| Pages in topic: < [1 2 3 4 5 6 7 8 9 10] > | Clients / large translation companies now talking about pooling linguistic data. Should we be there? Thread poster: Henry Dotterer
|
|---|
| I get enough of that during regular work | Jul 24, 2008 |
Henry D wrote:
What if, with ProZ.com participating in this group, it were possible to make the pooled resource available to ProZ.com members? In that case -- with translators getting back access to the linguistic assets they create -- would some balance be restored?
Honestly Henry, I regret it enough when I accept to use 100% matches as is from a translation memory supplied by a customer. A scenario with pooled translation memories with 147 different translators and 87 different levels of quality makes me shiver! 
| | | |
the quality would be so variable and the application of outsourced TMs for single clients so limited in scope that I don't see that happening. Nice idea, but I don't think it would work overall. Maybe from a terminology point of view, although, again, the quality would have to be checked and that's expensive...
| | | | | Thanks, RobinB | Jul 25, 2008 |
Of course that doesn't mean that a limited amount of sharing doesn't go on, for example in the form of seminar and workshop content and handouts (which of course come at a modest price). And we share IP with the handful of freelances we work with because there's a mutual benefit.
The term IP, as you use it here, are you talking about linguistic assets, or something more? (Or is that proprietary?)
It would seem to me that if TMs, etc. are differentiators, those who share, thereby enriching their resources, would get ahead. No? I don't see it happening this way. The owners would be giving away quite a lot of what makes them successful in the market, and what would they get in return? More of what they already have? The reusability of much of this IP is heavily limited in any case. For example, we estimate that less than 5% of financial reporting TM content for a customer is reusable across other customers.
Interesting to know.
There's certainly a place for initiatives such as the TDA or tmmarketplace.com, but I see these more in the commodity localization space than in high-end narrative translation.
Makes sense.
If you believe that ProZ should somehow get involved in translation data exchange, I personally think that you could consider helping to promote interchange format standards such as TMX, TBX and XLIFF (and especially the latter) and encourage their take-up by tools manufacturers so that users (above all translators) can concentrate on buying or leasing (SaaS) tools with features that best meet their own requirements, rather than being forced into using - generally expensive - captive systems that use proprietary data formats. A more open market for translation tools would, I think, represent a significant step forward in user choice and overall tool adoption, and contribute significantly to growth in tool market maturity.
I agree.
Andre can tell you what LSPs at the Englewood conference think about the current tools market situation.
I'll ask. Thanks!
| | | | mediamatrix (X)
Local time: 13:32
Spanish to English
+ ...
| Well, since you ask... | Jul 25, 2008 |
Henry D wrote:
Actually, for once I don't disagree with you on a topic (not yet, anyway... knowing you, you'll have more to say!)
It would be a pity to disappoint you, Henry, by not saying something more on this topic - after all, despite the references to 'suspects' and past quarrels, we have actually agreed on several things in the recent past, especially in the area of project management.
The likely impact of TAUS on the livelihoods of 'real' translators has tended to dominate this thread, so I'll try to get things back 'on topic' by again answering Henry's second question. And as a bonus, I'll answer the first question as well.
Henry asked:
1. Does this topic matter to your business? How?
No, not in the slightest. As far as my career in translation is concerned - and even more so with regard to my current business - TAUS and the whole CAT kaboodle is coming (at least) thirty years too late. And when I quit, my business will die with me. But if these things had come along thirty years ago, I would have been pushing my (then) employer to get involved (sigh...).
Henry then asked (edited version...):
2. Should ProZ.com join as a company, maybe with a designated attendee or two (translator and small company) from among our members?
Short answer: No, I don't think you should.
Why? Well, since you asked, here comes the long answer:
Let me first explain where I'm coming from in this matter. In a 'past life' - and for nigh on 25 years - I was employed by what is in effect an 'industry association'; an extremely successful one in fact whose brand has been a household name across Europe for over 50 years and which has had significant impact on the television viewing and radio listening habits, as well as the domestic media technology, of almost the entire world population. That 'association' is, itself, an influential founder member of several other industry associations.
