Restarting W3C Volunteer Translation Tracking

Dear Translators,

As many of you know, back in January 2017 [1], W3C had to stop the
maintenance our translation database where volunteer translations of W3C
specifications and documents had been tracked.

We are grateful to see this has not prevented this great community from
developing and publishing translations of W3C Recommendations and other
W3C documents.

Having had the time to review the limitations and gaps of our previous
tool and process in this space, we would like to offer a new approach to
the management of volunteer translations, which would offer stronger
support to volunteer translators and an improved ability to W3C to
promote their translations.

This new proposed approach is described in this message; at this stage,
we are looking for input and feedback on this proposal (with some
specific questions at the end of this mail, but all feedback is
welcome), as well as for a few volunteer translators who would be
willing to work more directly with us as we experiment with and set up
the infrastructure for this new system.


The main orientations of the new system would be:
* making it based on github - all W3C groups develop their
specifications on github nowadays, and anchoring the translation work on
the same platform provides opportunities for better alignment with the
specification process, as well as a natural platform for collaborating
on the development and maintenance of translations

* setting up some minimal level of quality reviews on translations - the
volunteer translation program has in the past been abused by spammers,
which among thing meant that the value of the otherwise excellent work
done by most of our translators would be diluted by some less reliable work

* having W3C automatically publish a copy of the translated documents
under its own control - this is to ensure that the useful work done by
translators remain available to the community even after the said
translators are no longer interested or in a position to keep it
available on their own site


In more details (although a lot of these are expected to evolve as we
build and experiment), the new workflow would roughly be:

* a translator signals their intent to translate by raising an issue in
a well-defined github repository

* as an opt-in, other translators of that language get notified of that
intent (and possibly, signal their intent to help with the effort)

* once the translator(s) finishes the translation, they bring the
translated document via a pull request to a well-defined repo

* a team of identified reviewers for that language get notified and are
expected to make a high-level review of the document to ensure a minimum
level of quality (e.g. avoid spam, low-quality automated translation);
optionally, these reviewers can provide more detailed feedback as
non-blocking requests for enhancements

* via a github automated check, a number of automatic validation are run
on the translation (presence of a disclaimer, HTML validity, ...)

* once these automated checks pass and at least one reviewer validates
the translation, the translation is automatically published by W3C and
linked from relevant W3C pages

* the same repository where the translation document was brought via a
pull request is used to maintain the translations as changes in the
original document or mistakes in the translation get reported


I have set up a repository on the W3C github space where I expect to
document progress and issues of this proposal as it moves forward:
  https://github.com/w3c/translation-management
I invite interested people to watch the repository to keep track of its
progress.


As mentioned above, we are very much interested in feedback on this
proposal before we get too deep into experimentation. If you have any
concern with this new workflow, please let me know - there is still
plenty of room for adjustment based on input we will receive.

There are more specific questions we have already identified in which
feedback would be great appreciated (either on this list, on the github
repository linked above, or as private mail to me).

Based on preliminary conversations with some of the translators, we've
heard some concern about requiring the use of github for the translation
process, since not all translators are familiar with this platform.
We're particularly interested to hear how broadly applicable that
concern is.

We are also considering to suggest (and maybe overtime, require) a
dedicated template for translated documents, which would associate them
more clearly with W3C, and would also include a pre-defined disclaimer
that the only authoritative source is the English version. How much of a
burden would this represent to current translators?

To help guide the redesign of the "intent to translate" workflow, we're
also keen to hear how in practice translators pick which document to
translate, and in particular, what motivation drives their choice of one
document vs another.

Are there additional improvements to the translation process that we
should consider as we're looking at overhauling?

Finally, we are looking for a few volunteers to help test-drive some of
the early implementation work of the project in the upcoming few months.
Please let me know by private email if you are so inclined.

Many thanks,

Dom

1. http://lists.w3.org/Archives/Public/w3c-translators/2017JanMar/0000.html

Received on Wednesday, 12 September 2018 08:17:20 UTC