I have historically found Google's translations (to English; I'm not
competent to evaluate any translations away from English) to be
surprisingly good. I know (on-line) someone who works for Google (near me
in meatspace as it so happens), and was discussing this with xir.

What I found out is that Google leverages their _search_ technology for
doing translations: They look for matching phrases in the two languages
where they've indexed what is known to be "the same" document previously
translated-and-verified by humans, and use the results from that before
going down to individual words. You, the searcher, are given the
opportunity to "improve a translation", which results in your preferred
translation for a phrase being weighted slightly "heavier" for future
translations involving those words and phrases. As a result, you'll get
generally better results translating an entire article than you will if you
just pick out a few short phrases.

In some cases, where there are translations of "the same" document in
multiple languages that include a few "key" languages like Esperanto and/or
Lojban, they'll use that "key" language as a "bridge" to find the phrase or
word. Bridge languages are chosen because they have the lowest levels of
ambiguity in the absense of additional contextual material. One might
suppose that if Orwell had fully defined Newspeak, it would be suitable as
a bridge language for texts and ideas that were acceptably orthodox under
Ingsoc (since everything else would be simply crimethink).

On Tue, 25 May 2021 12:54:35 +0100, Timothy Collinson
<xxxxxx@port.ac.uk> wrote:

>.. I keep meaning to mention here, something I found at the weekend.
>
>Dobry Den!
>
>You may be aware, I've probably muttered about it here before, that very
>occasionally I do a search for Czech language Traveller material.  A bit
>random really but as I've learned a fair bit, why not combine interests!
>
>Such searches have't yet produced anything beyond short paragraphs *about*
>Traveller, usually on store web pages selling (an English) core book or
>some such.
>
>However, this weekend I had started thinking about constructing a
>vocabulary list for Traveller terms in Czech and did some searching.  Thus,
>I stumbled across:
>https://cs.xcv.wiki/wiki/Traveller_(role-playing_game)
>which is as extensive a thing I've ever found and gives one or four of the
>keywords I was interested in.
>
>I was very pleased to find that I could understand large swathes of it.
>Mainly because I know what it's talking about so well!
>
>Better yet, it is a translation of a page I know well:
>https://en.wikipedia.org/wiki/Traveller_%28role-playing_game%29
>
>so I didn't even have to whack it though Google Translate to be able to
>produce a side-by-side (parallel text) document for my own enjoyment and
>word learning.
>
>However, the ease (well, for certain values of 'ease') with which I read it
>and it's closeness to the original, make me wonder if it hasn't been
>translated so much as whacked through Google Translate to create the
>webpage.  Unfortunately my Czech isn't good enough to be able to tell.
>
>(Of course, it's possible that there's a halfway house between that and
>someone having translated it, and that is that someone put it through
>Google Translate and then 'tidied it up'.  A bit like I did with that Rue
>nee Curios adventure for Amber Zone a while back that started life as a
>relatively incoherent bit of AI generated text.)
>
>So if anyone speaks, well - reads, Czech well enough to be able to confirm
>how good/bad the actual Czech is, I would be more than interested.
>
>And of course, if anyone does know of any Czech Traveller
>rules/adventures/etc that I've missed, I'd be doubly interested.  It's
>quite interesting as a thought experiment to consider what words you might
>search for that would, in fact, locate such material.  I mean you won't
>find my own 'Spindrift' by searching for <Traveller> and <English> for
>example.  I found the above by searching for <treti imperium> and
><cestovatel> but I can probably do better when I get some time.
>
>And if said person is interested in collaborating on a keyword list, I've
>made a start (with the help of Google Translate).  I can spot a few
>infelicities of the automation but really need a native speaker to get from
>'ball park' to 'this would work'.
>
>tc
>
>(And yes, I was surprised and not a little delighted to get to the bottom
>of the page and find I was a couple of the references!)
>-----
>The Traveller Mailing List
>Archives at http://archives.simplelists.com/tml
>Report problems to xxxxxx@simplelists.com
>To unsubscribe from this list please go to
>http://www.simplelists.com/confirm.php?u=wfcoddcMsNO4BxJM3qWTs4LGDbHIt4XD

®Traveller is a registered trademark of
Far Future Enterprises, 1977-2020. Use of
the trademark in this notice and in the
referenced materials is not intended to
infringe or devalue the trademark.

--
Jeff Zeitlin, Editor
Freelance Traveller
    The Electronic Fan-Supported Traveller® Resource
xxxxxx@freelancetraveller.com
http://www.freelancetraveller.com

Freelance Traveller extends its thanks to the following
enterprises for hosting services:

onCloud/CyberWeb Enterprises (http://www.oncloud.io)
The Traveller Downport (http://www.downport.com)