time: 13:56
German to English
+ ...
In regards to percentage of deduction Mar 15, 2014

Whether to give discounts to a client, or what to charge for a translation period, is the decision of the translator. A while back some company decided to create software and advertised it to the customers of the professionals who would be using that software, promising those customers that the professionals would give deductions. It shows some ignorance about the nature of translation. There is no reason why translators should honour a policy that a software developer coined on their behalf.

Have you actually determined that you need a CAT tool? Are you translating high volume technical material requiring terms to be translated the same way each time?


Direct link Reply with quote
 

Tomás Cano Binder, BA, CT  Identity Verified
Spain
Local time: 20:56
Member (2005)
English to Spanish
+ ...
It is important to discuss CAT and MT Mar 15, 2014

I think that the concerns raised by other colleagues make total sense. Without a clear picture of CAT and MT (even if some tools combine them), we will be talking different languages and any opinions won't have a base. I am among the ones who think you do not have a clear picture of what a CAT tool is about.

A vast majority of professional translators today use some kind of CAT tool, but also a vast majority of them dislike MT and will not use it.


Direct link Reply with quote
 

Christine Andersen  Identity Verified
Denmark
Local time: 20:56
Member (2003)
Danish to English
+ ...
Last week my CAT helped me through an 'impossible' job. Mar 15, 2014

I don't use MT, but I am always in charge when I use my CAT. There is only human translation in the TM, some of it from colleagues as well as my own. (I can see their initials, and know when to avoid some of the Canadian's suggestions for UK law for instance, but I know he is very reliable for technology.)

Last week one of my clients sent me a 12 800 word contract in Norwegian with law and technology, and a TM.
It translated into 17 500 English words, which is quite normal for that language pair, and she needed it a week later.

That is not my favourite language pair, so it was a lot more than I would have attempted in the time without the advanced CAT features. Although there were practically no matches and very few repeats, I could use the concordance (in the TM), where I could see suggestions from trusted colleagues as well as my own former work. Multi-Term and the AutoSuggest feature were particularly useful, saving me typing the long names of parties to the contract dozens of times, and once I had entered the technical terminology in the database, it kept track for me of the small but critical differences in the terminology, saving me checking every time. It saved time on proofreading too.

I have to admit that I keep coming back to Trados Studio after trying other CATs, simply because I prefer Multiterm for my purposes.
There are few useful repeats in most of my work, but there are references to earlier work for the same client, and the concordance is absolutely the quickest way of finding the terms in context. Otherwise I enter fixed terminology in the database and it can be inserted, correctly spelled, at the touch of a couple of keys.

You can usually set the percentage for fuzzy matches on most if not all CATs, and perhaps yours is set too low for the work you do.

I don't know about Hungarian at all, but MemoQ was created by Hungarians, so I would guess it is not entirely hopeless for Hungarian, especially as it is so popular generally.

It does take time to learn about all the features of an advanced, expensive CAT, but the producers have been forced to listen to translators. Believe me, many translators would not spend their hard-earned money on a CAT if they did not think it was worth it, although different people prefer different models.

Don't write them all off, but get to know what you need for the work you do, and how to adjust the settings to suit your way of working.

Maybe you should look at a different CAT, taking advantage of the free demos and really trying them out. The expensive ones can be customised, so it is worth persevering and finding something you like.

Disconnecting machine translation while you do it might help you to sort out the best settings. Please excuse the pun, but MT may be clouding the issue by introducing attempts by non-professionals and crowd-sourcers who contribute to the databases. It depends a lot which databases you are linked up to, of course. MT also has problems with homonyms, at least in Danish, that humans would simply never confuse. Composite verbs cause headaches and false matches too...

I have, as I said, negligible experience of machine translation, but I have seen some of the results, and can imagine some of the weird and unusable 'matches' come from there.

Keep trying with your CAT, because I think the St. Jerome style of translation without technical help is a rapidly dying art! I can still do it, when I get medical records as scanned PDF images, but I seriously miss my CAT.

Best of luck!


[Edited at 2014-03-15 21:40 GMT]


Direct link Reply with quote
 

Peter Simon  Identity Verified
Netherlands
Local time: 20:56
Member (2013)
English to Hungarian
+ ...
TOPIC STARTER
Some clarification Mar 16, 2014

Dear colleagues,

Thank you very-very much for your answers. Under your keen eyes, I'm now able to make a few clarifications about my problems.

As I've pointed out, I have been using CAT tools only for a short time because earlier I lived in Hungary and mainly worked as a teacher, where I couldn't afford to spend a full month's salary on any of those tools, let alone more than that. And as a result, I'm finding out, to my consternation, that what I now use is of very little value simply because I haven't been using it during the decades of my intermittent translator's life.

As the cat is already out of the box, I can admit that I first downloaded the free version of WordFast and did a course on it (that's what Sheila noticed), thinking that what I've learnt there would be useful for the other one that I bought with the group discount on Trados 2014. I bought the latter because by far the most ads on various translators' sites require the applicants to use it (so no, it didn't seem that I can be free to use a CAT, and no, I may not have enough means to try and buy MemoQ as well and begin it all from scratch there).

