Pages in topic:   < [1 2 3 4 5 6 7] >
"Mini-contest" launched (on new "beta" contests platform)
Thread poster: Henry Dotterer
Dr. Matthias Schauen
Dr. Matthias Schauen  Identity Verified
Germany
Local time: 15:26
Member (2007)
English to German
Finalist determination: segments vs entries Oct 31, 2012

Hello,

how were the finalists determined? I seems to me (by looking at the ratings of the non-finalists) that they were determined solely by their average segments rating, not by their entry rating. If yes, my rating the entries seems to have been useless work. How was this computed exactly?

Cheers,
Matthias

[Edited at 2012-10-31 13:38 GMT]


 
Alexander Onishko
Alexander Onishko  Identity Verified
Russian to English
+ ...
The option "Rate Segments" is still active Nov 2, 2012

Hi! The option "Rate Segments" is still active. I think this has no sence in the Final phase?

 
Henry Dotterer
Henry Dotterer
Local time: 10:26
SITE FOUNDER
TOPIC STARTER
How finalists were determined Nov 2, 2012

Dr. Matthias Schauen wrote:
how were the finalists determined? I seems to me (by looking at the ratings of the non-finalists) that they were determined solely by their average segments rating, not by their entry rating. If yes, my rating the entries seems to have been useless work. How was this computed exactly?

Both forms of rating were taken into consideration. In a manual process, we looked at the leaders from each method. Most of the finalists were leaders by both methods. Where the situation was less clear, something like a weighted mixture of the two ratings was used. (So if twelve people rated an entry on a segment-by-segment basis, and only four rated on an entry basis, the segment rating was given more weight.)

In two language pairs out of the sixty (the two that were delayed), there was not a tight correspondence between the two rating methods. Here we looked deeper when considering which entries had a chance to win. Not wanting to rule out an entry that had a chance to win, we included the top entries from each rating methodology, and then to be extra sure, I added in a few extra that were notable in certain ways. (This is why Spanish has 9 finalists -- two more than the usual limit of 7.)


 
Henry Dotterer
Henry Dotterer
Local time: 10:26
SITE FOUNDER
TOPIC STARTER
New idea: Composite "best translation" Nov 2, 2012

Alexander Onishko wrote:
Hi! The option "Rate Segments" is still active. I think this has no sence in the Final phase

Good question. Segment-based rating will not have any impact on the selection of winners. Sorry for any possible confusion -- we will add a note to that effect on the page.

Segment-based rating is still on intentionally, however. The purpose is a new feature that I hope people will find interesting and that will add another dimension to the contests. Segment-based ratings make it possible to create a composite "best translation" -- the "community translation", if you will -- composed of the segments rated most highly. That will appear soon.


 
Katalin Horváth McClure
Katalin Horváth McClure  Identity Verified
United States
Local time: 10:26
Member (2002)
English to Hungarian
+ ...
You are joking, right? Nov 2, 2012

Henry Dotterer wrote:

Segment-based rating is still on intentionally, however. The purpose is a new feature that I hope people will find interesting and that will add another dimension to the contests. Segment-based ratings make it possible to create a composite "best translation" -- the "community translation", if you will -- composed of the segments rated most highly. That will appear soon.


Henry,
You are joking, right?
Or are you planning on having future contests for translating texts that are made up of standalone segments, such as the Yogi Berra quotes?
Otherwise, I can't see how could you be serious about proposing that any continuous text, where the sentences must connect to each other, and the entire text must have a (God forbid!!!) good, natural flow, consistent style and terminology could be put together from segments translated by different people to create a "best translation"? It is like asking a bunch of kids which was their favorite animal head, the cat's the dog's, the goats, etc. Then, which one is their favorite body? The fish's body? The horse's? Which limbs do they prefer? Hooves, paws, wings, fins? How about the tail? Then put the favorite pieces together - what do you get? A monster.

Please, go back to any of the previous contest texts, and try that for the three top texts. This does not make sense.

I thought ProZ.com was aiming for promoting the highest standards in the profession, how can an idea like this even conceived in such a context?

I just hope that we, the participants of this mini contest are not assisting unknowingly in a crowdsourcing experiment.

Katalin


 
Henry Dotterer
Henry Dotterer
Local time: 10:26
SITE FOUNDER
TOPIC STARTER
Nope Nov 2, 2012

Katalin Horváth McClure wrote:
Henry Dotterer wrote:
Segment-based rating is still on intentionally, however. The purpose is a new feature that I hope people will find interesting and that will add another dimension to the contests. Segment-based ratings make it possible to create a composite "best translation" -- the "community translation", if you will -- composed of the segments rated most highly. That will appear soon.

You are joking, right?

Not at all! Let's see how it is when it comes out... if you're not curious don't look...


