Jeff Beatty, Arky, George Fiotakis and Zbigniew Braniecki (Gandalf)
Localization Contribute Group
Meetings every other Wednesday at 6 pm pacific.
- Next meeting: Contribute/L10n/Sprint_06_2012
- Notes from previous meetings
- 2012 Q4
- Area focussed mentorship
- Middle East, Africa, SE. Asia etc.
- Set up automated response to inquiries using stock responses.
- L10n team health assessment
- Create health assessment form (i.e., what questions do we want to ask, what info do we need from them, etc.)
- Set up work week in December.
- Shared IMAP folder of responses/follow-ups
- Success stories: New contributors who have successful contributions coming from Get Involved page.
- Part of blog? Add names to https://l10n.etherpad.mozilla.org/socialmedia.
- Set up github account for slide decks.
- Use a repo with existing personal accounts for now.
Audit existing channels where potential contributors find information (or should be able to find information) about how to get involved and then optimize to make sure the channels are effective.
For an area of the project that has been active for so long, such as l10n, there will be many old pages with out of date information on them that should be retired.
- Audit all legacy paths in wikis (example: Google Mozilla localization & you'll find legacy articles in the first 5 hits).
- Create plan to drive traffic to new contribution paths and away from legacy paths.
How can we use events as offline channels to help people get involved?
- Identify l10n-related events for recruitement opportunities.
- Leverage the ReMo program -- help bring in more technical Reps and help non-technical Reps promote l10n opportunities by creating material for them to use (videos, hand-outs, etc.)
- Encourage more localizers to join ReMo.
- Make sure key community events have someone from l10n going to talk about how to get involved. For example, see this MozCamp Asia Report post about the effective coding session.
- Create a 'How to get involved with Mozilla Localization' flyer that can be handed out at events and made available for people visiting Mozilla Spaces.
There is a history of students at universities and other schools contributing to Mozilla coding projects as part of a class assignment. This is something that could be considered for students in localization related fields.
- Create list of academic institutions/programs located throughout the world from which to recruit.
The best practice for dashboards is measure the number of people expressing interested in contributing to your projects, the conversion rate of the number of those people that start contributing, and project specific metrics about activity and health of community. The Get Involved dashboard has information about the first two metrics and project specific dashboards may need to be created for the other data.
Using the dashboards:
- Look at the conversion rate and use that to experiment with alternate contribution paths to find the optimal way to bring people into your projects.
- Conversion rate is measured by mapping a contribution path, identifying key points and looking at how people are proceeding along that path. What are the right conversion points for new l10n contributors?
- Look for people who have recently stopped contributing and contact them to either encourage them to contribute again or learn information about why they stopped that would help other people stay involved.
General Community Building Tools
Track ideas for tools we want to create to help with community building.
Match Volunteers with Opportunities
- [Insert name here] is currently fielding inquiries about l10n from the Get Involved page. Can the response we're sending be optimized? Can we tell who is successfully getting involved? Can we reach out to people who haven't successfully contributed and ask them why they didn't proceed?
- Research how other teams (e.g., BrowserID and User Engagement) might perform L10n recruiting in tandem with the existing L10n communities, eventually integrating them into the fold.
Provide documentation and other materials, such as videos, that help people learn how to successfully use contribute to Mozilla.
Existing presentations about l10n could be pulled together into one easy to reference location that could be used by Reps and others.
Identify the key milestones in the contribution process and help people along the way either with manual or automated processes. For instance, Bugzilla is now sending email to people who have had their first patch approved by a reviewer.
- Aaudit all of the steps a new contributor takes to be successful and identify where to introduce manual or automated assistance.
Create a scalable process for identifying and recognizing key active contributors with swag, invitations to events and work weeks, etc.
Have a plan for getting newly active contributors into the phonebook and use relevant tags (ex, firefox, l10n, etc). This will allow us to reach out to experienced contributors with specific opportunities as they come up.
Audit and update the existing module ownership information for l10n modules. Identify process for recognizing new contributors and creating new modules/owners/peers.
Q: Can you identify all of the contributors on your team (both paid-staff and volunteer-staff)?
A: Paid staff = Yes, since there are only eight of us. Volunteer staff = There is no way :-) We have hundreds of volunteer contributors that are working on L10n. Searching the L10n tag using Mozillians only returns a hanful of results. We could do better at promoting signing up on Mozillians and using the L10n tag. The Mozillians directory really isn't the best way to understand our community. With that said, we've been able to understand/identify our contributors through various surveys and which we use when tailoring campaigns, projects, or documentation to the communities as a whole. Jeff
Define Contribution Opportunities
Q: Can you point someone interested in contributing to your project to a list of available contribution opportunities?
