How unrealistic are Transit's fuzzy matches?
Thread poster: Hans Lenting

Hans Lenting  Identity Verified
Netherlands
Member (2006)
German to Dutch
Jan 16

Just how unrealistic are Transit's fuzzy matches?

78 percent match

78 % match, really, STAR AG?


[Edited at 2021-01-16 09:38 GMT]


 

Endre Both  Identity Verified
Germany
Local time: 04:10
Member (2002)
English to German
There are worse offenders Jan 16

See here for a few funny samples.

The one you posted appears quite legitimate compared to the above, even though it's useless in terms of workload reduction – the matching algorithm is geared towards sentences, not lists of terms.


 

Hans Lenting  Identity Verified
Netherlands
Member (2006)
German to Dutch
TOPIC STARTER
What makes up a sentence? Jan 16

Endre Both wrote:

the matching algorithm is geared towards sentences, not lists of terms.


Though list of terms are quite common in technical translations, I understand what you mean.

The thing is, Transit doesn't know what makes up a sentence. All it can do, is look for an uppercase letter at the start of a sentence and a full stop, question mark or exclamation mark at the end. So I added these, and hey, the matching percentage is still 78 %:

78 percent match

BTW: I don't mind if I cannot use the match, since there are better ways to be productive than by using fuzzy matching (I have even set the insertion threshold to 99 % ...). However, these useless matches are used to reduce the rate.


 

Heinrich Pesch  Identity Verified
Finland
Local time: 05:10
Member (2003)
Finnish to German
+ ...
Transit still around? Jan 16

Never except reductions for less than 98% matches.

Jorge Payan
 

Hans Lenting  Identity Verified
Netherlands
Member (2006)
German to Dutch
TOPIC STARTER
Yes, still around Jan 16

And still very advanced, compared to some other tools (Studio, memoQ). New features are silently being added to Transit, but old flaws (buggy SQL support etc.) aren't fixed, which is a pity.

Besides that, there are some design decisions that aren't good for freelance translators (e.g. no export of projects as XLIFF). Open standards aren’t a one-way street. If you benefit from XLIFF, you have to offer its export too.

There exists an overview of all features that have
... See more
And still very advanced, compared to some other tools (Studio, memoQ). New features are silently being added to Transit, but old flaws (buggy SQL support etc.) aren't fixed, which is a pity.

Besides that, there are some design decisions that aren't good for freelance translators (e.g. no export of projects as XLIFF). Open standards aren’t a one-way street. If you benefit from XLIFF, you have to offer its export too.

There exists an overview of all features that have been added since the original release of Transit NXT, but I cannot find it on the web. It's quite impressive.


[Edited at 2021-01-17 06:07 GMT]
Collapse


 

Dan Lucas  Identity Verified
United Kingdom
Local time: 03:10
Member (2014)
Japanese to English
Was always wary of it Jan 16

Hans Lenting wrote:
And still very advanced, compared to some other tools (Studio, memoQ).

I have heard that a lot, but I have to say that I didn't really see anything - for my usage - that stood out compared to Studio, and a lot that was more cumbersome. It's approach and UI were certainly, er, idiosyncratic. I used it frequently (sometimes several times a week) but not deeply over a period of about two years for a client that was clearing a huge backlog of untranslated material. Like you, I was never convinced by its fuzzy calculations, so I never really trusted the work that came my way via Transit. I also didn't like the licensing setup.

If I had been able to use Transit for the same sort of work for which I use Studio I would have been able to make a fairer apples-to-apples evaluation. However, things went the opposite way, and the client switched to the Studio toolchain, which in itself tells you something.

Regards,
Dan


 

wotswot  Identity Verified
France
Local time: 04:10
Member (2011)
French to English
Still my CAT tool of choice Jan 17

I agree with Hans about XLIFF (Transit STILL treats it as a "project format" rather than a file format), but for all its faults, Transit is still my favourite CAT tool, after eleven years of using it on a daily basis, sometimes in combination with Studio, which I use mainly to convert XLIFF files or SDLPPX projects to TTX with Legacy converter for translation in Transit.
Why? For 5 main reasons:
- lifetime licence
- macros, which for me save a lot of time to-ing and fro-ing bet
... See more
I agree with Hans about XLIFF (Transit STILL treats it as a "project format" rather than a file format), but for all its faults, Transit is still my favourite CAT tool, after eleven years of using it on a daily basis, sometimes in combination with Studio, which I use mainly to convert XLIFF files or SDLPPX projects to TTX with Legacy converter for translation in Transit.
Why? For 5 main reasons:
- lifetime licence
- macros, which for me save a lot of time to-ing and fro-ing between keyboard and mouse
- "horizontal layout" option: easier to see complete adjacent segments containing very long sentences (notoriously omnipresent in French, legal French in particular), less reliance on preview, IMHO it's a more "natural" layout than the side-by-side layout
- the load and save options in Find/Replace, with which I've built up an extensive library of F&R items over the years, including many regexes (unfortunately in STAR grep, Transit's proprietary regex flavour), whereas in Studio you would have to create your own external F&R regex library and copy and paste each time
- the "bubble windows" option for many windows, saving valuable screen space, whereas in Studio you have to use the mouse and click on tabs to see different windows. Studio devotees say this is easily solved with a second monitor, which I have, but I prefer to use the second monitor for external resources and my permanent browser window (and Alt-tabbing/Ctrl-tabbing in Windows 10/Microsft Edge is such a pain, much more user-friendly in Firefox).