 
Alexander Onishko
Alexander Onishko  Identity Verified
Russian to English
+ ...
*** Nov 2, 2012

Segment-based rating is still on intentionally, however. The purpose is a new feature that I hope people will find interesting and that will add another dimension to the contests. Segment-based ratings make it possible to create a composite "best translation" -- the "community translation", if you will -- composed of the segments rated most highly. That will appear soon.


Hi, Henry!

Sorry to say that, but I don't think it's a good idea. Really. "What is too good is not healthy" as one Ukrainian proverb says.

In my humble opinion the number of dimensions, which the contest already has, might be quite suffisient. No need for a new one.

[Edited at 2012-11-02 20:38 GMT]


 
Henry Dotterer
Henry Dotterer
Local time: 10:26
SITE FOUNDER
TOPIC STARTER
Thanks, Alexander Nov 6, 2012

Alexander Onishko wrote:
In my humble opinion the number of dimensions, which the contest already has, might be quite suffisient. No need for a new one.

Thanks, Alexander. I was wondering that myself, whether this will be too much. And you may be right. On the other hand, there are many people whose interaction with contests is only sporadic. For example, some people who did not submit entries in this contest will come in now to vote for a best entry, and not come back again. For them, I guess it will be a mildly interesting feature to look at, in passing.

Plus, I think people will get a kick out of having one of their segments selected as best, even if they did not win or even make the finals. (I would.)

I opened a new thread specifically for feedback on "composite best translation". See: http://www.proz.com/topic/236768 I'm particularly interested in hearing opinions on the overall quality of the "composite best"s.


 
Alexander Onishko
Alexander Onishko  Identity Verified
Russian to English
+ ...
*** Nov 8, 2012

Henry Dotterer wrote:

Alexander Onishko wrote:
In my humble opinion the number of dimensions, which the contest already has, might be quite suffisient. No need for a new one.

Thanks, Alexander. I was wondering that myself, whether this will be too much. And you may be right. On the other hand, there are many people whose interaction with contests is only sporadic. For example, some people who did not submit entries in this contest will come in now to vote for a best entry, and not come back again. For them, I guess it will be a mildly interesting feature to look at, in passing.

Plus, I think people will get a kick out of having one of their segments selected as best, even if they did not win or even make the finals. (I would.)

I opened a new thread specifically for feedback on "composite best translation". See: http://www.proz.com/topic/236768 I'm particularly interested in hearing opinions on the overall quality of the "composite best"s.


Hi, Henry!

On having a look on that new feature I must say that it looks not bad. However, maybe more feedback from users should be collected in order to determine whether this feature should be included into the next contests.


 
Alexander Onishko
Alexander Onishko  Identity Verified
Russian to English
+ ...
by the way... Nov 8, 2012

... seems that the Winners page is not working

http://www.proz.com/translation-contests/winners

I definitely remember that i was able to open it before.


 
Matt Petrowski
Matt Petrowski
United States
Member (2008)
English to Spanish
+ ...
Winners page restored Nov 8, 2012

Alexander Onishko wrote:

... seems that the Winners page is not working

http://www.proz.com/translation-contests/winners

I definitely remember that i was able to open it before.


Sorry about that, the page is now linked to and visible again.

Matt


 
Roland Nienerza
Roland Nienerza  Identity Verified

Local time: 15:26
English to German
+ ...
the trickiness of shenanigans - Nov 9, 2012

Katalin Horváth McClure wrote:

Henry Dotterer wrote:

Segment-based rating is still on intentionally, however. The purpose is a new feature that I hope people will find interesting and that will add another dimension to the contests. Segment-based ratings make it possible to create a composite "best translation" -- the "community translation", if you will -- composed of the segments rated most highly. That will appear soon.


Henry,
You are joking, right?
Or are you planning on having future contests for translating texts that are made up of standalone segments, such as the Yogi Berra quotes?
Otherwise, I can't see how could you be serious about proposing that any continuous text, where the sentences must connect to each other, and the entire text must have a (God forbid!!!) good, natural flow, consistent style and terminology could be put together from segments translated by different people to create a "best translation"? It is like asking a bunch of kids which was their favorite animal head, the cat's the dog's, the goats, etc. Then, which one is their favorite body? The fish's body? The horse's? Which limbs do they prefer? Hooves, paws, wings, fins? How about the tail? Then put the favorite pieces together - what do you get? A monster.

Please, go back to any of the previous contest texts, and try that for the three top texts. This does not make sense.

I thought ProZ.com was aiming for promoting the highest standards in the profession, how can an idea like this even conceived in such a context?

I just hope that we, the participants of this mini contest are not assisting unknowingly in a crowdsourcing experiment.

Katalin



Hello Katlin!

