Translation - art & business » Money matters »
Pages in topic:   < [1 2]
Fuzzy matches - the effects of percentage reductions.
Thread poster: Simon Taylor (X)
Samuel Murray
Samuel Murray  Identity Verified
Netherlands
Local time: 03:46
Member (2006)
English to Afrikaans
+ ...
Always use the tool that produced the match analysis, or... Aug 25, 2016

Mirko Mainardi wrote:
Simon Taylor wrote:
What is the definition of a match?

Basically, what we are paid when working with a CAT tool and a fuzzy matrix depends on some CAT tool developer's "critical business secret". Funny, isn't it?


That is true, but it is also understandable. The most basic matching is simple, but since we're dealing here with language, pure mathematics don't always produce suitable results, and so CAT developers have to spend a lot of time and resources to research and develop fuzzy matching that produces more useful matches for their users. If a CAT developer were to publicly tell exactly how he calculates his fuzzy matches, his competitors would be able to copy it at very little cost, if they thought that that developer's match engine is a good one.

What's more, a match analysis is only accurate if the translation is done in the CAT tool that produced that match analysis. If one CAT tool serves a match but another one doesn't, then the first tool's analysis will show that there is a match, but the other tool will not produce a match when the time comes. That said, some CAT tools consider certain matches to be "high" fuzzies even though other tools may consider them lower fuzzies.

I've taken a silly test text and checked how a number of CAT tools do their matching:

WFC
singing dancing talking (first segment)
sing dance talk = 58% match with "singing dancing talking"
singer dancer talker = 66% match with "sing dance talk"
unsing undance untalk = no match
dissing disdance distalk = no match

WFA
singing dancing talking (first segment)
sing dance talk = 52% match with "singing dancing talking"
singer dancer talker = 60% match with "sing dance talk"
unsing undance untalk = no match
dissing disdance distalk = 50% match with "unsing undance untalk"

WFP 3x
singing dancing talking (first segment)
sing dance talk = no match
singer dancer talker = 75% match with "sing dance talk"
unsing undance untalk = no match
dissing disdance distalk = no match

WFP 4x
singing dancing talking (first segment)
sing dance talk = no match
singer dancer talker = no match
unsing undance untalk = no match
dissing disdance distalk = no match

Trados 2015
singing dancing talking (first segment)
sing dance talk = 91% match with "singing dancing talking"
singer dancer talker = 91% match with both "singing dancing talking" and "sing dance talk"
unsing undance untalk = no match
dissing disdance distalk = no match

OmegaT
singing dancing talking (first segment)
sing dance talk = 40% match with "singing dancing talking"
singer dancer talker = no match
unsing undance untalk = no match
dissing disdance distalk = no match

Virtaal
singing dancing talking (first segment)
sing dance talk = no match
singer dancer talker = 75% match with "sing dance talk"
unsing undance untalk = 71% match with "sing dance talk"
dissing disdance distalk = no match
 
Mirko Mainardi
Mirko Mainardi  Identity Verified
Italy
Local time: 03:46
Member
English to Italian
Interesting Aug 25, 2016

Rudolf Vedo CT wrote:

philgoddard wrote:

I've never understood why anyone should buy software not to improve their productivity by doing less work for the same money, but to charge their customers less.


Well, brilliant marketing on the part of CAT makers, for one. Translators initially believe that it will improve their productvity but ultimately cave in to pressure from agencies, who are sold the same software with the promise that it will allow them to pay their translators less.

Though even SDL seems to have abandoned trying to convince now-jaded translators that there are any net productivity gains to be had. At a Trados 2015 presentation I attended last year while in Europe, the reseller didn't try to convince us we could make more money by upgrading. Instead, they drove home the point - not once but several times - that if we upgraded we would get more work because we could charge less.

CAT tool making is an odd business to be in, for sure. Basically it's a zero-sum game. Anything they come up to benefit one type of customer (translator or agency), results in "money left on the table" for the respective other customer, resulting in permanent conflict between the two. Increasingly it seems like the only way to "win" is not to play at all.


Interesting analysis (with which I largely agree). However, specifically about the part in bold, if you read the OP in the thread started by SDL about their upcoming 2017 version (which mechanically repeats the sales pitch from their website), you'll notice statements such as "SDL Trados Studio 2017 will bring revolutionary innovation to help you translate faster and ensure high quality translations", "Our upLIFT technology will help you translate new segments faster", or, from their website "Studio 2017 will simplify tasks and make items easier to find - helping you translate faster and enjoy your user experience more". So, back to basics, it seems...

As for me, the main reason why I purchased licenses for Studio and MemoQ was actually not to "translate faster", but to be able to work with clients who use those tools (often with online projects), as I believe Studio+MemoQ cover a substantial market segment in that respect.

Unfortunately, this also means experiencing the things you mention and entering a vicious circle where you are more or less "forced" to keep buying new "upgrades" even if they add zero value to you (and can actually make you lose time to bugs/issues and/or to get used to new interfaces, features, etc.), just because some of the clients you're working with are "upgrading" themselves...
 
Pages in topic:   < [1 2]


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

Moderator(s) of this forum
Jorge Rodrigues[Call to this topic]
Andriy Bublikov[Call to this topic]
Rita Pang[Call to this topic]
James Heppe-Smith[Call to this topic]

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

Fuzzy matches - the effects of percentage reductions.







TM-Town
Manage your TMs and Terms ... and boost your translation business

Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.

More info »
Trados Studio 2022 Freelance
The leading translation software used by over 270,000 translators.

Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop and cloud solution, empowering you to work in the most efficient and cost-effective way.

More info »