From what I have observed, there are two categories of involvement in these associations. At one end of the table there are the 'founder members', often large influential bodies in their own right, who for reasons concerned mainly with their own financial and operational security decide to harness the assets/ideas of others less influential or, if the 'others' have nothing worth taking then at least 'fence them off' and make it difficult for them to carry on their business. Key tools to this end are royalties/licence fees or what-have-you on the new-fangled 'standards' they intend to 'encourage'. At the other end of the table, of course, there are those 'other' members who join in through fear of what might be going behind their backs and decide to 'send someone along to see what's happening'.
The 'big boys' know what they want - and how to get it. They have funds (a life-saving substitute for real expertise) sufficient to enable them to dedicate a small team within their organization to making sure the association achieves the 'appropriate' end result. Three or four individuals, working full-time of course, will often be sufficient. The 'appropriate end result' often translates (sic) as 'imposing their way of doing things, at minimum long-term cost to themselves and regardless of the long-term investment needed by the 'small fry' when they (poor blighters...) eventually find themselves marginalised technologically and financially'. As a gesture of 'bonne volontée' towards the free market, they will sometimes (despite the annoyance caused to some of their erstwhile founder-member peers) deliver 'open standards' covering all but a few apparently insignificant details; they will leave the standard just vague enough to ensure that with mimimum extra investment (and with the certain, prior, knowledge of where those vagueness-filled loopholes will miraculously appear, as if from nowhere) they can impose a proprietary 'interpretation' of the standard that effectively leaves them with the technological (and economic) monopoly they set out to achieve on day one.
Just occasionally - and I've seen it happen - the big boys get a surprise. They find themselves face-to-face with a small-fry 'peer' who gives them more than they bargained for. If the small-fry's offerings are genuinely useful they will generously adopt them with a smile, assuming for themselves the credit for having been 'open to input from all comers'; of course, these features will become part of the open standard thus minimising the return on the small-fry company's investment. To minimise the risk of problems they may also buy the small-fry company's expert (if not the entire company).
Small-fry representatives who have nothing to offer will simply be ignored. Except when it's their turn to host a meeting, of course; they will be expected to be as generous as the big boys with lavish entertainment, airport limos, the best hotels, excursions for the delegates' partners, etc. etc.
Anyway you get the picture, I guess. Now, let's imagine Proz.com in this scenario. Big-boy? - or small-fry? Come come! Let's not be modest: 'Big-Boy!
Well now, let's see how this works in practice. We've already established in this thread that the Proz.com company can only represent itself. Oh dear! most of the staff here are computer boffins, administrators; oh - and a PR man; mustn't forget him! It stands to reason - the CEO is on record as saying that his primary business is web-hosting so it's only natural that hands-on in-house high-tech translation awareness is in rather short supply.
No matter - perhaps we can make up for that with dollars! Ooops! - the CFO has already divulged the budget ...
Henry D wrote:
Several thousand euros, plus some travel and possibly other costs for whoever goes, ...)
Aie!!!! That won't go far in the hotel bar after a tough meeting of the representatives of 40+ of the world's biggest $$$$$-spinners! These fancy hotels charge you more than that just for the ice-cubes, let alone the whisky! We'd better make sure we leave before it's our round!
Hmmmm... Errrr ... maybe it would be sufficient - and still worthwhile? - if we sent someone along just to 'see what's going on' in TAUS. And who knows? - later, if we can impress those fellows with our products they might just listen to us. Now let's see, what do we have to offer - stuff really relevant to their project? Well, we've got the KOG! ... No .... perhaps we oughtn't to push that - even our own customers tell us its full of ?!*?#!@!# No, we'd do better to tell them about wikiwords. Hmmm ... perhaps not ...
Well even if we have nothing to offer we can still spend our budget making sure there's someone there just to listen to these guys extolling the virtues of their pre-conceived idea of translation paradise! That's it! We'll just sit in the corner, taking note. Ah! - but actually they're already telling us quite a lot on the Web, so why waste time and money that could be better spent on stuff that's really important to our own customers? Like fixing the KOG .... Mumble mumble mumble... dammit! should've got that sorted out years ago. If we'd done that we'd have something to bargain with in TAUS - we'd be able to play conkers with the big boys!
