Thanks for the comments, Erik.

The purpose of the video was to show initial ideas to the community to start getting feedback from the beginning.
Since translation involves a repetitive workflow and any minimal change may affect its fluency, we think it is better communicated with movement than with static images. 

However, the video just covers part of the scope for the UX improvements planned for translation tools. 
Some relevant scenarios are still missing and we are working on them, these scenarios include translator sign-up, searching for translations, proofreading and also page translation.

Both proofreading and page translation have some aspects in common that make it challenging from an interaction design perspective: for these scenarios, compared to message-based translations, users want to read (and compare versions) in a wider-than-paragraph scope (ideally with final formatting), but act on a smaller scope (translating sentence by sentence). In addition, concepts such as message description do not apply, and suggestions may have an even reduced scope since previous translations of the same exact sentence are probably lacking (but spelling-suggestion-like completion for parts of the sentence may be helpful for speed and quality).

We are preparing UI concepts that support the above ideas, and we'll be sharing them as early as possible.


On Tue, Aug 28, 2012 at 8:56 PM, Erik Moeller <> wrote:
On Thu, Aug 23, 2012 at 3:06 PM, Pau Giner <> wrote:
> I wanted to share a short video that illustrates some ideas for the user
> experience improvement of the Translate extension.
> The video is available at  (links to
> different language-specific versions of the prototypes are available at the
> video description)

It's really gratifying to see this kind of comprehensive prototyping;
thank you for that, Pau and Arun.

The current design/UX efforts seem heavily focused on the message
localization use case, vs. the page translation use case. For example,
if you translate,_June_2012 using
Special:Translate, currently the interface gives you a full view of
the entire text of the page you're translating. In the proposed UI,
those paragraphs would be abbreviated to a one-line representation.
Indeed it seems like we should increasingly move away from terminology
like "messages" and the message-centric view for page translation.

What's the current thinking of improving the workflow for page
translation, or preventing usability regressions in that area?

Erik Möller
VP of Engineering and Product Development, Wikimedia Foundation

Support Free Knowledge:

Design mailing list

Pau Giner
Interaction Designer
Wikimedia Foundation