I had looked at this pretty much as you do here - that it would not make much - if any - sense to propose voting to be done on a segment by segment or sentence by sentence scheme also for "normal" texts, i.e. texts that are not composed of isolated sentences like the source material of this mini venture.

But I found out after a little while that the intent of this mini contest - with a quite appreciable input effort of Henry - is to try to come to a certain way of a) somehow objective and b) somehow time-efficient way of rating.

The last two contests before this mini/beta exercise had a fairly nice system of rating by allowing to comment on individual sentences out of the complete context. The problem was then - and hence the title of this posting - that the discussions became in some cases very intricate and in many cases very polemic. And this, together with the options of agreeing or disagreeing with any comment given, made it in the end rather difficult to summarize or to quantify a clear judgment for a submission.

Your objection to the idea of extending the segment by segment rating from this contest with a compound of individual remarks to a continuous text touches in an interesting manner again at the core conundrum of the translation process - the famous question of how precise a target text has to be and of how free it could be allowed to be in order to be called a good translation.

While on the one hand it will be already technically quite troublesome if not altogether impractical to present and to vote the sentences of source and target of a somewhat longer text individually it is on the other hand indisputable for me that in case that someone considers a translation as flawed he should not just express "an opinion" but should demonstrate his point very concretely. And in doing this he should be able to single out the mistaken sentences very precisely - and rate them according to his view.

For me it is a positive - because more practical - approach to allow voting sentences, whether some or all remains to be seen, individually. Just voting - on this scale of 1 to 5 - saves the ultimately no more manageable process of bickering and verbal wrangling. This type of voting is easy to be done and permits easy and clear quantifying of a qualitative judgment.

But as I see it the contest designers and framers should make up their minds. If they allow voting on segments or sentences individually this voting should be compounded into the final voting of the piece.

It makes absolutely no sense for me that a voting for an entity given by 4, 5 or 6 people can outweigh the voting of 150 and more people on sentences. So far I see no solution for this problem. If the segment voting is finally ignored, as it looks to me, it does not make sense to offer it at all.


 
Katalin Horváth McClure
Katalin Horváth McClure  Identity Verified
United States
Local time: 10:26
Member (2002)
English to Hungarian
+ ...
You completely misunderstood Nov 12, 2012

[quote]Roland Nienerza wrote:


Your objection to the idea of extending the segment by segment rating from this contest with a compound of individual remarks


Roland,
I am not talking about the voting.
I am talking about the "experiment" or "new feature" what Henry calls "composite best" whereby the system creates a complete "translation" using individual sentences taken from different submissions. This time, since the source is NOT a piece of continuous text, but 14 individual quotes, it is OK, as I said, it is simply like 14 sub-contests, each for one quote.
However, Henry was talking about using this in the future, for other rounds of contests.
I question this idea, because every translator knows that translation is not about translating isolated sentences. We normally translate continuous texts, and it is important that it has a natural flow from one sentence to the other, consistent style and terminology, etc. etc. It is clear to every professional translator, just the same way it is clear that translating a sentence is not about looking up the individual words in the dictionary.

Since ProZ.com supposed to be a place for professionals, I question how this "feature" would promote professionalism and the value of quality work.

By the way, there is another thread about this feature, so perhaps the discussion should continue there:
http://www.proz.com/forum/prozcom_translation_contests/236768-new_contest_feature:_composite_best_translation.html

Katalin


 
Alessandro Marchesello
Alessandro Marchesello  Identity Verified
Italy
Local time: 15:26
Member (2005)
English to Italian
+ ...
Has this contest ended without winners? Nov 14, 2012

Hi everybody.

I'm just wondering why winners haven't been announced yet considering that the final voting phase ended more than 48 hours ago.
Is there any problem with vote counting?

Thanks in advance for your reply.


 
Henry Dotterer
Henry Dotterer
Local time: 10:26
SITE FOUNDER
TOPIC STARTER
Sorry for the delay Nov 14, 2012

Alessandro Marchesello wrote:

Hi everybody.

I'm just wondering why winners haven't been announced yet considering that the final voting phase ended more than 48 hours ago.
Is there any problem with vote counting?

Thanks in advance for your reply.

Hi Alessandro. Sorry for the delay. I was out for personal reasons for a few days, this is the reason for the delay. Hopefully we'll be in a position to announce the winners (at least in most pairs) tomorrow.


 
Pages in topic:   < [1 2 3 4 5 6 7] >


To report site rules violations or get help, contact a site moderator:

Moderator(s) of this forum
Lucia Leszinsky[Call to this topic]

You can also contact site staff by submitting a support request »

"Mini-contest" launched (on new "beta" contests platform)






Trados Business Manager Lite
Create customer quotes and invoices from within Trados Studio

Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.

More info »
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 »