| موضوع میں صفحات: < [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17] > | The global TM I want... دھاگا پوسٹ کرنے والے: Felipe Gútiez Velasco
|
|---|
Samuel Murray 
نیدر لینڈ
Local time: 08:02
رکن (2006)
افریکانسسےانگریزی
+ ...
Felipe Gútiez wrote:
A global TM should be quick too.
What is your Internet speed?
Has anyone 32 MB bandwith?
I think most South African freelance translators have around 48.8 kbps and they pay per minute. I'm fortunate to have 384 kbps and I pay per byte. But you're not targeting South Africans, are you?
| | | |
Samuel Murray wrote:
From my experience (and this may differ from yours) smaller companies tend to care more for their translators, and provide better opportunities for continued improvement. The translators who work for large companies are those who couldn't get jobs at smaller companies. 
Exactly. Small companies do care A LOT about what they translate and who does it, as they need to have it right from the beginning. They know that paying a bit more saves them a pile of money in the long run.
| | | | | I entirely agree | Sep 3, 2008 |
Marie-Hélène Hayles wrote:
I've been following this thread from the beginning, though I haven't contributed. Essentially (leaving aside the issue of copyright and confidentiality), I am against the idea of a global TM, due to the quality issue....
I entirely agree with your reply Marie-Hélène!
| | | | | More than 4000 readings more than 200 postings... because we repeat a lot | Sep 3, 2008 |
Felipe Gútiez wrote:
More than 4000 readings more than 200 postings
We just have to repeat things a lot because you apparently don't listen. We are all saying the same from the very beginning. This is one more reply, adding to the "success of this posting"...
| | |
|
|
|
| History is important too... | Sep 3, 2008 |
Is there anybody spezialising in history?
I wonder how was the transition from picture to alphabet.
The transition from oral to written.
The transition from writting to print (Güttemberg).
The transition from print to ....again oral(Dragon...:-))?
I think it is very important to learn the lessons of the history. Try not to repeat the same mistakes.
What can history teach us for a global TM?
| | | | | Multidisciplanirity, fresh looks.... | Sep 3, 2008 |
I saw in a TV programm that a Silicon Valley tycoon (or genius) decided to stop swimming on money made with software and started a completely different thing: photovoltaics.
He just used logic and money:
-he looked for the best scientist in internet (in this case it was a Chinese)
-he looked for the best managers available (some ex-IBM) and
-he put them together and spend some billion euros or so
He had absolutely no idea about photovoltaics and he found new... See more I saw in a TV programm that a Silicon Valley tycoon (or genius) decided to stop swimming on money made with software and started a completely different thing: photovoltaics.
He just used logic and money:
-he looked for the best scientist in internet (in this case it was a Chinese)
-he looked for the best managers available (some ex-IBM) and
-he put them together and spend some billion euros or so
He had absolutely no idea about photovoltaics and he found new ways of solving the problems almost for every question he had to overcome.
Logic, multidisciplinarity, fresh air.
Is there anyone who has ABSOLUTELY NO IDEA about translation?
If you know someone (a relative, a colleague, a friend), please try this:
show him/her the questions posed by a possible global TM (in a general way, not very precise as to keep his/her free of mind, naivety or like you want to name it) and, if you want, tell us in this thread what his/her proposals were.
I wonder if we will find lot of fantastic new approachs.
I propheshy a swift from a culture of companies to a culture of fori or platforms. And this will happen quick, very, very quick. ▲ Collapse
| | | | | The lessons from history | Sep 3, 2008 |
Felipe Gútiez wrote:
What can history teach us for a global TM?
If you want something done properly, do it yourself.
| | | |
Charlie Bavington wrote:
Felipe Gútiez wrote:
What can history teach us for a global TM?
If you want something done properly, do it yourself.
