Contextual search engines Thread poster: gabritran
|
|---|
| | Tom in London
United Kingdom
Local time: 06:55
Member (2008)
Italian to English
I use Linguee and Reverso and *sometimes* find them useful in my language pair, though only in comparison with other tools and only in particular subject areas. They can't be used on their own. In terms of their web technology they do seem a lot more advanced than Kudoz and the Proz glossaries. Maybe that's one of various things Proz might like to consider when they get around to updating the website (an enormous task, I imagine, but an important one if Proz wants to stay at the top!)
[Edited at 2015-02-06 13:01 GMT]
| | | | Luca Tutino 
Italy
Member (2002)
English to Italian
+ ...
I am often disturbed by the sheer number of low quality hit that my searches tend to find when I google for bilingual documents. Save for rare exceptions, contextual engines bring low-quality results in three respects:
- They often make it more difficult to narrow down a search to a bilingual document in a special field,
- Their results mostly look just as guess attempts or poor translations, and
- Even when they appear to be good, they can very rarely be considered as ... See more I am often disturbed by the sheer number of low quality hit that my searches tend to find when I google for bilingual documents. Save for rare exceptions, contextual engines bring low-quality results in three respects:
- They often make it more difficult to narrow down a search to a bilingual document in a special field,
- Their results mostly look just as guess attempts or poor translations, and
- Even when they appear to be good, they can very rarely be considered as positive evidence of proper use of the terminology.
In the long run, for complex terminological searches the total loss of time and energy caused (even if only by the need of adding -site: clauses) can be considerable. I believe that I never resolved a term question using them.
This might be a specific problem of my language combinations (EN>IT, FR>IT, NL>IT). I have seen good EN-FR colleagues rather happy about them. In any case, I am afraid that in the present state of the art the damage is higher than the advantage in my case.
Despite all the noise, I see the web is a precious resource for any language professional. I used to dream about something like it since when the first modem arrived in my editorial department, in the late Eighties. But discerning the quality of the resources is always crucial. Today, among the multilingual resources, Wikipedia is mostly a great addition. Alibaba is pure pollution. Glosbe, Linguee and Reverso are rather poor as far as IT as target language is involved. ▲ Collapse
| | | | Tom in London
United Kingdom
Local time: 06:55
Member (2008)
Italian to English
Luca Tutino wrote:
- Even when they appear to be good, they can very rarely be considered as positive evidence of proper use of the terminology.
Linguee is good because it presents the hits in the context from which they are taken, and you, as the user, can improve the engine by selecting the hit you think comes closest to what you were looking for. I'm very impressed by Linguee, at least in my language pair. I recently had a problematic job that included a lot of legal terminology, and Linguee proved useful because it suggested the correct legal terms, used in a context that was credible.
| | |
|
|
|
I use Linguee, but am no fan of Reverso. You need to be careful with it though and of course do some further research.
| | | | Luca Tutino
Italy
Member (2002)
English to Italian
+ ...
Tom in London wrote:
and Linguee proved useful because it suggested the correct legal terms, used in a context that was credible.
Yes, legal terms can be considered as an exception: they are generally good. But I am under the impression that better resources are available already in this field. (Even if my favourite, Eur-lex, recently became less efficient for us.)
The inclusion of source url is a positive aspect, as far as the sources prove useful.
Linguee might become a very good resource in time. I hope so. But a the moment, for me, its noise is higher than the useful information. And I wonder if unqualified user votes can be considered as a path towards quality.
| | | | Michael Beijer
United Kingdom
Local time: 06:55
Member (2009)
Dutch to English
+ ...
| Michael Beijer
United Kingdom
Local time: 06:55
Member (2009)
Dutch to English
+ ...
| EUR-Lex > EUR+Lex | Feb 6, 2015 |
Luca Tutino wrote: Tom in London wrote:
and Linguee proved useful because it suggested the correct legal terms, used in a context that was credible. Yes, legal terms can be considered as an exception: they are generally good. But I am under the impression that better resources are available already in this field. (Even if my favourite, Eur-lex, recently became less efficient for us.) The inclusion of source url is a positive aspect, as far as the sources prove useful. Linguee might become a very good resource in time. I hope so. But a the moment, for me, its noise is higher than the useful information. And I wonder if unqualified user votes can be considered as a path towards quality.
Hi Luca,
There is another (better) way to access the EUR-Lex data (in addition to the TMX route I sketched above):
http://anyterm.info/eur%20lex/index.php?jezik=angl&beseda=
The link seems to get broken by the Proz forum software. "eur" plus "lex" is being changed to: "eur%20lex". Just change the "%20" back to a plus sign (+) in the URL and it ought to work again. Not sure what is going on.
Michael
[Edited at 2015-02-06 22:43 GMT]
| | |
|
|
|
Thanks for your comments. I’ve just downloaded TMlookup which can be extremely useful indeed. Michael, did you really manage to use the contents of
http://opus.lingfil.uu.se/? It's a huge amount of data.
I'm under the impression that contextual search engines such as Reverso Context are constantly being improved so let’s wait and see...
[Modificato alle 2015-02-07... See more Thanks for your comments. I’ve just downloaded TMlookup which can be extremely useful indeed. Michael, did you really manage to use the contents of
http://opus.lingfil.uu.se/? It's a huge amount of data.
I'm under the impression that contextual search engines such as Reverso Context are constantly being improved so let’s wait and see...
[Modificato alle 2015-02-07 09:57 GMT] ▲ Collapse
| | | | | TMLookup generate content | Feb 7, 2015 |
gabritran wrote: Thanks for your comments. I’ve just downloaded TMlookup which can be extremely useful indeed. Michael, did you really manage to use the contents of http://opus.lingfil.uu.se/? It's a huge amount of data.
Here's a newer version of TMLookup with a couple of new features and minor bug fixes: https://dl.dropboxusercontent.com/u/16377950/TMLookup_1.25_win.zip
I haven't gotten around to putting it on the website yet.
As Michael said, TMLookup can search all that you can throw at it. Everything you can get from OPUS and then some. The performance depends on what your search for. The simpler the search and the fewer the hits, the faster it is. E.g. let's say you have 10 million TUs in your DB. If you search for a single term for which there are less than 300 hits, you'll probably get your results in 0.05 sec to 1 sec. For a search term with multiple words and a few thousand hits, maybe 0.5 to 3 secs. Search for something that generates a hundred thousand hits and you'll be sitting there for a minute waiting for your search to complete. Performance obviously depends on your hardware as well, but by and large, 10 million TUs will work fine with reasonable searches.
It seems to be slower if there are search terms in both search boxes, so use the highlight option to dissect your hits instead of running bilingual searches when you can. Regex search is very slow by nature, so don't use it on large databases unless there is no other option.
Searches can be sped up by loading the database into memory, but that will only work with smaller databases (less than 3GB, probably less than 2GB).
Personally, I prefer offline databases to web-based services (which should hardly come as a surprise to those who know that I spent a significant amount of time over the last two years compiling enormous TMs and writing software that can exploit them). If you have the data on your own computer, you can usually search faster and more conveniently, and you have more control over the process (including what exactly goes into your database and what tool you use to search it). Of course some types of data may not be available for download so you may be stuck using a web service.
| | | | | |
I use Glosbe but I don't think it's very good because if the word you are looking up is not a very common one, you get zero results. I prefer Sensagent.
| | |
|
|
|
Reverso helped me a lot with my English-Russian-English translations.
[Edited at 2016-09-04 07:54 GMT]
| | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Contextual search engines | 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! » |
| | Protemos translation business management system | Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!
The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
