[zanata-users] Discussing RFEs [was: List-forking. A clarification about Zanata admin rights]

[Sankarshan Mukhopadhyay]

On Sat, Sep 7, 2013 at 5:21 AM, David Mason <damason at redhat.com> wrote:
> I agree that a review queue is probably not the best way to handle suggestions - these are evolving ideas, with plenty of room for improvement. The concept I have in mind at the moment is that suggested translations by someone with low reputation would be presented as an option for a more experienced translator to copy as a starting point or to use as-is, something like the way translation memory is presented at the moment. The intention is that the more experienced translator would be able to tell from looking at the suggestions for the first few strings in a document whether the suggestions from a particular user are of high enough quality for them to use, so that the overhead from lower quality suggestions is not high.

[I loved the narrative and the thought; snipped it here for brevity]

I think I can now understand what you are intending to achieve and, it
is a somewhat unique spin on the process of contributor participation.
A bake-off between contributors leading to a file that may have
contributions from all "candidate contributors" hasn't been the
typical method of coaching new participants in a community. It could
actually work out very nicely by generating a fantastic set of new
terms which can then be included in the terminology (used by the
translation memory system). For example, I was recently poking around
a new entry-level phone from Nokia and, I realized that the terms used
for a few UI items (for Bengali) were much more in tune with what is
the colloquial form of the language. The next thing was to remember to
modify the existing terminology and update it. The process flow that
you describe takes this one step further - it would be possible to
maintain the attribution (original contributor so to speak) of the
term.

I'll write a bit more about how file assignment, new contributor
on-boarding and, reviews happen.

New contributors to a language are few and far between. And, from the
ranks of those who sign up only a few remain as long term
contributors. So, language coordinators have two tasks - [i] ensure
stickiness and, [ii] ensure that whatever has been the random
contribution, that has been reviewed and if found acceptable,
included. No one wants to throw away whatever little has been received
as contributions.

(This process may or, may not be followed as is by all languages. What
I write is somewhat specific to what we do.)

Absolutely newbie contributors are asked to attempt a small subset of
strings from a file that is not terribly complex eg. not SELinux or,
Virt related. They are asked to go through all the English strings in
order to obtain the context of the file/application and, perhaps the
use cases (we do ask them to have a machine image for the OS with the
application; but often that is not very readily available). Once they
do submit their contributions, the reviewer is expected to go through
the entire set of strings before reviewing. The reason we ask both
parties to read through the file once is to prevent any strange
translations ie. words which seem correctly translated but are
obviously wrong when placed in context of the entire application or,
the specific action in the application. After a few quick cycles of
this, the contributor, if (s)he wants to continue, are provided with
the rights to commit to the repository themselves and participate in
the review sprints.



-- 
sankarshan mukhopadhyay
<https://twitter.com/#!/sankarshan>