Re: [gnome-cyr] Re: gtksourceview/po/ru.po: =?utf-8?b?0K3QudGH0KLQnNCb?=? WTF?

[danilo gnome org (=?utf-8?b?0JTQsNC90LjQu9C+?==?utf-8?b?INCo0LXQs9Cw0L0=?=)]

Здраво Микхаил,

Long answer follows.

Данас у 1:11, Mikhail Zabaluev написа:

> Hello Данило,
>
> On Sat, Apr 17, 2004 at 11:06:02PM +0200, Данило Шеган wrote:
>>
>> >> That puts the issue out in the open: if you don't go for
>> >> transcription, you lose out completely, because newcomers won't know a
>> >> thing you're talking about, and they even won't know how to read it or
>> >> call it (is that ???н-т-т-р??? or ???е??ч-ти-ти-ар???).  I.e. they can't read
>> >> it, and they can't learn it.
>> >> 
>> > Yes, it's a key point. Newcomers (or any other people) don't have to
>> > know anything than their native language and alphabet.
>
> Yeah, just as they can't learn anything about that wacky foreign program
> "Microsoft Word" that boasts its English title on them. Oh wait... nevermind.

I don't follow.  I'm talking about them knowing how to read it.  So,
they see "HTTP" in a translated program, for instance in setting a
HTTP proxy.  Let me use Serbian example here, to demonstrate my
point:

  Изаберите HTTP посредник: [       ]  порт: [8080]

Now, they call their friend to ask for help, who's proficient with
computers, and ask him what?

"What's this 'n-t-t-r' thing?  What do I put there?"  The friend
becomes confused, doesn't know what's "n-t-t-r", and says: "I dunno,
I'll have to come over to help you there".  Time lost, productivity
lost, everyone loses.

When they see "Microsoft Word", they see the letters that are not
Cyrillic: w, r, f, d.  It's quite easy for everyone to deduce that
it's not their native language.  So, at least they're prepared (as
English proverb goes: "Forewarned is forearmed").

Besides, I think Microsoft should use translated names as well.  But,
they're not free software, so I can't do anything about it (not that I
care about non-free software).  Gnome *is* free software, and we can
do the best thing with it.

> Face it, newbies don't know a thing about "HTML" and they don't
> care, until they need it and/or read about it somewhere. In Russia,
> this means they read the name in Latin. Dmitry says it's somehow
> wrong, but it's the fact that can't be helped in any other way than
> a totalitarian regime (and we don't want that again). For proof,
> try any Russian search engine.

Just the way to prove your point.  You don't read my mails, and try
to fight some totalirism that isn't even there.  Yeah, I understand,
I come across it all the time in Serbia as well (everyone seems to
think that there's one true solution, while I constantly claim that
there isn't; sometimes you must use Latin script, but in most cases,
you need not).  Please excuse me if I over-generalized wrongly.

You need to re-read my mail.  I'm trying to show you that you can't
read the text supposedly "translated", unless you know it in front.
And if you know it already, why would you need any text at all?

Again, tell me what am I typing right now: HТTP.  Be careful, I may
have used either Cyrillic, or Latin.  So, read it out loud, in a
correct way (hint: it's Latin H, Cyrillic T, Latin T, Latin P).

It's a very practical problem, with several practical solutions.  You
may say that it's a non-issue as much as you wish, but it's easily
provable that it isn't.  Just by looking at it, did you know what
letters were Cyrillic and which weren't?  Of course not.

(Note that you were talking about HTML, where such issues don't come
up, and I was talking about HTTP.)

>> I agree with your point.  But I was aiming at *others* who disagree,
>> and I was aiming at them with extremely practical point that even if
>> they *did* know the Latin alphabet, there's no way for them to know how
>> to read HTTP if they don't already know about it and what it
>> represents.
>
> That, I think, is a problem better solved by introductory books
> rather than space-constrained UI items of a widely used computer
> application.

Complete nonsense.  I need an introductory book on HTTP in order to
learn how to read a single word "HTTP"?  That's so much nonsense that
I can't even believe I'm reading it.  What's text for if not for
reading?  Why don't we switch to using random images instead?  We
remember text by memorising it read in our native language (it happens
with foreign languages as well, once we become sufficiently good with
them).  We don't remember glyphs along with the script ("ok, this was
ha-te-te-pe in Latin").

>> I'm talking about glyphs looking the same.  HTTP looks exactly the
>> same as НТТР (eg. ???Нови Труд Тво??е Руси??е??? :).  Not many programs
>> provide features such as M-x describe-char Emacs does, which might be
>> used to differentiate between the two.
>
> This justifies what? Does any, say, German l10n provide
> pronuncation hints for "HTTP" in order to prevent some naive Germans
> to refer to that as "ash-te-te-pe"? Does any random application have
> to include a mini translation assistant in its day-to-day UI?