As for Transit's fuzzy percentages (or Studio's for that matter), I've given up trying to fathom them out!
Collapse


 

Dan Lucas  Identity Verified
United Kingdom
Local time: 03:10
Member (2014)
Japanese to English
Pros and cons Jan 17

wotswot wrote:
- lifetime licence

The cost of which they do not publish, but I can only assume is frighteningly large, given that licences for shorter periods can cost several hundred euro each. I'd far sooner pay a smaller amount for a non-time limited licence that I can use for many years than pay a huge amount up-front. I know many people still using Studio 2015 without apparent problems.

- macros, which for me save a lot of time to-ing and fro-ing between keyboard and mouse

This is true. Studio has support for plugins (important, in my opinion), but not macros at the user level. There are third-party workarounds that help, notably AutoHotkey, a scripting language can be applied to virtually any Windows application. I'd file this in the "nice to have" category.

Personally I would prefer SDL to spend some man-hours on fixing or improving other areas of the application that need attention, rather than spend time on adding a macro function. Which brings us nicely to...

- "horizontal layout" option

Something I believe people have been requesting for many years. There is still no sign of it. This is a major problem for some people, and surely it wouldn't be that hard for SDL to implement?

the load and save options in Find/Replace, with which I've built up an extensive library of F&R items over the years

I think this is a new feature in Studio 2021. If so, it took them long enough. It's just administrative functions of the kind you'd give to the developer intern to tackle over the course of the summer.

- the "bubble windows" option for many windows, saving valuable screen space, whereas in Studio you have to use the mouse and click on tabs to see different windows.

Horses for courses. Having briefly experienced (perhaps like yourself) the world of Dos applications back in the late 1980s and early 1990s, when user interfaces could vary dramatically according to the whim of the developer, I appreciate the standard Windows approach. I dislike quirky interfaces intensely, so Transit's "bubble windows" were an annoyance for me - and you can't change an UI wholesale.

Personal choice, at the end of the day. Arguably Transit has had its day, but it's a mature and capable application. Better to have many options for CAT software rather than just a couple. It's been interesting to watch Memsource make significant inroads among my client base in just a couple of years. I suspect SDL's market position isn't as secure as we might have thought even five years ago.

And that's before we consider what the acquisition of SDL by RWS means for Studio in the long term. More resources or less? Higher prices or lower? More dynamic development or benign neglect? Worrying times.

Regards,
Dan

[Edited at 2021-01-17 12:51 GMT]


 

wotswot  Identity Verified
France
Local time: 04:10
Member (2011)
French to English
Pros and cons continued Jan 17

Dan,

"The cost of which they do not publish, but I can only assume is frighteningly large,..." (lifetime licence)

Agreed from their point of view, but as far as I'm concerned, Transit Professional has paid for itself many times over in 11 years, whereas I've had to pay 4 times for major Studio upgrades (2009, 2011, 2015 and 2019). Apparently, 2021 integrates an advanced filter function that was previously a Studio add-on, but I can live with Transit's filters.
... See more
Dan,

"The cost of which they do not publish, but I can only assume is frighteningly large,..." (lifetime licence)

Agreed from their point of view, but as far as I'm concerned, Transit Professional has paid for itself many times over in 11 years, whereas I've had to pay 4 times for major Studio upgrades (2009, 2011, 2015 and 2019). Apparently, 2021 integrates an advanced filter function that was previously a Studio add-on, but I can live with Transit's filters.

Point taken about "quirky" interfaces and the standard Windows approach, although I resent the way Microsoft are gradually killing off all the nice things about Windows 7, introducing often unwanted features in Windows 10, virtually forcing users to use Onedrive and imposing online Office account controls for a one-off Office 2019 purchase (downloaded version).

I wasn't aware of the acquisition of SDL by RWS, can you tell us more?


Richard
Collapse


 

Dan Lucas  Identity Verified
United Kingdom
Local time: 03:10
Member (2014)
Japanese to English
No substantive detail yet Jan 17

wotswot wrote:
I wasn't aware of the acquisition of SDL by RWS, can you tell us more?

All I know is what was reported in the media and by specialist firms such as CSA (comment here) a few months back.

Dan


 


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

Moderator(s) of this forum
Maya Gorgoshidze[Call to this topic]

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

How unrealistic are Transit's fuzzy matches?

Advanced search






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 »
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 »



Forums
  • All of ProZ.com
  • Termsøk
  • Jobber
  • Forumer
  • Multiple search