Poll: Would you pay to use machine translation software if it were customized for your areas of expertise? Лице кое објавува дискусија: ProZ.com Staff
|
|---|
This forum topic is for the discussion of the poll question "Would you pay to use machine translation software if it were customized for your areas of expertise?".
View the poll results »
| | | | neilmac
Шпанија
Local time: 07:03
шпански на англиски
+ ...
This question seems rather loaded. Who wants to know?
Anyway, I already "pay to use machine translation software" (NB: use, not abuse). I was using an early version of Systran around the turn of the century, but it wouldn't work on more recent operating systems. When I eventually forked out the cash for a more recent version, I was disappointed to find that the simple copy and paste method of the old one had been "upgraded" into a fiddly, irritating nightmare to use. After a couple ... See more This question seems rather loaded. Who wants to know?
Anyway, I already "pay to use machine translation software" (NB: use, not abuse). I was using an early version of Systran around the turn of the century, but it wouldn't work on more recent operating systems. When I eventually forked out the cash for a more recent version, I was disappointed to find that the simple copy and paste method of the old one had been "upgraded" into a fiddly, irritating nightmare to use. After a couple of days persevering without seeing any benefit or convenience, I stopped using it. In fact, the other day I removed it from this PC completely, because it was causing a conflict with Wordfast.
A couple of years ago I started using GT4T as a helper tool and I still use it, albeit judiciously, and only for brief text fragments. I find it can be interesting to compare its results with my own efforts, but certainly wouldn't recommend it as a panacea.
PS: I don't know about the "customized for your areas of expertise" part... it's not an issue for me.
[Edited at 2014-04-11 08:43 GMT] ▲ Collapse
| | | |
I carefully try to specialise in areas were MT is simply not suitable.
It is part of a deliberate strategy, not to compete with MT, because that is hopeless, but to find the jobs where MT cannot compete with human translation.
I dislike post-editing and am getting very fussy about proofreading - if the text is not already almost faultless.
As in many other professions, there are still enormous areas where machines are not going to take over in the foreseeab... See more I carefully try to specialise in areas were MT is simply not suitable.
It is part of a deliberate strategy, not to compete with MT, because that is hopeless, but to find the jobs where MT cannot compete with human translation.
I dislike post-editing and am getting very fussy about proofreading - if the text is not already almost faultless.
As in many other professions, there are still enormous areas where machines are not going to take over in the foreseeable future, and that is where translators need to set their sights. ▲ Collapse
| | | |
because technically, only rubbish is output and the effort to complete the review is more than translating in the first place.
Which company has paid to have this poll placed??
| | |
|
|
|
Elina Sellgren 
Финска
Local time: 08:03
Член (2013)
англиски на фински
+ ...
I am suspicious but if someone one day shows me software that would actually be useful like this, I might try it.
I first misread this as talking about CAT tools and selected 'yes'...:D
| | | | Thayenga
Германија
Local time: 07:03
Член (2009)
англиски на германски
+ ...
Michael Harris wrote:
because technically, only rubbish is output and the effort to complete the review is more than translating in the first place.
Which company has paid to have this poll placed??
My thoughts exactly.
Christine Andersen wrote:
I carefully try to specialise in areas were MT is simply not suitable.
So do I. Who has heard of an MT software that can produce even halfway decent results in literature or poetry? A CAT tool comes in handy regarding technical translations, something I cannot honestly say about MT.
| | | |
My brain is a “trillion times” more sensitive to context, which is everything for meaning, than any machine translation software (customized or not!). Anyway, my brain needs constant intellectual stimulation…
| | | | | mebbe, mebbe,.... | Apr 11, 2014 |
Although I answered "yes" just to screw with the nameless, presumably corporate, body who placed (paid for?) the poll. Because the name of the person who suggested it is *usually* displayed, isn't it?
But akshully, I don't object to that. Would be nice to know who they are, at least afterwards. But here is surely an excellent place to conduct market research among translators. And we know the site is only in it for the money, and all that "community" BS went out the window l-o-n-g a... See more Although I answered "yes" just to screw with the nameless, presumably corporate, body who placed (paid for?) the poll. Because the name of the person who suggested it is *usually* displayed, isn't it?
But akshully, I don't object to that. Would be nice to know who they are, at least afterwards. But here is surely an excellent place to conduct market research among translators. And we know the site is only in it for the money, and all that "community" BS went out the window l-o-n-g ago.
But the "maybe" bit is true. Although it would have to be entirely disconnected from the web, which is my main objection to MT at the moment. And therefore it would probably be hopeless, since crunching statistical probabilities on huge volumes, too big for my laptop to hold or process, are how I believe it works now. But if someone invented a half-decent, presumably rules-based, standalone wossname, I might be interested.... ▲ Collapse
| | |
|
|
|
Anthony Baldwin
Соединети Американски Држави
Local time: 01:03
португалски на англиски
+ ...
| I do, sort of. | Apr 11, 2014 |
I use a CAT tool that has the ability to incorporate results from Google Translate (http://www.omegat.org ), over their API, for access to which one must pay, and I do pay. I can insert those results in a target segment with a simple keybinding (Ctrl-M for Machine).
I probably spend about $5 - $10/month on that.
I would never in a million years send a client something translated with Google Translate, a... See more I use a CAT tool that has the ability to incorporate results from Google Translate (http://www.omegat.org ), over their API, for access to which one must pay, and I do pay. I can insert those results in a target segment with a simple keybinding (Ctrl-M for Machine).
I probably spend about $5 - $10/month on that.
I would never in a million years send a client something translated with Google Translate, and am often so amused by its results that I split my sides with laughter.
At the same time, it still proves immensely useful for me.
