Re: Subject: [RFC] docs/zh_TW: defining a maintainable scope for the Traditional Chinese translation

[葉宸佑]

> If the current translation quality is acceptable and it's not
> troublesome to move it, I think we can just move it. On the contrary,
> if the translation is significantly behind, there is no need to go
> through all the trouble, just delete them.

I understand the appeal, and for some files I think you are right. But
I would like to separate two things that are getting mixed together:
relocating a file is cheap and mechanical, while deciding to delete one
is neither, and the two do not have to be decided at the same time.

The four relocations in section 2 are the cheap case: the files are
reachable, they are not the worst offenders, and moving them is a
rename plus an :Original: fix. I would rather just do that and revisit
their content later.

For deletion, my hesitation is that it is the one decision here we
cannot walk back. Marking a file unmaintained can be undone the moment
someone turns up to look after it; deleting it means the next person
starts from nothing. Right now zh_TW has one native speaker, so
"re-translate instead" is only more efficient if the re-translation
actually happens.

So I would suggest: deletion by exception rather than by rule. Rather
than "significantly behind means delete", we name the specific files we
think are beyond saving and say why. If you would like to propose a
list, I am happy to look at each one -- I suspect we would agree on
more of them than this exchange makes it sound.

> Tricky?

The six worst are mostly two clusters: admin-guide/mm/damon (5 files,
55 commits between them) and arch/arm64. Those are exactly the files I
would expect a Taiwanese reader to skip in favour of the English, so
they are strong candidates for whatever we decide "unmaintained" means
-- and, if you want to make the case, for deletion.

> However, the former might be clearer and more understandable for the
> readers.

Agreed, the notice belongs in the file. I will do it that way.

> And after process/, I'd like to do that part "Working with the
> development community".

Good -- that is the section of the index that matters most after
process/ itself.

Dongliang: Weijie quoted you above on re-translation being more
efficient. That was in a different context, so I would rather not
assume it carries over -- what is your view on delete-versus-mark for
the files outside process/?

Thanks,
Chen-Yu Yeh