Re: Emacs i18n

[Alexandre Garreau]

Le jeudi 22 juillet 2021, 19:32:49 CEST Yuri Khan a écrit :
> On Wed, 21 Jul 2021 at 15:55, Emanuel Berg via Users list for the GNU
> 
> Emacs text editor <help-gnu-emacs@gnu.org> wrote:
> > Translating huge books is just an immense wasted effort that
> > also have several negative effects that I've mentioned
> > already. For example this book
> > 
> > @book{introduction-to-algorithms,
> > 
> >   author     = {Cormen and Leiserson and Rivest and Stein},
> >   edition    = {2nd edition},
> >   isbn       = {0-268-53196-8},
> >   publisher  = {MIT Press},
> >   title      = {Introduction to Algorithms},
> >   year       = {2001}
> > 
> > }
> > 
> > It is 1184 pages!
> > 
> > And that format (book heft) isn't uncommon!
> > 
> > Are you going to translate this to French because some French
> > guys aren't good enough English readers?
> 
> [This one?][1] It has been translated into Russian and it is being
> sold at ~73 USD. Which is a lot for a book, and still it seems there
> is demand for it.
> 
> (As for me, if I personally wanted to own a copy of that book, I could
> probably buy it in English on Amazon for cheaper. I’ve seen enough
> books whose translations were inaccurate, inelegant or any combination
> of the above.)

Same for me, I regularely noticed that.  Yet I own french translations of 
some important books, such as this one (and I got it from another user of 
emacs, btw, met on emacs french irc channel), and first edition sicp (which 
was translated in my university).

But as nothing is perfect, absolute and rarely objective, there can be 
stuff you can sometimes like in old editions, for instance the first edition 
of sicp contains some additional exercises (some of which are unsolvable, 
yet amusing to try to solve), and I dislike the “object.notation” of third 
ed. of cormen, which does not appear in french edition which is from 
second ed.

More generally, if books used a tailor-made translation infrastructure, 
such as GNUnation, or anything allowing to translate only changed/added 
paragraphs, and only portion of them, that would make translation of new 
editions much easier… but I’m pretty sure it’s not done because (1) 
copyright unhelps, (2) it’s obviously already profitable anyway, even being 
horribly expensive, (3) the interest in it and profitability, especially in 
smaller audiance, may be a lot more chaotic and variable, so it may not 
remain always as interesting to translate.

But more generally a few errors and (correctable!) unaccuracies are worth 
the increased easiness reading.

Anyway *that* is the core, DH0 [0] point of it.  And you may attack it 
pretty efficiently (but maybe you don’t want that but just hope everybody 
feels like you more generally dismissing everything else) on its *own* 
playground (the political field) in this way: having a wide and rich enough 
linguistic sphere for benefiting from it is a privilege, that currently 
mostly benefit the most powerful imperialist nations (france, russia, 
china, japan, hispanophone and lusophone countries, arabic countries), 
dismissing most of other languages

So it would not even be efficient (*at all*) at protecting languages from 
extinction (beside that side effect that english would spread less fast and 
extinguish them more slowly).

A counter-argument, which, contrarily to the uneasiness of reading a 
foreign language, may by chance not be scientifically studied upon, is that 
it’s still easier to learn a similar language, from the same family, 
rather than just english.  Except Japan which is an isolate, and Iberic 
ones which are totally unrelated to american languages (well except 
vocabulary nowadays…), there are many arabic (maybe semitic in general), 
chinese, slavic, and latin languages out there… yet some very important 
families are missing from that imperialist game, such as austronesian, and 
many african, american, indian (and asian in general) ones.

Actually, if I wanted to approach from adressing your core, DH0 [0] point, 
of shared culture, I may say that centralization (except maybe in some 
“utopia” where all languages are extinguished) is not the most efficient way 
of doing it… just as broadcasting everything, in a horizontal, 
decentralized and most importantly *indirect* network such as the internet 
(or more generally human relations), is less efficient as *multicasting* in 
a tree-shaped way: translating most important stuff to main languages of 
most important linguistic families, that would then be learned by most 
people may be easier, as for the fact that most of the human population 
still is *ignorant* (it’s not necessarily elitism, it’s just a fact, not a 
researched goal or an immutable truth) about science and technology, we 
could maybe in this way approach more the goal of making them *aware* of 
what’s important in it.

Most important knowledge is not the one that change most often.  Classical 
mechanics are still used, and it’s good that books on it are translated in 
most of languages (in it could be in *all* languages that’d be great).  

I’m *sure* translating Knuth would be wonderful, once he’s dead (I hope he 
authorizes it, but under american copyright law maybe he’s not even 
allowed to free his own work after publishing so well…).

[0] http://www.paulgraham.com/disagree.html