I work largely with complex legal and technical materials, and despite what others have said, no, MT is not great at technical translations, and for legal matters, well, that's where I get the most laughs.
Nonetheless, when I have 1200 segments where I simply need to convert numbers from R$ 98.345,78 to R$98,345.78, being able to plug in the Google MT results with a quick bang of a keybinding, rather than typing can save me hours upon hours of work.
And there are often short, simple segments where the output is appropriate, and, likewise, being able to simply insert it with the keybinding is very efficient.
And even the segments where the MT output is laughable, it IS sometimes even more efficient to insert it and properly arrange the grammar or otherwise repair it than to completely type in the proper translation. Note, I say "sometimes", because other times the repair is so extensive, as others have indicated, that using the MT output would only take more time. Nonetheless, there ARE such instances where MT saves me time.
I dare say, on some projects it likely cuts my time by up to as much as 30%, and that is valuable, and certainly worth the $5 to $10 it costs me. ▲ Collapse
| | | | Triston Goodwin
Соединети Американски Држави
Local time: 23:03
шпански на англиски
+ ...
I don't even like paying for CAT tools.
| | | | Al Zaid
Соединети Американски Држави
англиски на шпански
+ ...
| Wouldn't touch MT... | Apr 11, 2014 |
...with a ten-foot pole...
MT works for people who need a text translated -maybe for understanding- and are willing to take the risk.
But for us who are responsible for the output accuracy and quality, it is really cumbersome to review MT to see whether it's been done properly or not. I thoroughly review all my translations or have them reviewed by the occasional colleague (I pay them back by reviewing theirs), but that's a totally different matter...
And there's what Ch... See more ...with a ten-foot pole...
MT works for people who need a text translated -maybe for understanding- and are willing to take the risk.
But for us who are responsible for the output accuracy and quality, it is really cumbersome to review MT to see whether it's been done properly or not. I thoroughly review all my translations or have them reviewed by the occasional colleague (I pay them back by reviewing theirs), but that's a totally different matter...
And there's what Charlie Bavington says... the most decent ones are actually on-line...sucking feedback and information from what you translate and revise.
So, for now, I'm in Rage Against the Machine.
[Edited at 2014-04-11 17:51 GMT] ▲ Collapse
| | | | | Wrong question | Apr 12, 2014 |
This is a typical MT provider question:
"Would you pay to use machine translation software if it were customized for your areas of expertise?".
I actually don't care if the software is customized or not, all I am interested is "output that is good enough to be useful" for my purpose. Unfortunately I have not seen any output from any statistical MT system that fullfilled this requirement "good enough to be usefull".
The longer and the more I get into statistical M... See more This is a typical MT provider question:
"Would you pay to use machine translation software if it were customized for your areas of expertise?".
I actually don't care if the software is customized or not, all I am interested is "output that is good enough to be useful" for my purpose. Unfortunately I have not seen any output from any statistical MT system that fullfilled this requirement "good enough to be usefull".
The longer and the more I get into statistical MT, the clearer the picture. Statistical MT is many years away from being useful for me. And customized - there are no working customized MT engines in medical/pharma in my language pairs.
Would I be prepared to pay for using a MT software that produced output that in my eyes fulfills the requirement "good enough to be useful"? Sure, I would be prepared to pay for it.
And I am actually looking for a system - Please, any MT vendors that believe that their system works (and I mean works - not can be customized to fulfill my requirements) for my/our language pair(s) English > German, Dutch > German, English > French, German > French, domains: medical technology, medical IT, pharma (SPC/PIL), ICF conatct me.
I am prepared to spend serious money on a "working" system.
Before contacting me, please remember that I am not a risk capital outfit that can be tricked with some marketing buzz, I want to see results. ▲ Collapse
| | |
|
|
|
Muriel Vasconcellos (X)
Соединети Американски Држави
Local time: 22:03
шпански на англиски
+ ...
As I have said here many times, I have been using machine translation for over 30 years on a regular basis. I might consider paying a small fee, but only on the condition that I could update the dictionaries myself.
There are times when MT speeds up my work, but I am also quite fast when I'm translating from scratch, and it's less stressful for me not to have to be constantly manipulating text on the screen. When I translate naturally, I close my eyes and do the sentence in my head... See more As I have said here many times, I have been using machine translation for over 30 years on a regular basis. I might consider paying a small fee, but only on the condition that I could update the dictionaries myself.
There are times when MT speeds up my work, but I am also quite fast when I'm translating from scratch, and it's less stressful for me not to have to be constantly manipulating text on the screen. When I translate naturally, I close my eyes and do the sentence in my head while touch-typing the my solution. It's more relaxing for me, especially as I'm going blind in one eye.
[Edited at 2014-04-12 07:21 GMT] ▲ Collapse
| | | | | for eye protection | Apr 12, 2014 |
Muriel Vasconcellos wrote:
As I have said here many times, I have been using machine translation for over 30 years on a regular basis. I might consider paying a small fee, but only on the condition that I could update the dictionaries myself.
There are times when MT speeds up my work, but I am also quite fast when I'm translating from scratch, and it's less stressful for me not to have to be constantly manipulating text on the screen. When I translate naturally, I close my eyes and do the sentence in my head while touch-typing the my solution. It's more relaxing for me, especially as I'm going blind in one eye.
[Edited at 2014-04-12 07:21 GMT]
I type too slow, so it MIGHT BE useful to have a "raw" version to save the time and protect my eyes.
| | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Poll: Would you pay to use machine translation software if it were customized for your areas of expertise? | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
| | Pastey | Your smart companion app
Pastey is an innovative desktop application that bridges the gap between human expertise and artificial intelligence. With intuitive keyboard shortcuts, Pastey transforms your source text into AI-powered draft translations.
Find out more » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