A: We've placed a more concerted effort on identifying contribution opportunities and lowering the bar for entry. We created this contribute wiki to point to specific ways people can contribute. We've been working on re-writing our wiki doc as well to clarify the actions that new contributors should take to join in the effort. Jeff
Map Contribution Paths
Q: Are there clearly understood steps someone can follow to go from knowing nothing about your project to successfully contributing?
A: This wiki page is meant to accomplish just that. It outlines a technical, task-oriented workflow for anyone interested in getting involved to follow. Unfortunately, it can be a bit daunting due to the large amount of information that a new contributor needs to know in order to get involved. I think we could improve in working to simplify the introdctory process. Ultimately, that's what we hope this new guide will help to accomplish. In addition, the World Ready program Stas has recently launched is meant identify potential mentors. As this program progresses we'll find points in the process in which to add a mentor for individual contributors. Currently, Arky is filling the mentor role for new contributors who are starting an entirely new localization. Jeff
Establish Goals and Metrics
Q: Can you measure participation or contributors today? If so, what metrics can you track? What goal or metric would you like to achieve for Q1? Alternatively, what metrics would you like to get in place for Q1?
A: Right now we have multiple dashboards for tracking each localization team's progress on a per project basis but not for individual contributors. The high volume of contributors makes it very difficult to track each individual contributor's efforts. It's especially difficult because of the diverse manners in which an individual contributor is able to contribute to the project. Some dashboards (e.g., for wb parts) can be modified to track some individual contributions, but that metric wouldn't be representative of all individual contributions. To put it simply, yes, we have the tools to measure a team's contributions but not an individuals.
Our main focus/goal in this next quarter, in terms of engagement, is to increase the size of each L10n community. We're attempting to do that by lowering the bar to entry (through new tools development and clearer doc), creating a mentoring program (like World Ready), and providing a more centralized and clear platform for project tracking and joining the communities (elmo project). Jeff
In this community profile, I want to point out some of the l10n drivers' approaches to managing the l10n community, its efforts, why this inhibits our ability to track their lifecycles, but why we'd like to keep it this way.
We largely allow l10n teams to govern themselves. Each team creates its own organizational structure, manages assignments and tasks according to its needs, and selects where and how it would like to contribute. We teach them correct principles and provide a solid infrastructure, they do govern themselves. At times we step in to re-teach and remind, as well as provide them with new information should that info affect their workflow.
Here's an example of how self-governing teams make it difficult to track contributions. Hg is a difficult tool to learn. It's also challenging to aqcuire the necessary permissions to commit and push to product repositories. For these reasons, many localizers prefer that only one team member have hg commit access. This means that while the whole team may have participated in producing the localized build of Firefox 15, the only statistical evidence of contribution is linked to the ID of the localizer who performed the hg commit. False data like this is common, however, we would prefer it to requiring that everyone working on localizing products learn hg and gain commit access. You know what they say about too many chefs in the kitchen.
Freedom of choice
You may have noticed that we have a large selection of l10n tools that our available to our localizers. While this does create interoperability problems for us, it also allows many l10n teams to work in whatever way they choose. In addition, few l10n tools contain PM utilities or metrics utilities. We still prefer our l10n teams have the freedom to choose because it enables more people to get involved in l10n without any technical ability requirements nor internet access requirement (both of which could halt the progress of several of our l10n teams).
Our primary division amongst l10n tools is online vs offline:
L10n teams using online tools do so for the following reasons:
- Little to no technical understanding is needed.
- Easier to collaborate when there's a good internet connection.
- Provides a central "beta" repo in which suggested translations can be reviewed, voted on, and approved before they're pushed to the locale's repo.
L10n teams using offline tools do so for the following reasons:
- More challenging, gives a sense of "hacking" a localization.
- Only option available for regions without a good internet connection.
- More portable, as all work is stored locally.
- Allows team to determine which l10n tool they want to use. Some teams use full CAT (Computer-Assisted Translation) tools which allow them to create and maintain terminology and translation memory all in a single tool. Others prefer to use the command line and a text editor.
Plethora of projects
There are many, many projects to localize. A localizer has the following l10n project options when they join Mozilla:
- one or more products,
- Mozilla web content,
- SUMO articles,
- MDN articles,
- and any and all marketing campaigns that come down the pipeline.
Each of these projects has a different workflow, and even sometimes fragmented workflows depending on whether the localizer is able to work on- or offline. Even then, it's hard to determine where the division of responsibility lies for localizing projects like SUMO or MDN (Who monitors the localizer's activity for either project, the l10n drivers or SUMO/MDN? Should the product team be monitoring a localizer's activity if they're localizing product or should we?). We would prefer, however, that l10n teams determine where the locale's needs lie and that they ensure that a localizer is participating in the projects they're passionate about. That being said, the obvious consequence is that we're unable to track a single localizer's contribution through every single possible contribution pathway.