Remove Opera from BCD

[estelle]

Opera is not a company focused on its desktop browser. The browser is chromium-based, and we already have Edge and Chrome, both chromium-based, listed. It may make sense to merge Chrome, Edge, and Opera, and possibly other Chromium-bases browsers like Brave into a Chomium/Blink column.

Opera is a marketing ( or possibly crypto) company at this point. They don't even post browser release notes. The only post on their developer site since 2019 is about wallets. Being listed in caniuse and mdn's bcd is, at this point, keeping Opera on anyone's radar.

Maybe put the UC browser or Duck Duck Go browser in their place? The UC browser for Android has greater market share

[jpmedley]

The Chromium forks are not identical. I believe there are cases where Opera has removed features from its version of the fork. There's no reason Edge couldn't do that as well. Samsung is even weirder. They fork a version several years behind the current one and have been know to merge features from later versions.

Are collapsible columns possible?

[jpmedley]

And Edge is the only Chromium fork I know of that uses Chromium's version number.

[queengooborg]

We actually ended up removing UC Browser a while back due to the lack of maintenance, and because its market share is surprisingly still smaller than Opera's (0.82% vs. 2.22% as reported by Statcounter):

However, this is for global market share across desktop and mobile alike. When comparing mobile market share, their usage is much more neck-and-neck:

[queengooborg]

Overall though, I'm +0.5 on removing Opera, especially given their track record of privacy issues. But the reason I'm +0.5 and not +1 is that Opera still has a significant market share, and I'm skeptical to remove data for a browser with that much usage.

I'll keep this issue open though for additional discussion -- I'd love to hear what others think!

[estelle]

Is it possible that their market share is as high as it is because they are in the BCDs of so many sites, including ours, w3schools, and caniuse?

When I add up all the user numbers for opera, v10.1 to 83, on caniuse I get 0.88%. I do not know why there is such a discrepancy between statcounter and caniuse

Added note: Opera is at 5% in Europe. So maybe we can't get rid of them. I wonder why they're so high. Maybe the free VPN, or maybe b/c Opera used to be from there. FF is at more than 33% in Africa, while Safari is at 2%.

Added Note 2: https://jobs.opera.com/ - they have a lot of crypto roles, but only one browser-related role open

[queengooborg]

I do not know why there is such a discrepancy between statcounter and caniuse

What's even more odd is that CanIUse apparently gets its info from StatCounter...huh.

[lukewarlow]

I guess this might be better placed on MDN content repo, but my two cents is that the opera data isn't useful to MDN viewers and is in fact misleading at times. I've often heard oh this isn't supported on opera that's a shame. When in fact it is supported it's just that the release data for opera (or opera android isn't up to date), or historically but thankfully fixed by mirroring the support data simply hadnt been updated when opera came out a month or two after Chrome and Edge. We also can't automated the release data for opera because there's no machine readable release data.

Edge has meaningful differences from Chrome with many APIs. Samsung internet at least historically has too (WebAuthn and WebXR come to mind). I'm unaware of any APIs that have been implemented in Opera but not chromium, or any that have been in chromium and not enabled in opera. Brave, Vivaldi etc don't exist on MDN (or BCD for that matter) and I don't see why Opera is special in that regard. (I understand opera does have the historical presto data but that's not meaningful for most MDN consumers)