At least, since they're closing access from alternative sources, they're slowly succeeding in driving me off facebook. At least if I now have to move to an app, I'll pick one that supports end-to-end encryption (which facebook+xmpp+otr did fine, btw, until they killed it).
I wonder how quickly things will turn in the coming decades, as the incumbent network silos continue to screw users on the altar of data mining and ad revenue, and anyone can start a competing product for free (up to ~thousands of users at least, and at that point, maybe you can find a monetization strategy or get investment to keep going until your product becomes another "big thing").
As far as I can gather, XMPP is pretty much a dead end at this point, Google Wave is open, but also dead as a spec. So what'll presumably happen, is that eventually someone will find a new, sane, extensible and open protocol, and it will take over. We now know some things that other protocols were missing: easy serialization, sane hypermedia, sane extensions, sane push, sane presence, sane async and sync modes (mail and im w/server side archiving) end to end encryption, and some form of trust or other way to mitigate spam. I'm guessing it'll be json based with perhaps swappable transport. Probably should be encapsulated in message-level authenticated encryption. And it'll probably be another five years before we see it realized.
The author of Conversations recently wrote up the state of XMPP in 2016: https://gultsch.de/xmpp_2016.html
So, though you may not be using it anymore, XMPP is far from a dead end.
[ed: submitted that to hn, as I'm sure many people would be interested in an update on XMPP from someone that actually works with XMPP implementations:
Oh right, I didn't know they have my whole Android address book before. I recommend you to download the data Facebook has about you (they don't give all, but a fair amount) https://www.facebook.com/help/302796099745838
I was actually surprised at the sheer amount of my personal life contained in that file.
It is Apache wave. And nothing is dead if a community wants it to be alive.
I really, really would like to see this happen, but if the past decade and a half is any indication, the ones taking over will be newer centralized, ad monetized systems. Most of the non-tech people don't care much about privacy or lock-in. I keep trying very often to explain to people why privacy is important and why platforms like FB/WhatsApp, etc., are not good. It's an uphill battle though, and most people don't get it or feel it's too much of a hassle to change their habits and that of the people in their networks.
Maybe that's an idea for an open source Android demo app - one that requests a certain set of (or all possible) permissions, and then simply displays the status of those permissions (granted/blocked).
I am an occasional user of Facebook Messenger, and using the mobile website was perfect: I don't care if notifications are delayed or even lost (I'll see the messages eventually anyway). But there's no way I'm going to install a separate app for something I use maybe once a week, tops.
If it stops working from the mobile website, then I just wait until I'm back at my computer to read fb messenges.
But I hope you can imagine why a chat team does care if a technology doesn't allow them to send timely and reliable notifications.
It is often better to have no feature than a bad feature. I've never used the mobile web messenger, but if it is unreliable, only works in a few browsers, it makes sense to remove the product.
Sure, that's why they have the app. But it's not a reason to completely turn off messaging from the website. Heck, they could even turn off chrome notifications completely and the website would still be completely functional.
1) you can never assume a recipient ever has their phone on, with them and in a non-silent mode 2) users can disable push notifications for individual apps anyway
That said, the reasons you mention are pretty edge cases. Instead of just hating on it, just stop and think about it for a moment without prejudice. When you email someone you don't expect them to read it immediately, but when you send a text you at least expect them to know they received something, no? (If you say no to this, I really have nothing else to say anymore, you would be just denying the reality) That's what Facebook wants to be. They're competing with SMS, not email.
Also, if they're supporting desktop clients I again call shenanigans on your argument.
Shenanigans!
I am amazed how many people on this thread are attacking me saying "Android Chrome supports push notification" when I said "Chrome doesn't support push on iOS". Are you even listening to what I say?
It's not that we don't hear you, it's that the reason you're in your position is because you're using a locked down platform that's trailing all other vendors in web technologies intentionally.
> Even if Chrome had push notification on iOS
This is what I said. And this applies to rest of the paragraph. Also open standard has absolutely nothing to do with this, and that was exactly my point. You guys think just because it's open standard it's all open and good, but my point was it is not. The reality is Google and Apple are effectively duopoly when it comes to anything push notification related (GCM is the gatekeeper for Android, APNS is the gatekeeper for iOS). So it doesn't matter if a technology is open standard, it wouldn't even matter if it was completely open source. The reality is your push MUST go through either Apple or Google.
Facebook has already implemented this. It is in use today on mobile web. If Safari on iOS (which powers "Chrome" on iOS) is updated to support the Push API tomorrow it will start working. Facebook has already implemented it.
As for the rest we're not even talking about web at all. Back to your original post:
> The web (html documents fetched via http requests) were not built with mobile devices in mind, especially websockets. In the mobile era, all messaging is based on push notifications which must go through the device vendor (APNS or GCM), and trying to support the legacy tech in this landscape is just not worth it.
"Legacy tech" isn't necessary to support push notifications on web. Yes, Safari on iOS is woefully outdated for now (which again is what powers "Chrome" on iOS). But it is currently supported on Android (across multiple browsers), soon Windows, and it's possible Safari on iOS will add support in the future https://onesignal.com/blog/when-will-web-push-be-supported-i... Facebook is already using this today.
I work for a living and carry a work phone that is managed due to compliance requirements. I could pop on Facebook and check the occasional message. Facebook makes the mobile web experience as user hostile as possible, but it's sort of works. Now I'm stuck, as are about 20,000 of my coworkers.
FB is a real lightning rod topic.
Says you. I'm perfectly happy not knowing about my FB messages until I go look for them. I DON'T WANT them pushed to me. If they want to send me something that needs to be read quickly, use email. Or gasp phone/text/whatsapp (I use the latter begrudgingly because whatsapp has pretty much replaced texting in lots of Europe).
Except native apps aren't the future :/