I think is a quotation from Al Capone? (just a joke):-)
| | |
|
|
|
sent twice
[Edited at 2008-09-03 13:52]
| | | |
Felipe Gútiez wrote:
Why should professional top translators be against an entity consisting of thousand of translators collaborating in a global TM? What is wrong with it?
Perhaps because professional "top translators" don't personally know thousands of other "top translators" nor are they familiar with their work.. so in your utopic money-making global scheme, the so-called "top translators" would have to blindly trust other alleged professionals whose quality they have never corroborated. Top as they might allegedly be, I have to add "professional my ***".
Felipe Gútiez wrote:
Is that not better than having to deal with thousands of disorganised TMs which the client (who in most cases have absolutely no idea about how a TM works or what ambiguity or homonimy means) imposed us by force now or in the future?
Oh my, you are absolutely right!! A global TM created by thousands of allegedly top translators is bound to be FAR more organized and trustworthy than the TMs created by a company, which will probably include units by only 20 or so different people.
Felipe Gútiez wrote:
Top translators could be just revisors (like for example in the European Union or in big translation agencies). Medium-level translators could be the translators receiving inmediate feedback in online collaboration to questions/partial translations/.
And who, praytell, would be all-knowledgeable and omnipotent one who decides who is a "top translator" and who isn't in this global, thousand-translators' money-making scheme of yours?
I wouldn't touch a TM created by a thousand different people, with a thousand different styles, a thousand different criteria, a thousand different preference, etc, with a ten-foot pole, even if it every single one of those people were a tried and true, 1000% corroborated "professional top translator".
"Too many cooks spoil the broth", as the saying goes. It's just a matter of common sense.
Andrea
[Edited at 2008-09-03 13:57]
| | | |
Charlie Bavington wrote:
Felipe Gútiez wrote:
What can history teach us for a global TM?
If you want something done properly, do it yourself.
Greetings
| | | | | The history of this posting | Sep 3, 2008 |
Felipe Gútiez wrote:
I think it is very important to learn the lessons of the history. Try not to repeat the same mistakes.
Well, the history of this posting tells me that it is useless to spend more time in it... You simply don't listen and keep asking more and more questions whose answers you barely care about...
So..... cheers everyone! See you in another topic.
| | |
|
|
|
| |
Hi, this is a draft summery of the thread. I think that Samuel could do a much better summary. Not everybody have such an ability for summarising in an unbiased way like him. I cannot do it that good. But here is my try (the first 4 pages) Any improvement is desired and welcome
Background
Many of us run across phrases and sentences that we've translated probably dozens of times: legalese, manuals, ... See more Hi, this is a draft summery of the thread. I think that Samuel could do a much better summary. Not everybody have such an ability for summarising in an unbiased way like him. I cannot do it that good. But here is my try (the first 4 pages) Any improvement is desired and welcome
Background
Many of us run across phrases and sentences that we've translated probably dozens of times: legalese, manuals, companies' statements, certificates, etc. Basically terms, phrases and sentences very generic that we have to translate constantly
Thread from Henry about this topic
http://www.proz.com/forum/prozcom:_translator_coop/110929-clients_large_translation_companies_now_talking_about_pooling_linguistic_data_should_we_be_there.html
I posted in a thread in Spanish about wether or not give up our TMs
Thread in Spanish
http://www.proz.com/forum/spanish/113411-¿es_legal_entregar_las_memorias_de_traducción.html
I realised that we needed a new thread about the global TM we want….
The idea is to get a common idea of what us, translators want and need from a global TM that could be useful for us. This thread is to get some feedback for this idea
Is like choosing the extras for our personalised car, bought in Internet, of course: air-conditioning, 7 seats, etc,
First question: the name.
In Spanish could be "el baúl de los recuerdos". Buscando en el baúl de los recuerdos. The name was considered by some people a great name. http://www.youtube.com/watch?v=rudBC9cfJtM (look at you tube about Karina, buscando en el baúl de los recuerdos).
