[hgbook] Italian version available

[Giulio Piancastelli]

On Thu, Sep 17, 2009 at 6:53 AM, Stefano Tortarolo
<stefano.tortarolo at gmail.com> wrote:

> Hi Giulio,
> 10 days later... wow! Sorry for replying so late.

It's OK. I'm still reading. ;-)

>>> I think we should stick with terms already adopted, 'fusione' is weird
>>> but if svnbook uses it maybe we should use it as well.
>>
>> I don't think that being the first term to be used says anything about
>> its quality or whether it is the most apt to the job. I'd like to talk
>> about these issues rather than following a translation "just because"
>> it came before. If there is a better word, I don't think I should not
>> use it "just because" someone else used another word before.
>
> It's not a matter of quality, it's a matter of what people are used to
> hearing and using.

This is a possible, "pragmatic" in some sense, point of view. But, in
my opinion, it sets the bar too low. People are used to hear a lot of
terrible uses of the language; that's not a good reason to promote
those uses.

On the other hand, to me it seems quite odd that someone would just
like to be feed the English word, putting no effort in understanding
at least what the word means, and doing the necessary brain-work to
relate the English term (which gets used every day by issuing
commands) to the Italian meaning. I can understand, but if you decide
to turn to a book, I believe you expect the book to treat the topic
with a certain language accuracy, both native and technical. Even
technical books are not casual conversations with a co-worker, from
this point of view.

In this sense, a possible fault of the translation is not the use of
"changeset" as you pointed out in a previous message, but the use of
"hook" and "template". *These* are much harder to explain, if not by
the translator's fault in finding a good Italian term to use as a
substitute (or accepting the less-than-average terms found). Have I
mentioned that translating is hard? ;-)

> Another solution, in my opinion, could be writing a side note
> explaining how other VCSs translated the same term, and then using our
> new term.

This is a sensible advice. I have updated the relative issue (*) on
Bitbucket, so as to remember the advice when I try to resolve it.

(*) http://bitbucket.org/gpiancastelli/hgbook-it/issue/2/

However, for completeness, I'd like to note how, in paper publishing,
even the change of terms in the translation of two different editions
of the same book is not annotated at all. Case in point, Learning
Python: one edition used "descrizioni" for list comprehensions, the
following edition used "espressioni", but there's nowhere to be noted
that readers may have already got used to the previous term instead,
since that's what made "the literature" until the new edition hit the
shelves.

(On the other hand, paper publishing seems to be much more permissive
about "technical" translations sometimes -- or lack thereof -- perhaps
following the "pragmatic" approach you outlined at the beginning.)

> I agree with you, but we need to take in consideration the fact that
> lots (most?) of our users have to deal with the English terms as well,
> that's why I still believe that keeping the English terms, along with
> the Italian ones, would be the best solution. What I'm suggesting here
> is to use Italian terms, without hiding completely the English ones.
> This could be a fair compromise, I guess.

I'm considering it, as the opened issue shows. However, let me say
(again) that the English term is not completely hidden, because it's
still used *on the command line*, where you can't use anything else.
There's the command, hg merge; and then there's a chapter, in Italian,
that talks about that command, referencing it as "hg merge" of course,
and explaining in Italian what that command does. So, since the book
language is Italian, and the explanation of the command's
functionality is in Italian, the translator uses "unire" instead of
"merge" in the book's text.

>> Ah, yes. I have looked at the Italian translation of Mercurial while
>> translating the code snippets. I used some parts of it, but I changed
>> (in the book) the parts that I didn't like.
>
> The current translation is far from being perfect, that's why some
> time ago I asked (in vain) for help on this ml.

I think software translation is hard, because you see most of the
messages out of context; also, it's probably way less attractive than
translating books, articles, or general-purpose documentation.

> Btw, I like 'aggiungo' and 'etichetta'.

Yup. :-)

If you plan to still work on Mercurial's translation, you may check
the book for other translations of more messages. There are some that
got their very first translation in the book, because the i18n file
for it was missing them. I can't really remember the details, but I
think that a comparison is in order for the (next?) person who'll work
on the software translation.

> As long as we define somewhere (in that case in hg help update) that
> 'update' means 'aggiornare', I think it's absolutely correct to say
> 'Aggiorno a ...'.

So be it. The first line in Mercurial 1.3.1 "hg help update" starts as
"Update the repository's working directory to the specified revision"
and I can't possibly see how that "Update" could *not* end up being
translated as "Aggiorna".

> The main point in translating a command line-based software like
> mercurial is to ensure that new users will be able to use a specific
> command correctly, even though their English skills are very low. In
> order to achieve this, in my opinion it's very important to keep a
> strong relation between the Italian translation and the actual hg
> command.

To me, this is achieved by explaining what the command does, as
clearly as possible, *in Italian*.

Regards,
Giulio Piancastelli