Because I've learned a few things on the WordFast webinar, I thought I can easily make my term-bases from what sits in my excel-file I've been using for learning some new languages. To my surprise, I had to learn that Trados doesn't give this opportunity as it comes, I'd have to buy other programs consistent with it ... and I said no, I have no more money to fork out on this. Then I learned that there's one free program that does it, and the result can be used for Trados. OK. Then I work through my file and use it.

I also assumed form Trados ads that the program automatically suggests words as I'm going ahead with my translations. No, it doesn't, not without the TermBase. Only, Trados ads fail to mention this 'small' glitch while polishing the image of their product to lure new users.

So now I can only rely on the BeGlobal Community to help me. But that's actually an MT. I'm weary of add GoogleTranslate to the TM basis as I'd have to pay for it and I'm more-or-less sure that I wouldn't get much better results for my money than for the free version on the Net. So that's all I can rely on, that's where my fuzziness came into my posts regarding MT and CAT. I've been building my Translation Memory only for a couple of months, which hasn't allowed me much opportunity to really use the Automated Translation feature, which already shows me fuzzy and more exact matches. So I have very limited use of the concordance opportunity Christine mentions and I can't even see where Multiterm could be that she mentions.

One strange thing with BeGlobal is that when I do Dutch to Hungarian, it refuses to work - it says it doesn't exist in the pair I'm trying to apply. Does anyone know what's wrong there? Dutch is not a small language and if it works between E and Hun, it should work from Dutch to Hun as well, I innocently suppose as Google and a few others do work, for better or worse. Perhaps I'm wrong.

The translation jobs I've been doing regularly were initially done before I started with Trados, but with my new additions, I already see a couple of benefits. But because most of the items on the Memory are sentences, those never repeat themselves. Only some names and the dates sometimes return on their own, that's when the CAT works for me as a CAT. But it doesn't purr as nicely as I'd hoped.

Actually, I've never done a job with 'fuzzy matches' as part of a contract, but recently I've been offered one involving checking machine-translated stuff where the contract mentioned deductions for more-or-less full and less full matches. I don't know how that would be calculated but because it would affect what I'd get, I got a bit nervous and tried to ask about how it may be done in reality. This is not my ignorance, but comes as a result of the novelty of the idea, exacerbated with the experience with Google and similar on-line MT tools, which I've quickly grown to despise for the hillarious results I get into Hungarian, also with BeGlobal. That's where the examples in my blog post come from.

Perhaps it is now clearer why I've washed together MT and CAT tools considerably - in my limited experience, the CAT tool I work with hasn't been much more than an MT. So I see that I have to apply the advice Sheila and Christine have given me and be very patient in building all up from the limited basis I have. I hope that when I finish my excel file and manage to turn it into term-bases based on word lists rather than sentences in my TM, they become a lot more beneficial.

Thank you all for your help. Cheers,

Peter


Direct link Reply with quote
 

LegalTransform  Identity Verified
United States
Local time: 14:56
Member (2002)
Spanish to English
+ ...
The problem is... Mar 16, 2014

...that translators established our base "per word" rate based on the fact that on average over the course of a year's worth of documents, there would be some repetitions, some fuzzy matches, some "easy" words. Now clients want to remove all of this and think that translators should then still charge the same base "per word" rate.

No, no and no! If clients are going to unilaterally change the parameters of which units are charged for and which are not, then translators must also change the parameters of how much is charged for what is left.

In the end, the price is still the same whether translators charge .xx WITHOUT CAT deductions or .xx + .xx WITH deductions, there is no difference.

Edited: per Peter's suggestions.

[Edited at 2014-03-16 17:22 GMT]


Direct link Reply with quote
 

Peter Simon  Identity Verified
Netherlands
Local time: 20:56
Member (2013)
English to Hungarian
+ ...
TOPIC STARTER
Sorry, Jeff, Mar 16, 2014

I don't understand what you're talking about, viz. what you mean by 'we', 'they', and 'you'.

Thanks, Jeff, it does make sense this way. Good points.

[Edited at 2014-03-16 17:29 GMT]


Direct link Reply with quote
 
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 »

Percentage deducted for Matching

Advanced search







memoQ translator pro
Kilgray's memoQ is the world's fastest developing integrated localization & translation environment rendering you more productive and efficient.

With our advanced file filters, unlimited language and advanced file support, memoQ translator pro has been designed for translators and reviewers who work on their own, with other translators or in team-based translation projects.

More info »
Across v6.3
Translation Toolkit and Sales Potential under One Roof

Apart from features that enable you to translate more efficiently, the new Across Translator Edition v6.3 comprises your crossMarket membership. The new online network for Across users assists you in exploring new sales potential and generating revenue.

More info »



Forums
  • All of ProZ.com
  • Term search
  • Jobs
  • Forums
  • Multiple search