Oh well. Maybe we'd do better to concentrate on crowdsourcing...
MediaMatrix
| | |
|
|
|
| Thank you, mediamatrix! | Jul 25, 2008 |
Wow, you did not disappoint at all! I'm serious and I mean that in a good way.
Of course a handful of your speculations aren't accurate (it's not our style to sit in corners, for example), but a lot of your post was really insightful, not to mention hilarious. I had a pretty good laugh. Thanks for taking the time.
Since we aren't actually resorting to "crowdsourcing", at least not in the way that you may mean it, what would you advocate? How can we best serve the intere... See more Wow, you did not disappoint at all! I'm serious and I mean that in a good way.
Of course a handful of your speculations aren't accurate (it's not our style to sit in corners, for example), but a lot of your post was really insightful, not to mention hilarious. I had a pretty good laugh. Thanks for taking the time.
Since we aren't actually resorting to "crowdsourcing", at least not in the way that you may mean it, what would you advocate? How can we best serve the interests of ProZ.com members (ie. our customers; this is not feigned altruism), including those with their careers ahead of them?
Apart from working on the KOG / wikiwords, which we are doing again now... finally... ▲ Collapse
| | | | RobinB 
United States
Local time: 12:32
German to English
Henry D wrote: The term IP, as you use it here, are you talking about linguistic assets, or something more? (Or is that proprietary?)
Sorry, I'm talking about the intellectual property embedded in linguistic assets, specifically translation memories (TMs), as well as things like bilingual corpora, other bitexts, termbases and translator training materials.
Robin
| | | |
Henry D wrote: Giovanni Guarnieri MITI, MIL wrote:
big companies pooling their resources...
to drive rates even further down and make big fat profits for themselves...
they get our TMs... What if, with ProZ.com participating in this group, it were possible to make the pooled resource available to ProZ.com members? In that case -- with translators getting back access to the linguistic assets they create -- would some balance be restored? I don't know whether this would be possible in TAUS. I asked, but if I understood correctly, a policy has not yet been formed. The impression I got is that this concept had not yet been considered.
that would be helpful. I think you should join and see how it goes. It's better to be there to influence people and processes from the inside. If you then find the whole thing a waste of time, you can always withdraw.
The portal idea is interesting, although it would present some problems, especially from the IP point of view, as Robin pointed out in his post. Another problem would be related to the privacy. It will be a heck of a job removing all the references to clients names and products. Who's going to do that? I don't think big companies would be very happy to see their confidential information published on a public website. The objective is very clear: a classic case of 'let's give a little away (public portal) but let's keep all of the important data to ourselves', so we can gain an even more dominant position in the marketplace. Let's face it: these people are not charities, they are in for the big bucks.
“This collaboration is a major breakthrough for the translation industry”... come on, we are not stupid...
[Edited at 2008-07-25 08:59]
| | | | Gerard de Noord
France
Local time: 19:32
Member (2003)
English to Dutch
+ ...
| I like the tone of this discussion | Jul 25, 2008 |
Ralf Lemster wrote: 2. Should we join, maybe with a designated representative or two (translator and small company) from the community? Given the expected uproar amongst translators (there won't be a rational discussion, methinks), I would keep this low-profile, but it would be good to keep an eye on developments. Best, Ralf
As for the discussion in this forum, I'm glad Ralf has been wrong until now. The actual uproar might start in a year or two, when we'll start receiving pre-translated documents entirely consisting of 95% matches.
We, the translators, are the only party who seem to know the actual value of (our own) TUs. Other parties have different interests. I've been translating with Wordfast since 2001 and have created and gathered an impressive corpus of TMs. The reusability of all that material is limited, TUs get stale so fast. So many of the promising green, yellow or golden matches my clients send me need editing.
I'm not a Luddite, my background is IT, and I'll probably remain a translator in the next two decades or so. The industry will keep changing and, after some consideration and reading all the contributions to this thread, I think Proz.com (Henry D himself) should try to get on board of this initiative, were it only for the good reason Robin has mentioned: promoting interchange format standards. Another requirement for the data pool should be that every TU has loads of attributes to filter them, otherwise the whole project would be nothing but detrimental for us translators.
