2018-03-12 16:39 GMT+09:00 Alastair Houghton via Unicode <unicode@unicode.org>:
> On 11 Mar 2018, at 21:14, Marcel Schneider via Unicode <unicode@unicode.org> 
> wrote:
>>
>> Indeed, to be fair. And for implementers, documenting themselves in English
>> may scarcely ever have much of a problem, no matter whatʼs the locale.
>
> Agreed.  Implementers will already understand English; you can’t write 
> computer software without, since almost all documentation is in English, 
> almost all computer languages are based on English, and, to be frank, a large 
> proportion of the software market is itself English speaking.  I have yet to 
> meet a software developer who didn’t speak English.

Somewhat digressing from the topic, but I'd like to make some comment
on this part as I smell a persistent myth among some, hopefully small number
of, software engineers in Anglosphere.

First, the fact that computer languages are written using English
words doesn't mean that programmers are supposed to have proportional
English knowledge. Take the word of Matz, the creator of Ruby
language: "The English skill is a super-powerful rare card (in the
career path of a Japanese engineer)!" He then continue that you should
be in keeping with most up-to-date overseas info/trend in order to be
a high-tier engineer and so on. It's far from "requirement".
http://eikaiwa.dmm.com/blog/3826/

I've also read somewhere a memoir of a middle-aged programmer who was
already into BASIC in childhood. One day he thought he'd written off a
"great" program and printed it on paper, but to his surprise, an
auntie who took a look at it immediately decoded the program and
roughly understood what it was meant to do; she knew English, and he
didn't.

Programming as such, is just like a Chinese room replaced with
English, where you sit inside a cramped room night after night,
communicating with a computer by typing in English words the bulky
reference guide teaches you. Most East Asian countries are blessed
enough with a tremendous number of translated technical publications
(e.g. O'Reilly) each year, not to mention firsthand writings in their
own languages. So the documentation is easily available if you don't
speak English the language.

Second, that English is lingua franca doesn't necessarily mean the
English spoken in the wild is. The aviation industry is another field
which employs English as the common language, but they exert utmost
effort to maintain the system working. Namely, they have a controlled
word set with semantics as disambiguated as possible, called
ASD-STE100, for technical documentation, such as maintenance manuals,
to minimize errors caused by limited English knowledge. Unicode, on
the other hand, is merely written in a free style used when English
speakers who (almost) graduated from college write to English speakers
who (almost) graduated from college. Having such level of proficiency
being a non-native speaker isn't something trivial, unless someone is
constantly in contact with English-speaking community. (And
programming community isn't contained inside English-speaking
community at all.)

That said, I agree to almost everything Alastair said after. If I have
to add one more thing, a monolingual writing is usually too tightly
coupled with the language, more than engineers may believe, even if
the writer carefully chose their words to be context-neutral. Thus
it's hard job to say no more and no less than the original text in
another language, especially when exactitude matters. It's one of the
problems prevent from fully automated translation being a thing, I
guess.

Best regards,

Yifan


2018-03-12 16:39 GMT+09:00 Alastair Houghton via Unicode <unicode@unicode.org>:
> On 11 Mar 2018, at 21:14, Marcel Schneider via Unicode <unicode@unicode.org> 
> wrote:
>>
>> Indeed, to be fair. And for implementers, documenting themselves in English
>> may scarcely ever have much of a problem, no matter whatʼs the locale.
>
> Agreed.  Implementers will already understand English; you can’t write 
> computer software without, since almost all documentation is in English, 
> almost all computer languages are based on English, and, to be frank, a large 
> proportion of the software market is itself English speaking.  I have yet to 
> meet a software developer who didn’t speak English.
>
> That’s not to say that people wouldn’t appreciate a translation of the 
> standard, but there are, as others have pointed out, obvious maintenance 
> problems, not to mention the issue that plagues some international 
> institutions, namely the fact that translations are necessarily non-canonical 
> and so those who really care about the details of the rules usually have to 
> refer to a version in a particular language (sometimes that language might be 
> French rather than English; very occasionally there are two versions 
> declared, for political reasons, to both be canonical, which is obviously 
> risky as there’s a chance they might differ subtly on some point, perhaps 
> even because of punctuation).
>
> In terms of widespread understanding of the standard, which is where I think 
> translation is perhaps more important, I’m not sure translating the actual 
> standard itself is really the way forward.  It’d be better to ensure that 
> there are reliable translations of books like Unicode Demystified or Unicode 
> Explained - or, quite possibly, other books aimed more at the general public 
> rather than the software community per se.
>
> Kind regards,
>
> Alastair.
>
> --
> http://alastairs-place.net
>
>

Reply via email to