German uses Latin script, and doesn't have this problem.  They don't
need to read it as "eich-tee-tee-pee", because they can read it
without a problem, and *unambigously*.  Put "HTTP" in front a random
sample of people in Russia, and let them read it.  I believe you'll
see what I mean.

It wouldn't be a problem if we read this as "ха-те-те-пе", but the
problem is because many won't read it that way, but more likely as
"н-т-т-р".

When you want to do comparisons, don't compare languages using the
same script (such as German and English), but rather, try with
English and Russian.  When was the last time you saw Russian word
untranscribed/untransliterated in English book/newspaper, except as
an example of "a Russian text" or something?

[FWIW, I have yet to come across a Serbian fellow who pronounces FTP
as "eff-tee-pee"; everybody pronounces it like ФТП]

>> Just ask a 10-year old kid to *read* this: ???C++??? or ???Язык C???.  Ok,
>> everybody who is aware of ???C programming language??? (C as in ???see???)
>> won't have a problem here, but this will introduce a lot of problems
>> for everybody using a translation.
>
> Let's get back to the matter at hand.
> Please provide any realistic example where the said kid _needs_ the C++
> highlighting mode in Gedit before he/she _knows_ about the language.

I used a kid only as the most probable case of someone not used to
reading "C" as "си", or whatever.  You may replace it with any adult
not commonly using computers.  Yet, I tend to treat kids as best
non-corrupted language lawyers.  I want to make my translations
understandable and readable by kids as well.

Yet, what if a kid sees "C++" there, and wants to learn more about
it.  It goes to the library looking for "S++", asks friends about it,
and nobody knows anything.  What then?  Why shouldn't it be easier?
In Serbia, if I write "Це++" there, whoever the kid might ask, be it
an expert, or not, will know how to direct the kid: "go buy
Stroustrup's book", or something.

And what do we lose with this?  That's the question to be asked,
since we gain nothing except problems with keeping the Latin name.

>> If you were writing introductory book for C, you would inevitably
>> have to write a couple paragraphs explaining that what everybody
>> thinks of as С (S, as in ???Слово???) is actually ???See??? or
>> ???Си???.
>
> It may come off as a surprise, but there is such practice in books
> (although it usually won't take two paragraphs). And the practice is
> good. For books.

Yeah right.  And this is in what ways better (good?) than not having
to explain it at all?  Or perhaps having a paragraph or two explaining
that "Це++" is written as "C++" in English?  What's the difference?
I see only the gain: readers know right away how to pronounce it, and
don't have to bother.

So, this "good" practice sucks no less than explaining the original
English name itself.  Please, try to prove otherwise.

>> The simple proof of all this is ТеХ itself.  How many times did you
>> see it called ???текс????  I saw it too many times.
>
> Erm... And for native English readers this "mistake" is less common how
> exactly?

Not less common at all.  That's what I'm talking about.  Knuth made
it harder for *native* English readers with using Greek alphabet in
the program name.  *You* are making it harder for *native* Russian
readers with using Latin alphabet all around.  You're simply agreeing
with me.  Though, I'm sorry to hear that you like making it harder
for your fellow country men.

I do respect Knuth for all his efforts and genius, but I think this
decision of him sucks big-time (and I simply love TeX as the program).

For a more tolerant approach, look at Содиподи (the web page of it
contained the name written in Cyrillic).  Yet, they didn't insist on
that being used everywhere, instead, they recommended "Sodipodi" to
Latin-based scripts.  That's what I appreciate, and recommend myself.

> I believe that depends on the definition of 'unambiguous'.
> Try to search for 'броузер' and 'браузер' in any Russian search engine
> (http://ya.ru) to see what I'm talking about.

"Browse" in English means to look around without a goal.  "Браузер"
or "броузер" are unstandardised transcriptions.  They're evil, I
agree.  But if you translate it (in Serbian, Разгледач, verb,
разгледати), you have no problems.  You can also translate it
differently.  Eg. we use "Web reader" (Читач веба) instead of "Web
browser", "File browser" (Разгледач датотека), etc.

I also agree that such crappy transcriptions and translations have
caused many people to dislike all translations in the first place.
That's why some "totalitarian" people who translate/transcribe
everything do harm to those who'd do it properly.


Sorry for being somewhat over the line, but with many similar
"arguments" coming from the Serbian side as well, I get easily
excited ;)  I have nothing personal against you, so please excuse
some harsh words, and answer only to arguments (I tried to make
it all backed up with arguments, but it all depends on the reader as
well, so what seems as an argument to me may seem like a hoax to you :).

Cheers,
Данило