The generics of translation

Matt Mackall mpm at selenic.com
Mon Jul 2 13:57:30 CDT 2012


On Mon, 2012-07-02 at 18:34 +0200, Adrian Buehlmann wrote:
> On 2012-07-02 17:58, Martin Schröder wrote:
> > Am 02.07.2012 17:43, schrieb Thomas Arendsen Hein:
> >> "vereinen" klingt hier für mich leicht unpassend, das Substantiv
> >> wäre ohnehin "Vereinigung" (nicht "Vereinung"), daher mein
> >> Vorschlag: "vereinigen" und "Vereinigung"
> > 
> > Beide haben ziemlich genau die gleiche Bedeutung. "Die Vereinung" ist
> > tatsächlich sehr ungebräuchlich. "Das Vereinen" wiederum nicht.
> > 
> > Ich bin weiterhin für "vereinen" oder dafür, "zusammenführen" zu behalten.
> 
> I didn't really want to take part in your decision making processes, so
> please just take my posting as an idea. After all, there was a call for
> ideas... :-)
> 
> And perhaps this should be taken off-list, if the discussion continues
> in German. Matt doesn't seem to like discussions in non-English on this
> list.

Ugh. Moving this discussion off-list is _exactly_ counter to my point. 
Instead, you're practically painting me as someone who insists on
English just because he arbitrarily hates foreigners or something.
Thanks!

If you're writing in a language that only a subset of Mercurial
translators can read, then every group of translators is just going to
have discussions about the same generic issues:

- should we copy jargon used by other SCMs or perfect our own?
- should we translate metaphorically or technically?
- which of the above is more important?
- etc.

EVERY word choice decision, even though it looks like it's just
language-specific trivia on the surface, is going to be rooted in more
generic principles. Formulating those principles 11 times each in 11
languages is the wrong process. And just choosing words based on what
feels right at a given moment is also the wrong process. Instead, we
should actively look for the generic issues, formulate one set of rules,
and apply them consistently within and across languages, and iteratively
improve the rules as we discover corner cases.

As a practical matter, that simply can't happen if we discuss these
issues in a language other than English.

Here's an example proposal:

--------

When translating a Mercurial-specific term like "pull":

- distinguish between usage as a literal command name (untranslated) and
as jargon (translated)
- translate the components of the metaphor (ie "push and pull a stack of
papers") to a natural set in the target language so that people can
benefit from the metaphor
- if there's no obvious translation of the metaphor, consider using a
transliteration[1] 
- terms that are likely to be translated back to the original English
term aid in cross-language communication (eg German<->Japanese)
- be wary of translation of Mercurial-specific terms to related jargon,
such as "pull"/"clone" -> <download>[2] as this may give users an
incomplete or wrong impression and result in confusing bug reports:

  Q: I did a download and my changes were missing
  A: What do you mean by download??

When translating a non-Mercurial-specific term like "patch":

- look for a widely-established precedent as used by other similar tools
for the technical term "patch"
- if multiple translations are well-established, favor the one that more
closely matches the metaphor (ie "patch a tire")
- otherwise, if a transliteration is well-established, favor it
- if no translations are well-established, translate the metaphor as
above

When translating pure jargon like "grep":

- favor transliteration

Keep a glossary at the head of the .po file

Focus on tracking the stable branch

[1] Note that "transliteration" is generally the same as "exact copy"
for languages using a Latin alphabet
[2] Note that if <download> was the optimal translation of "pull", the
English term probably would have been "download".

-----

Discuss, please.

-- 
Mathematics is the supreme nostalgia of our time.




More information about the Mercurial-devel mailing list