|Open||Petrb||T172919 Implement new review filters into Huggle / Major UI refactoring|
|Open||Petrb||T225928 UI refactoring: Visual signals are hard to interpret|
|Open||Petrb||T225929 UI refactoring: Some helpful aspects are hard to identify|
sitting here ROFLing to myself
if I thought the icons I whipped up in photoshop in 2007 would:
(a) still be in use ten years later, and
(b) be analysed in a WMF usability presentation
I probably wouldn't have bothered releasing huggle in the first place...
(they made perfect sense to me at the time, and I wasn't up to the task of making any kind of AI to determine faith-ness of edits which seems to be happening now)
Hi Gurch, nice to hear from you after all these years!
Why "wouldn't have bothered releasing it"? It's a great thing and many people still find it very useful :)
Regarding icons, there isn't really a big need to change something that works, and these icons did work pretty well for their purpose, so I guess there was no need to change them in any way. This presentation was based on Wikimedia new editor engagement survey, probably with intention to make Huggle more friendly to new users (or rather to make its users aware that whoever they are about to revert is a new user). They are also addition to the existing icons, not replacement. Nobody really has any issue with the icons you made back then.
I am still not myself decided on if all this new stuff as it is suggested is really worth all the work, but some of the ideas aren't bad.
Just for fun I downloaded the old (2.x) huggle and it only needed two code changes to make it work (https, and screen-scraping the login page). I was pleasantly surprised.
I think their proposed icons look a bit too cutesey for a change management program... the people using it aren't the new users they're trying to engage with! Rather they're concerned about people being put off by instantly-reverted edits and template warnings. Although editing volume has declined since 2007, it's hardly on the level where you can respond personally to every vandal, so it's a necessary evil. The problem of people using something like huggle to revert well-intentioned edits, that's more of a social problem than a software one.
They make a good point about icons being colour-sensitive (which I did consider changing). But then they propose icons which are also colour sensitive, so I dunno.
It's centered around the "Objective Revision Evaluation Service" thingy (which is extremely useful if it works, but the fact that such things don't work perfectly without strong AI is why huggle exists in the first place). Needless to say if ORES had existed before, it would have worked differently. The main "good edit" finder was always the user whitelist (I find it amusing that when they introduced the "extended confirmed" user group, they picked the same thresholds I'd implemented for whitelisting in huggle). ORES is obviously a lot more sophisticated than that. Whereas the existing icons are based on things that can be objectively determined (such as what warnings a user received). They're mainly there to account for policies on Wikipedia that long predate huggle (four warning levels => Adminstrator intervention against vandalism => blocking). I think to achieve what WMF is getting at, those policies would also need to change (and at this point they're so entrenched I can't see that happening).
I never put a "revert and send friendly message" button in because it was never a common enough action that it needed a special button to combine the two. You would select "Revert" and then "Send user message" and pick a template from the list. "Revert and warn" got combined into one button because that's 95% of what you do with huggle. Extra buttons clutter up the interface (which used to be a lot more complex and you've done a good job at simplifying it).
Still weird to see people analysing things like this though. (Someone wrote an academic paper about huggle once. It referred to AIV as an "obligatory passage point [...] in the language of network-actor theory".)
BTW just to clarify why this task is open for so long time with no update - it's simply not clear to me how to, step by step, implement most of this stuff that was proposed into Huggle's interface. Perhaps this should be split into multiple separate tasks that each defines what should be changed.