Unlike Google, which look much more arbitrary in their shutdowns, Mozilla typically kill projects that are already on their last leg, or replace them with something objectively better.
Yes, they made some mistakes and adoption was poor. But it worked well and the underlying protocols are something the internet still desperately needs.
It was DELIGHTFUL, and while I understand the rationales for why it was canned, I still harbor some sour grapes.
I know I played around with a proof of concept implementation internally, but never deployed it publicly, which was a decision based primarily on their own messaging.
[1] EG the JS implementation and the dependency on persona.org were both temporary stop-gaps to enable fast-prototyping and use in other browsers, but would be replaced with a "real" implementation in the browser that supported multiple providers before broad roll-out.
IMO, the main mistake was that it used a centralized authentication system to "bootstrap" the protocol - basically a fallback auth source back when nobody supported the protocol. In theory, after the protocol was adopted widely, they could shut down the Mozilla Persona servers. In practice, it ended up being perceived as "Mozilla Login" and the branding (dropping BrowserID) didn't help.
It didn't have to be this way, although I understand the reasons for it. The fallback auth system could have simply been a piece of software infrastructure that you (the site operator) were expected to run locally. It would have had two downsides: 1) it's a fairly complex piece of software that sends email and interacts with the website, so it would have been hard to make portable, and 2) it would mean doing the email auth dance separately on each site (like we do already).
The main advantage, though, would have been that it wouldn't depend on Mozilla running servers. The project could live on.
Honestly, if I had free time, this is one of the things I would be working on. The protocols are already written up, it just needs software.
Take Sunbird. The source tree is organized as a set of backend modules, with various products including as many of them as they need and then throwing up the main frontend window on top (although SeaMonkey really has diverged its UI in quite a few places). Sunbird was the standalone calendar app, but it was decided that the calendar maintenance team was too small to maintain a standalone calendar and instead shifted their UI focus to only supporting SeaMonkey/Thunderbird integration (aka the Lightning extension). No code that actually did anything was removed--I was the one to actually hg rm it here: https://hg.mozilla.org/comm-central/rev/f2e605ee5475
As you mention, MXR was phased out in favor of DXR, which is itself being phased out in favor of searchfox. (I too worked on DXR, back in 2011). Prior to MXR, there was LXR (as in, lxr.mozilla.org), but I'm not sure it's worth mentioning as a separate project since it's basically an unmaintained fork of the LXR project.
While we're on the topic, there's a lot of "missing" webspaces here. Bonsai was a webtool to display atomic CVS commits, Tinderbox was the UI for seeing if the tree was broken (on fire) or not (this later gets supplanted by TBPL (tinderbox/pushlog), itself rewritten to not be a pile of hacks as treeherder), and there's definitely a buildbot clone/fork/I don't know floating about somewhere in the mix.
There's various versions of things like the addons website or support websites that I can't even enumerate off the top of my head as well. Some of these (like old addons stuff) was completely in-house, while others (like pastebin and etherpad) are hosted versions, potentially with some modifications, of other projects. None of these are listed, but do they even deserve to be listed? It's not consumer-facing, and I'm sure even many Mozilla developers would struggle to realize some of this stuff existed.
Thanks for the info about Sunbird, I had the understanding that some code was shared, but wasn't sure how much. I'll update it's description at some point.
You bring up good points about stuff like forks, internal tooling, and hosted versions. Up until now I haven't included forks and hosted versions that didn't make significant changes (which is why I included MXR and not the LXR instance). I'd like to include internal tools where possible, but I think the main obstacle is finding sufficient info.
I think where this differs from Killed by Google is that Mozilla has had less consumer services that were discontinued while they had an active user base. I'm happy to include old experiments, internal tools, etc., I really enjoy digging these things up.
Thank you again, I'll take a look at some of the services you mentioned!
But why? What's the purpose of a public website with a catchy domain that chronicles when internal-facing tools were killed (usually because they were replaced with something better)?
I understand the agenda behind Killed By Google, but right now your site just leaves me scratching my head.
(Disclaimer: I work for Mozilla on the ops team that ran many of these services)
No, they 100% killed Nexus.... Pixel line is in no way, in any shape or form a successor to Nexus.
Nexus was a reasonably priced unlocked reference phone targeted to Power Users
Pixel is an iPhone Competitor targeted to Carriers to be featured as an iPhone Alternative
Completely different markets, focus and goals
>>AngularJS? They took the framework in a different direction.
Again no, Angular is TypeScript, AngularJS is Javascript, these are not the same thing, and it is far from "a different direction"
Nexus One: $530 Nexus S: $530 Galaxy Nexus: $200 with standard 2 year contract (basically the same as any flagship phone at that time) Nexus 4: $300 Nexus 5: $350 Nexus 6: $650 Nexus 6p: $500 Nexus 5x: $379
Arguably Nexus 4 & 5 were the odd ones out with their relatively low pricing. The others (excluding the 5x) were pretty much on part with flagship phone prices at launch time.
Any new "external" feature is another matter, of course.