Regards,
Gerard
| | |
|
|
|
| @ Gerard and this thread | Jul 25, 2008 |
Gerard de Noord wrote:
The actual uproar might start in a year or two, when we'll start receiving pre-translated documents entirely consisting of 95% matches.
Gerard, in my experience (software/technical manual translation, especially in the case of some specific customers' projects), pre-translated documents with 95% matches (even if you yourself have produced a lot of them) are already happening and have been happening for a while, so maybe this is the reason why we sometimes already witness so much "uproar" in the forums, because such industry shifts have already been taking place and have already been affecting some people's daily work, work options or work situation in general.
But I totally agree with you that this has proven a very useful thread, in spite of Henry D's and Ralf's initial "fears" .
So many crucial aspects of the current trends in our industry have been mentioned, and in such an insightful way, I personally consider it a very useful reference of a lot of what lies ahead in the world of translation, especially technical translation. Definitely food for thought.
Cheers,
Ivette
| | | | | What about the confidentiality issue? | Jul 25, 2008 |
I don't really understand. All TMs are so full of absolutely confidential material. How can such resources safely be pooled?
As for the automation of translation, that is exactly what I am trying to do personally at this very moment. I have at last worked out that it is only possible in very particular circumstances to earn a professional living as a translator. The translation process is inherently very difficult. It is not possible, even at rates such as EUR 0.10, EUR 0.11 or EUR ... See more I don't really understand. All TMs are so full of absolutely confidential material. How can such resources safely be pooled?
As for the automation of translation, that is exactly what I am trying to do personally at this very moment. I have at last worked out that it is only possible in very particular circumstances to earn a professional living as a translator. The translation process is inherently very difficult. It is not possible, even at rates such as EUR 0.10, EUR 0.11 or EUR 0.12 per word, to earn any more than a cleaning lady if we do translations "by hand". I have therefore been strenuously trying to optimise the process with my CAT tools. I have come to the conclusion, in this respect, that creating very large Multiterm termbases, which are not confidential material, is the single most effective way to speed up the translation process. According to one or two previous polls, most users of Multiterm do not use it to the extent that it could seriously speed up the translation process for them. Actually building up extremely large termbases, optimised in the right way, with the right terms and expressions, and sorted out very neatly into topics, and then adjusting the percentage recognition match (in the case of inflecting languages) is the first way to earn at least twice as much per hour, or possibly three times as much. However, the termbases need to be as large as possible, and therefore represent a colossal amount of work, if created manually (as is ultimately the most useful way). I would be VERY interested in a project which would pool termbases. It is possible - and a good idea - to feed in whole sentences and expressions which occur regularly, but at least they are then not confidential sentences, relating to a particular case. I see the pooling of TMs as much more problematic, not only from the confidentiality point of view, but in view of many other aspects.
One thing is clear to me: we need to be able to translate pretty fast these days, otherwise the cleaning lady, the plumber, the electrician, etc. will always earn more than us. Therefore, I am all in favour of "automation" of translation. It is absolutely necessary.
Astrid ▲ Collapse
| | | | Catarina Aleixo
Portugal
Local time: 18:32
Member (2007)
Portuguese to English
| Unless the nature of language and culture changes... TMs are finite in their use | Jul 25, 2008 |
It seems to me that at the heart of this dicusssion is whether we will ever be replaced as translators. My belief is that we won't. the nature of our work may change, just as it already has with the introduction of translation software in the first place.
Luckily enough for us every world language, as far as I can tell, is infinite in its ability to express new ideas, and in fact even in its ways of rehashing old ones.
Unless these large corporations are happy to put across their mes... See more It seems to me that at the heart of this dicusssion is whether we will ever be replaced as translators. My belief is that we won't. the nature of our work may change, just as it already has with the introduction of translation software in the first place.
Luckily enough for us every world language, as far as I can tell, is infinite in its ability to express new ideas, and in fact even in its ways of rehashing old ones.
