[Gaim-i18n] Re: Bad language in Gaim

[Bjoern V. <bj...@cs...>]

Good morning all!

Ethan Blanton wrote:
> If we're keeping score (which I guess we aren't, because the scores are=
 meaningless), it took me about 30 seconds to answer it completely and co=
rrectly when I read your first email.
>  =20
But I don't wanted to make a survey with multiple choice answers. I=20
wanted to make a test which shows you (especially the developers) how=20
much work it can be to translate i18n strings in Gaim. Translating is=20
not like filling out a multiple choice test. Translating with Gettext is =

instead:

   1. search the an untranslated or fuzzy string (.po file editors make
      this with a key or menu item)
   2. decide, if it's possible to translate the string directly or with
      using the context
          * directly: translate it
          * with context: display the source code, search for comments
            there (you can skip this task because in the majority of
            cases there aren't comments!), search for explanation in the
            web, in Google, discuss here, in #gaim etc.
   3. spell check the translated string
   4. go to the next string (step 1)
   5. check the whole translation (syntax checks, compilation of Gaim,
      usage tests, test key strokes etc.)

Ethan, do you ever translated with Gettext tools?

I think a complete translation takes much time. The German Gaim 1.5.0 to =

2.0.0 translation caused some days work for two translators (Jochen=20
Kemnade and me). I think more context information (i18n comments in the=20
source) and some string changes for bad and too technical strings would=20
speed up the translations and would improve our motivation.

To see the current motivation of us translators look at:=20
http://gaim.sourceforge.net/i18n/

One problem is that we don't find new translators for the outdated=20
translations.

> > Master solution:
> > >1) What means "you can see the butterflies mating" in=20
> > >src/gtkaccount.c:1297?
> > Best answer:
> >=20
> >    Richard wrote: "It's in a callback function that's registered as a=

> >    pop-up handler for the port entry field in proxy options.
> >    Right-click on that field and you'll see the string. I'd assume
> >    easter egg. (NOTE: I had never seen this code before I looked at i=
t
> >    just now.)"
> >=20
> >    Unfortunately I was not able to test it. Right-clicking on the por=
t
> >    entry field shows a popup menu. But this popup menu only contains
> >    GTK+ things (cut, copy, paste, Input methods etc.), nothing about
> >    butterflies. May be it's "dead code" :-(
>
> It's not, it works fine on my machine.  I don't know why yours doesn't.=

>  =20
Ambrose Li wrote me an instruction (thanks Ambrose Li):

: You need to -
:=20
: - go to the Accounts window,
: - add/modify an account
: - make sure that proxy setting is "HTTP" (or maybe something else but
:   it has to has a "Proxy Port")
: - expand the proxy settings, then
: - right-click inside the text field of the "Proxy Port".


This works (I can see the butterflies). :-)

> > >3) What is a "TOC host" in src/protocols/toc/toc.c?
> > Best Answer:
> >=20
> >    Ambrose Li said: "A TOC server. TOC is an AIM protocol. A host is =
a
> >    computer. Literally a computer talking the TOC protocol, but the
> >    context requires it to be a server, and in the DOS/Windows world
> >    (since this is AIM) the word "host" traditionally only refers to
> >    servers."
>
> So, what do you propose we change here?  TOC is a technical term for
> which any euphemism is only going to make things worse, not better.
>  =20
I didn't say that we should change all these strings. A comment for the=20
translators would be good here. The "TOC" plugin is not compiled by=20
default. So it's nearly useless to have to translate the TOC strings.

> > >4) What does "SSL Handshake Failed" in src/protocols/irc/irc.c:432 a=
nd=20
> > >src/protocols/jabber/jabber.c:451 mean?
> > Best Answer:
> >=20
> >    Richard wrote: "The process of negotiating an SSL connection faile=
d."
> >    Read this long SSL handshake description from Microsoft
> >    http://support.microsoft.com/kb/q257591/
> >    (Netscape made the SSL protocol and more detailed information can =
be
> >    found there.)
>
> Again, what would you propose we change here?  The Windows/OSX/GNOME
> solution might be "just tell the user the connection failed, they
> don't need to know why", but this is not acceptable to me.
> "Handshake" is a technical term which is being used correctly.  It is
> not clear to me that there is a better translation for this, either.
>  =20
I repeat: I didn't say that we should change all these strings. A=20
comment for the translators would be good here.

> > >5) What is a "Channel HMAC" in src/protocols/silc/chat.c?
> > Best Answer:
> >=20
> >    Ambrose Li wrote: "Short answer (looking at my po file): Some kind=
 of=20
> >    user authentication token
> >    for the channel"
> >    Richard wrote: "... Looking at the source, it's immediately clear =
(given=20
> >    Channel Topic
> >    is right there) that this is talking about a "channel" as in "chat=
"."
> >=20
> >    I personally think that the channel HMAC is a unique id for a chat=
=20
> >    channel. It's possible to calculate such an unique id with the HMA=
C-MD5=20
> >    or HZMAC-SHA1 digest algorithms.
>
> That's not at all the case ... HMAC is, once again, a technical term
> with a specific meaning.  (Basically, a keyed hash function, such that
> two parties who have agreed on a shared secret can authenticate a
> chunk of data of arbitrary size by calculating a MAC across it and
> some function of the secret.)  A third time, removing this bit of text
> removes important information that clueful users may be able to make
> sense of.  What would you suggest?
>  =20
Thanks for explanation. I can not suggest anything here now. May be we=20
should look at other SILC clients.

Looking at other IM projects and their translation can be interesting=20
for us. Kopete for instance also has 2765 i18n strings (~2578 Gaim). So=20
the amount of strings is not the main special problem in Gaim-i18n. You=20
may not want to hear it, but Gaim's i18n strings are much more technical =

compared with Kopete. I also searched for difficult or deep-technical=20
i18n strings in Kopete but I didn't found much good examples. To be=20
honest, i18n comments are also seldom in Kopete - not good. Look yourself=
:

http://websvn.kde.org/trunk/l10n/en_GB/messages/kdenetwork/kopete.po?rev=3D=
519677&view=3Dmarkup=20
(British English)

> I think we're wasting a lot of time here to come to the conclusion that=
 some strings are just hard to translate.  :-P
>  =20
I can partially agree with your conclusion. But what is your suggested=20
action?

Do nothing (no additional context information - comments in the source,=20
no string changes etc.)? If yes, not Syd Logan is the lazy bum, but=20
Ethan Blanton. You can also call people "conservative" which are not=20
willing to make changes. ;-)

Read again what Clytie Siddall, Ambrose Li, Isriya Paireepairit, Vincas=20
=C4=8Ci=C5=BEi=C5=ABnas and me sgested to change.

Greetings, Bj=C3=B6rn