Situation:
-Big players in the question of TM sharing:
-TAUS, http://www.translationautomation.com/
-Wordfast very large translation memory, TDA, etc
- With 4,600 employees in 45 offices around the world and a network of 25,000 freelance translators skilled in more than 200 languages, the company is three times the size of its closest competitor in the localization business
Logoport from Lionbridge is something in the art (but private) www.lionbridge.com and https://freeway.lionbridge.com/default.aspx
More Info on Logoport:
http://www.reuters.com/article/pressRelease/idUS163067%2020-Feb-2008%20PRN20080220
In short, they add 2 million words per day, from work done for their own clients (they use Logoport internally as we use Trados, thus without generating any confidentiality issue) and the resulting macro TM has gone through some internal quality check. At least is of some uniform quality, which has allowed them to get and keep clients such as Bayer, Cisco, DuPont, GE, Google, IBM, Merck, Merrill Lynch, Microsoft, Morgan Stanley, Nokia, Pfizer, Sony, and Wal-Mart
- Netscape used to make a TM of common error messages available some time ago. More info about that?
-Small players too
-www.tolingo.com
Ideal features of the TM:
General picture:
- Could be paying, be it: per time, per Tus,etc. Some think it should´t be paying.
- Could be a feature of: Google, Wikipedia, ProZ, a new company, a cooperative of translators?. As an example of cooperative you can see MCC at www.mcc.es
- We could use open source or non-copyrighted files.
- A possible vision for a global TM is where a translator inputs his requirements online (eg subject field, type of document, etc), and the global TM either generates or selects a TM or TMs based on those requirements. The translator downloads the TM(s), uses it, and then contributes the TM again (if he has his client's permission).
- A global TM site could be a kind of social networking site, where groups of translators vet each other's work within the group, but in a way which makes the TMs globally available.
- If a translator can convince his client of the benefit of participating in a global TM, he could have the client sign (or accept via e-mail) a form of permission to donate some of the translation to the global TM
- One way in which a global TM might generate a TM is if the translator uploads his documents, the system does pre-matching and outputs a TM that contains 10 (or 100) possible matches for each segment in the source text. This should be done cleverly so that the TM doesn't contain 10 matches that are 90% similar to each other!
- There could also be several levels of checks performed on submitted TMs to ensure that identifying information is removed from the TM.
-Has the name and "impresum" of the person in charge for that TM, just in case it provides some illegal information
- very, very, very big (apparently Wordfast and Trados could have problems with a very big TM).
-quick. One of the reasons of the "apparent" failure of Logoport is that it is not quick enough. The speed of a TM is essentially determined by something other than the TM, basically the speed of the computer that hosts it. What about the Internet connection speed?
-searchable by creator, date, language, etc
-with lots of pictures or even videos for terminology. Pictures should be added as placeables. What means add pictures as placeble?
-with terminology tool
-with terminology tool scalable
-multilingual. Some think this would be not very useful.
-with lots of literary works translated (you can call the Cultural Heritage Unit at the European Commission to show interest, so that they can boost collaboration between the libraries everywhere) Some people does not want literature in it.
- TUs are labeled with their subject matters so they can be searched by subject matter so that the list of results can be refined using such a subject matter label
- the TM contains parallel texts from governments. It could conatin complete, aligned, official translations of the BGB (Bundesgeseztzbuch), HGB and other German legal frameworks. That would save a bit of research time, and for that sort of material should work with the "official" translations as long as they are not actually wrong
- Collaboration in a global TM could be all right with small, focused teams with competent people and a limited mission
Assesment features and grading:
- TMs would have to have a very strict standards to maintain a level of confidentiality and quality control. A QA team of translators (at least in the begging) could asses that, implementing a rating system like Viktoria suggested. Once a TM reaches certain score, the TM is made public for its members.
- the time at which the translator is in a position to grade the TUs in the TM is necessarily removed from the TM system itself, and the translator can therefore only grade the TM or TUs afterwards (or beforehand) using a separate process.
