Watchlists, watchlists, watchlists!
After a long hiatus, we're back! This week we are interviewing Addshore, who is an administrator and bureaucrat on the English Wikipedia as well as a MediaWiki developer. He currently works for Wikimedia Deutschland as...well, I'll let him explain.
What do you do at Wikimedia Deutschland?
- I am a Software Engineer primarily working as part of the TCB Team (TCB stands for Technischer Communitybedarf, which roughly translates to "Technical community needs").
- I also work on Wikidata and Wikibase!
One of the features you recently worked on was category watchlists. There were a few hiccups with the deployment the first time, but how did it go overall? If you had to do it over again, what (if anything) would you do differently?
- The initial hiccup was a security bug slipping through all of the review that happened on the initial patch, very unfortunate, but not much to change here other than more eyes on code/more review! Other hiccups, such as flooding CheckUser with basically useless data, perhaps could have been forseen, but again nothing really to change here. The initial deployment was basically, turn the feature on everywhere. In hindsight the second deployment, which slowly rolled out to various projects in groups, was far better.
- You can track the usage of the feature at https://grafana.wikimedia.org/dashboard/db/mediawiki-catwatch-feature
What cool new project are you working on next?
- Expiring watchlist items & general watchlist improvements & refactoring, phab:T100508.
- Watchlists have remained the same for years now, and we know what would make them better, (expiring watchlist items specifically as a request from the German community), as well as other changes! Probably going to be a rather slow process though.
As you know, there's been a lot of controversy over the past month, causing heated tensions in some cases. So lets hear it, where do you stand: short array syntax or long? (context)
- Having said that I have nothing against the long syntax...
- It makes sense to keep the code consistent with itself and I think the single patch changing all long syntax usages to short usages made much more sense than waiting N years for all usages to finally be gone spread accross 100+ patches..
Check back for the next Signpost on July 1.