On 5/30/2019 11:17 AM, William A Rowe Jr wrote:
On Thu, May 30, 2019 at 11:09 AM Patricia Shanahan <p...@acm.org> wrote:
Interesting. I just tried this on some of my own writing in the draft
FAQ. Translating English->Italian->Japanese->Basque->English transformed:
"Style and writing etiquette, such as how directly or indirectly one
indicates that another writer has made a mistake, also varies by
subculture and by gender. Most of us are, consciously or unconsciously,
more comfortable communicating with those who write the way we write."
into:
"Typing style and protocols are dependent on genre and genius, the way
in which other writers have directly or indirectly mistaken. We
communicate with people who write most of us comfortably or comfortably."
That was only three other languages. I have some work to do.
Japanese famously drops meaning. Words go unsaid, so there are multiple
meanings in a round trip. It's a major PITA for translation maintainers. I
wouldn't use it beyond conversational translation. Some other languages
have similarly radical differences with western language. Just to
demonstrate to yourself, try some arbitrary text simply between english,
japanese and back again to english. The failure is the tool in this case,
not the user.
This raises the question of what is a good sequence of languages for
this test.
English->Japanese->English did a bit better. It came very close on the
first sentence, keeping the essential meaning, but muddled the second
sentence:
"Style and essay etiquette, such as how directly or indirectly it
indicates that another writer made a mistake, also depends on the
subculture and gender. Most of us communicate with people who write the
way we wrote more comfortably or consciously."
---------------------------------------------------------------------
To unsubscribe, e-mail: diversity-unsubscr...@apache.org
For additional commands, e-mail: diversity-h...@apache.org
