- 2015 WM Hackathon: First time I tried to convince people by coding a prototype of a web app in (the presentation seemed to excite people a lot)
I wasn't able to get support for pursuing this so I tried to prepare more information and a more serious project that end of year and start in 2016.
- A more formal version of the project, proposed to be worked on the Reading Web
team (previously Mobile Web team)
- Prepared a presentation on early 2016 on the Wikimedia Developer summit:
WikiDev Summit 2016: Next Generation Content Loading and Routing
- Did a lot of research, and a prototype. Went equipped with data and information. We weren't given much time and the discussion afterwards shows the opposition of many based on opinions.
- loot-ui: Frontend server and webapp we
used to prototype and do measurements
- Documentation, with feature demos, architecture review, and code docs
- Mediawiki documentation: Projects/A_frontend_powered_by_Parsoid
- Prepared a presentation on early 2016 on the Wikimedia Developer summit:
WikiDev Summit 2016: Next Generation Content Loading and Routing
After all the effort and the response from the engineers and the lack of support from the foundation I lost my interest on the topic.
-
My colleague Jon Robson later that year did the same sort of thing:
- https://github.com/jdlrobson/weekipedia
- He went all in with just doing coding on his own, and took it really far. Basically replicate the whole mobile website to the point where he could run the browser tests against this new app and they passed.
- He got no support from the foundation or other engineers, like I didn't. He kept it going for a while, maintaining it and keeping it working (til 2018 or so) but I think he has lost energy.
-
Later I did a small SPA I used for prototyping New Readers features to perform user tests
- The most ambitious try as part of the Reading Web team was on 2017. I worked
hard with the PM to make a proposal for a new mobile website:
- mediawiki.org/wiki/Reading/Web/Projects/NewMobileWebsite
- We worked on it for one quarter
- https://github.com/wikimedia/marvin
- Best name ever
- We worked with PMs (see docs above, we made a roadmap too), designers (marvin blueprint), started socializing the project with engineers in the team and some other interested folks (presentation), we talked to the performance team, to other people in all hands before project...
- Things were going well, we had only 2 devs working on it but things were moving along, and then,
- Some people freaked out when they saw
T177057: Clarify production deployment requirements, outline and blockers
and, as far as I experienced:
- they got very defensive, and apparently started grassroots opposition without talking to us. I got an email directly from the CTO asking me to stop things and do an RFC (we were just prototyping and setting things up), and everything started to get very strange. Support from Product leads seemed to disappear.
- We ended up stopping the project because of all the opposition and blockers from people in tech and the lack of support from Product.
- https://github.com/wikimedia/marvin
After all this, a few months later, I moved to the Product infrastructure team and have been there for 1+ years, supporting API and backend work as an Engineering manager, in order to empower 3rd party API clients like apps, but also enable the possibility of a web PWA existing at some point in the future.