Re: Translating Emacs manuals is of strategic importance

[Stefan Kangas]

Po Lu <luangruo@yahoo.com> writes:

> The danger in accepting such translations is not a potential shortage of
> contributors _willing_ to take up their upkeep whom we might attract by
> distributing them under our own seal of approval, it is the motivation
> and ability of such contributors (many of whom have offered us their
> services) to carry such tasks to completion.  Without this we will land
> ourselves with a collection of incomplete manuals and accountability for
> their faults, which will place yet more strain on an already-burdened
> group of contributors, none of whom, I trust, can dedicate enough of
> their time to maintaining such manuals in just the languages widely
> spoken in technical circles (i.e. German, French, Russian, Hebrew), let
> alone each and every language that might see interest expressed in being
> the target of a translation of the Emacs manual, or that our users
> speak.

If I understand you correctly, your concern is that we will be
responsible for the translations.  To me, that would imply that we will
have to spend significant time maintaining them.  But I don't think it
will be like that in practice.  Instead, we have to make it clear that
the translation teams (often: interested individuals) for the respective
languages will be responsible for the translations.

Furthermore, only the English manual can be considered canonical for the
foreseeable future, since we cannot proofread translations like we do
with the English one.  It therefore makes a lot of sense to add suitable
disclaimers (dates, version numbers, etc.) to the translated manuals.

We could perhaps consider having several "tiers" of manuals: one
"nursery" for those are not yet meaningfully ready for distribution, one
"attic" for those that are no longer reasonably maintained, and one for
those that we actually do consider up to scratch.  The decision for what
goes where could be based on a dialogue between the maintainers and the
translators of various manuals.

> There is nothing to suggest that users cannot locate an outdated or
> incomplete manual over the Internet.  I don't dispute that they serve a
> worthwhile (if not crucial) purpose, my reservations are that we should
> not take _responsibility_ for them while being fully aware of factors
> that make for a very real chance that they won't deliver.  Then if the
> time comes when interest in them does increase such that it becomes
> sound for us to assume that responsibility, we can always revisit any
> decisions that might have been made up to that time.

I believe that the moment will _never_ present itself when we are sent a
patch containing a fully translated, high-quality and up-to-date manual
backed with an active team of translators, copyright assigned and ready
to be merged.  We have to build things up over time, and preferably in
communication with the translators themselves.

>From the outside, it is likely that the impression is that we are rather
uninterested in translating Emacs to other languages.  The fact that
there are fully translated materials distributed elsewhere speaks to
that.  If we double down on a workflow where we are happy to see manuals
translated, but do not want to distribute them, then that impression
will be cemented.

On the other hand, if we already had the infrastructure in place for
translations, it would be more clear that we are, in fact, interested.
This would be true even if we had only a small handful of manuals or
work-in-progress translations there: at least we would have a framework.

Our obviously central position in the Emacs world means that this in
itself could help promote these efforts, and hopefully also encourage
volunteers to take on translation work.

> Being realistic in our estimate of our own capabilities is not
> defeatism, nor is requesting that interested participants stage their
> work separately discouraging, which incidentally won't preclude
> obtaining copyright assignment for such work should it ever come to
> that.

It was not my intention to call your stance defeatist, but rather to
warn against it, since the current situation is neither fixed nor
impossible to change.  I apologize if I was being unclear.

I'm interested in hearing what people think about this, but if
Jean-Christophe and others intend to set up the relevant infrastructure
in emacs.git, I think we should try to encourage that.

I do think it is important that we are careful to set things up in such
a way that the translating work isn't too disruptive for regular Emacs
development.  Personally, I also would like to never have to see the
French translations in the course of my usage of Emacs: only users that
ask for it specifically, or use a French locale should see them.