Unless these large corporations are happy to put across their messages in precisely the same way as the competition, this sharing of TMs (or however they foresee doing it) will be limited in its scope.
We must not forget our abilities as writers and editors and cultural guardians when we translate. A TM, as huge as it may be, never brings that to the table and corporations that settle for pure reproduction of what has already been written will fall behind.
Automation may offer better translations to those who do not essentially need or have access to them for economic reasons. I foresee a ssituation in which translation becomes more commonplace, but where good translation is still required, just as the blogging world has provided a platform for those who have something to say, but has and will not ever replace well-trained and objective journalism because those who know the difference understand the flaws.
[Edited at 2008-07-25 20:24] ▲ Collapse
| | | | | Is this ethical? | Jul 25, 2008 |
Companies like ours invest huge amount of resources in creating these TMs and glossaries. In addition, many are created specifically for our clients who certainly would not take kindly to us "sharing" their materials with anyone. Confidentiality and our sense of ethic towards our clients property are what keep us in business and growing.
I understand the motivation for localizers and MT supporters to want access to these prized assets, but I do not understand how it could be in the ... See more Companies like ours invest huge amount of resources in creating these TMs and glossaries. In addition, many are created specifically for our clients who certainly would not take kindly to us "sharing" their materials with anyone. Confidentiality and our sense of ethic towards our clients property are what keep us in business and growing.
I understand the motivation for localizers and MT supporters to want access to these prized assets, but I do not understand how it could be in the best interest of a good translation agency to share them. It is a main differentiator in term of quality.
Am I missing something? ▲ Collapse
| | |
|
|
|
Nicole Schnell
United States
Local time: 10:32
English to German
+ ...
In memoriam | I am hesitant. | Jul 26, 2008 |
I was contacted by one of our clients earlier this year regarding this issue. It would save outsourcers a lot of money. However, how can I possibly take responsibility and liability for any translation that is based on, or has been created with the help of such sources?
Unrelated to this discussion with the client mentioned above: Two months ago we got quite burned when approximately 19k of "pre-translated" text turned out to be substandard and unacceptable. We fixed it, I swallowed... See more I was contacted by one of our clients earlier this year regarding this issue. It would save outsourcers a lot of money. However, how can I possibly take responsibility and liability for any translation that is based on, or has been created with the help of such sources?
Unrelated to this discussion with the client mentioned above: Two months ago we got quite burned when approximately 19k of "pre-translated" text turned out to be substandard and unacceptable. We fixed it, I swallowed the cost, and I am not too crazy about any recurrence in the nearest future.
It sounds like a terrific idea, but where is the quality assurance?
Great discussion, BTW. ▲ Collapse
| | | | LouisV (X)
Australia
Local time: 03:32
Dutch to English
+ ...
| pooling of translation memories | Jul 26, 2008 |
I can understand the trend towards automation and reducing the cost of the increasing amount of translation required in a global economy. I agree with others though, that the gradual decrease in translator rates indicates an oversupply of translators, rather than the opposite as claimed to justify this initiative.
However, this idea raises the old hoary chestnut of 'who owns the intellectual property to the body of work out there?'. Is it right that agencies get together and use o... See more I can understand the trend towards automation and reducing the cost of the increasing amount of translation required in a global economy. I agree with others though, that the gradual decrease in translator rates indicates an oversupply of translators, rather than the opposite as claimed to justify this initiative.
However, this idea raises the old hoary chestnut of 'who owns the intellectual property to the body of work out there?'. Is it right that agencies get together and use our intellectual product for their gain without our permission?
I'm sure whether this question has ever been tested in court, and if it has, I would be interested in the answer. Whilst I think that change such as this is inevitable, it would be foolish to happily contribute to the outcome with our intellectual property, when it is likely to benefit one side of the industry and not the other.
Keeping a close eye on it, however, would be a good idea.
Cheers
Louis ▲ Collapse
| | | | | Pages in topic: < [1 2 3 4 5 6 7 8 9 10] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Clients / large translation companies now talking about pooling linguistic data. Should we be there? | 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! » |
| | 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 » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