- Another problem is that some participants wil be eager to get TMs but not so eager to spend time doing gradings. A TM grading system would therefore have to be somewhat automatic. Grading should be somehow rewarded.
-Possible process. The translator uploads his current TM to the system, the system compares it to the TM that was previously downloaded by the translator, and makes a guess about which TUs from the original TM were probably used by the translator. The translator is shown these sentences on screen, he selects the ones that he remembers getting from the TM, and the system displays the original TUs to him to grade. The process should be automated (also with e-mail reminders from the sytem).
- the fuzzy that I chose can be evaluated from 1 to 5 by the translator using it and this evaluation is automatically reintroduced in the system so that a democratic quality control system has been created.
For this phrase, in this topic, in this context, this is the most acclaimed translation from this language into that language.
We can name it: crowd quality control system or democratic quality control system or translator-choice based quality control system.
- The translation units (or segments) should be labeled with confidence levels. For example two different confidence levels could be used. One would be established by a committee and another by the users of the TM (when you find a TU, you have the option to grade it and the average of all ratings collected is displayed along with the TU). This would be similar to download.com ratings: one by the editor and one by the users (I have a tendency to trust the latter more than the former, but with language, this may be reversed).
- a translator would first have to be checked in order to be allowed to publish
- tmx format or a format compatible. Some think is better an inhouse format.
- The system must be completely unable to record the TUs I sent to be translated and their translations. (Is it at all feasible?) Otherwise, the translator will have to ask the client's authorization to use it.
-
Arguments against a global TM:
- Clients forbid sharing TMS.
-It would violate confidentiality, which is clearly stated in the Codes of Ethics from different associations around the world
- Why should companies have to pay for databases (in case the global TM was paying), when they can manage them and use them for free?.
- Absolute security in Internet is impossible. A global TM would only make things even worse, that is more un-secured.
-Some people tried to bribe the author of the thread offering 30 US Dollar to drop it away. (or is that an argument in favour?)
- The idea of a single global TM tool is stupid and potentially illegal
- Except for very narrow, specific subject areas, all you can think of doing with such material is printing it out and using it as toilet paper
- There is doubts about private TM content being of much use to most good translators in specific fields. There are certain quirks of style which would surely drive many of them up a wall or at least require significant editing for style compatibility
- The result of such a patchwork would be a personal embarrassment to individual translators
- Even if all the end customers loved the idea of a "global TM", it would be a waste of effort except in a few narrow instances
Arguments in favour of a global TM:
-History/experience of tmmarketplace: www.tmmarketplace.com No judiciary process against.
-Generally speaking corporations don't care about translations, so that probably there would be no problems with copyright issues.
-Better doing/designing/controlling than "having to work with….TM imposed"
- We do need to do something but we need a starting point
- Most end clients consider their TMs as a residual product without interest to them
- Many end clients do not know that their translations are being made with TM tools
- Translators would improve visibility (like in KudoZ). The best graded authors of Tus could be contacted by end clients/agencies.
- We can see the same material for the same companies from half a dozen agencies or more - the only differences in many cases is what the middleman is willing to take for a cut and what resources the end customer has provided to that particular agent. Sometimes we're even talking about the same document to be translated... on those occasions a translator could figure he is competing against himself ▲ Collapse
| | | |
Samuel Murray wrote:
Yes, large corporations spend more money on translation in total, but that doesn't mean they spend more money on translation per capita. In fact, I suspect large corporations often spend less per capita than smaller businesses. And I doubt if the big corporations necessarily draw translators of higher quality.
That's exactly what I thought, too, when I read Felipe's post. Might I add that I am pretty sure that if those big spenders were not in the circuit, I am convinced that the global rate average would be at least 25% higher.
| | | | | موضوع میں صفحات: < [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » The global TM I want... | 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 » |
| | 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 
Translation news related to CAT tools
