Discuss the future of our Translation platform
Once sysadmin#17782 (closed) is done, it'll be a good time to have a conversation about several topics that are related to each other.
Strategy building
Goal: decide on high-level strategic goals that will inform the roadmap.
Past and present
Goal: ensure we're on the same page regarding the past, so we can then focus on the future.
Process: read this section and add your comments on the pad.
Long story short
- 2015: prototyping Weblate for Tails
- 2017Q4: we got a 11k€ grant approved to implement this in production.
- Stated goals:
- lower the technical barrier for new translators to help
- get translations into other languages such as Spanish, Arabic, Chinese, Russian
- For various reasons, this project did not succeed within the planned timeline & budget.
- Stated goals:
- 2018: we negotiated an extra 8.5k€ from Tails unrestricted funds for a new team to finish the bulk of the work
- June 2019: hefee & intrigeri complete their part of the work, release it to a wild, unsuspecting word, and hand over to Zen Fu for maintenance
- 2019: we realize that what we deployed is not exactly ready for prime time and is a PITA to maintain
- 2019-2022:
- Most of the dev/sysadmin capacity on this team is spent on fixing bugs, upgrading Weblate, and finally on making the backend easier to maintain & upgrade. All of which took significantly longer than anticipated, but eventually succeeded.
- The Translation Platform uses about half of our sysadmin capacity over the last 2.5 years. This has made us put other sysadmin projects on the backburner, which in turn blocks improving other areas of Tails.
- emmapeel is asked to be patient until the aforementioned backend work is done, before we can give her workflow and UX improvements proposals the attention they deserve. We understand this has been frustrating.
Impact on translations and translators
-
A few new translators for existing languages (https://translate.tails.boum.org/manage/users/?q=&sort_by=-profile__translated), yeah!
- OTOH, a few new translators have hundreds of pending suggestions waiting to be reviewed, which suggests the translator/reviewer ratio is not right.
-
We lost some historical translators but that's probably not related to the translation platform.
-
In the last 12 months, 3 languages that were already activated on our website are improving, 2 are regressing
Language 2021-02: core pages 2022-03-15: core pages 2021-02: full website 2022-03-15: full website de 44% 29% ↓ 26% 20% ↓ es 90% 69% ↓ 52% 44% ↓ fr 81% 91% ↑ 79% 94% ↑ it 60% 70% ↑ 31% 35% ↑ pt 39% 56% ↑ 18% 24% ↑ -
Languages activated on our website in the last 12 months:
- 1 new (Russian), which has seen no activity on Weblate since months
- 1 removed (Farsi)
-
We lack recent qualitative data from existing translators, that could help prioritize the next steps.
- For example: survey users to know how they feel compared to the old setup, or what they find difficult about using the translation platform.
- According to https://gitlab.tails.boum.org/tails/summit/-/wikis/2021/Team_reports#translation-platform, "Translators report it's much easier to use than the previous (Git-based) system".
- Gathering and relaying such information is part of the "Community management" role on this team.
- emmapeel reported such data on GitLab issues where she proposed UX and workflow improvements. It's not clear if the UX data that backs these proposals is still up-to-date.
Backend maintenance
It is now vastly easier to maintain and upgrade the Weblate application.
Team health
We lack recent self-evaluation of the different roles on the team.
What we know:
- Zen Fu reached the biggest milestones the team had set wrt. the backend. There's something to learn from the fact it took more work and a greater lead time than planned, which impacted emmapeel and other sysadmin projects.
- Since a year or so, emmapeel has not been as active as initially planned in Tails in general, and on this team in particular. We're aware of various reasons for this. We would understand if the "please hold on and be patient" moto had been a big demotivating factor.
- In 2021 we clarified decision making on this team, in an attempt to address tensions between emmapeel and hefee.
- hefee is happy to help Zen Fu occasionally, in a "Helper" capacity. Zen Fu is fine with this.
Where do we want to be in 2 years?
XXX
Resources:
- tier-1 languages
- sajolida's email sent on 2022-03-10 (
id:de4dba82-0e0e-5299-e381-a3f3b17155d8@pimienta.org
)
Understanding capacity
Goal: understand what the actual capacity of the team actually looks like
- General time commitments of the folks participating in the team
- Amount of necessary maintenance "KTLO" ("keeping the lights on) work
Roadmapping
This is informed by:
- the strategy decided above
- the KLTO work:
- upgrading Weblate regularly (sysadmin#17890 (closed))
- anything else?
Potentially useful links:
Team's needs
What skills and capacity does the team lack to execute the roadmap?
Misc
- What's the future of hefee on the translation platform team? Is the hand-over finished? Does it still make sense that hefee is a member of the team? Or formalize a "Helper" status like the FT is introducing (not formally a team member, not part of meta/self-organization/decision-making, the team can delegate them some work)?
Related issues
Blocked by:
-
sysadmin#17782 (closed) -
#18307 (closed) -
Gathering data about the benefits of the containerized setup (#18308 (closed))