Hacker News Reader

> This is a bad idea because a web browser should not become a chat or client as well.

Well, the browser already is a chat client for millions of people. It's called Facebook. So since the browser already is the only chat client tons of people use, why not make it a client for a Free/Open chat system instead of Facebooks chat system, that they control, censor, give access to to NSA, etc.

> Remove their like-buttons from your pages!

I agree we should do this. But that won't help people keep in contact with their friends. And so it won't change the thing I'm focusing on here.

> Well, the browser already is a chat client for millions of people. It's called Facebook.

"Is a" and "has a" are worlds apart.

If your goal is to have a free and open uncontrolled, uncensored chat-system that is NSA kinda proof then lets do that, Im all for it.

I just dont think XMPP fulfills the uncontrolled and uncensored and surveillence resistent parts. We need to look at web of trust, bitmessage and such instead, those are more fun problems than "to spite facebook".

But sure, any solution will benefit greatly if its bundled with the browser since that appears to be the only program most users actually use on their computing devices. Imagine if Tor and a tor-relay or even tor-exit-node came preconfigured with every Firefox and turned on by default. That would be fun.

You might very well be right. Using XMPP is not important to me. The important thing is the goal of moving as many people as possible over from closed to Free/Open systems.

Free would be nice - but it's open that's the key (and if you have open, somebody will make a free version anyhow)

Free as in Freedom :)

> Well, the browser already is a chat client for millions of people.

I'd rephrase it. Browser is turned into a virtual operating system for millions of people already. While it can be useful, I don't really like when that trend is pushed to the extreme and everyone rushes to run everything through the browser even without clear benefits. The point of that? I always prefer standalone chat clients (as well as e-mail clients and etc.). If some chat service can't interoperate with standalone clients, it's garbage.

Facebook can by the way, since they do use XMPP. Their problem is unwillingness to support federation. They are too selfish. So I never use their service (nor do I use their social network anyway).

You have to remember that just by being a person who read Hacker News you are not a regular person. I really don't mean to be harsh, but it doesn't much matter what you prefer or that you find this trend of running things in the browser problematic. What matters is what normal people do. People who don't even know what a "client" in computer lingo means.

Those are the majority, and if we want them to use Free and Open systems (which is my whole point here), then we have to create those systems in a way normal people will start switching from the closed systems and over to the open ones. And that is only going to happen if the open systems are easier to use and more accessible than the closed ones.

I am aware of the fact that Facebook uses XMPP (I mention it in the post). Still, extremely few people use it through a client. They use it through fb.com.

> Those are the majority, and if we want them to use Free and Open systems (which is my whole point here), then we have to create those systems in a way normal people will start switching from the closed systems

It's trickier than that. There are many open and federated XMPP services around, including those with web clients / interfaces available. Options are around. It's the sabotage of federation by "big" players which is a problem (or sabotage of interoperable protocols altogether).

XMPP/Jabber itself is not a bad idea at all. As well as federation. Companies who sabotage XMPP federation by not adopting it (M$, Yahoo, AOL etc), or backing off from it (Google) or proliferating new incompatible junk (Whatsapp and other monstrosities) are a bad thing.

That's ridiculous. Someone comes up with a spec, and if a company doesn't redo their systems to implement it and connect to others with it, they are sabotaging it?

They are sabotaging the only interoperable open and free protocol (which is an Internet standard developed through IETF). Did they propose anything better? Not even close. Even Google which boasts that their Hangouts are superior and thus can't use XMPP didn't open their protocol. So who cares if it's better then? E-mail is way far from perfect, yet it's open and interoperable. So everyone supports it. XMPP is the only developed analog for instant messaging. And all those selfish jerks of companies can't agree to support it. Or develop anything better if they don't like it.

You're not even speaking coherently. They aren't sabotaging it by not adopting it. That's just called... not adopting some arbitrary spec.

Being published by the IETF doesn't mean anything. The IETF has a spec for SIMPLE, too. Should we blame all XMPP networks for sabotaging that "standard"?

Why should a third party have to do something just because you don't like how they run their own networks?

> They aren't sabotaging it by not adopting it.

Oh, really? Inaction can be damaging as well. In this case - it's sabotage. Since you bring SIMPLE, which of these major closed networks support it?

Thats right, its only facebook and google internally that take advantage of XMPP while keeping the rest out. Thats what federation does

Counter-example: email.

Caveat: email as we know it developed in frontier times when things could grow organically and/or by general consensus rather than being driven by commercial needs.

The frontier days for this sort of communication are over for the most part. Anything that shows signs of being significant will get jumped upon by commercial entities who will try to direct it in a manner that helps their larger goals (and if we are more cynical, expressly in a manner inconvenient to their competitor's goals) potentially (in fact most likely) at the expense of the overall scheme (despite publicly doing it to "help the users/community/public".

Anything new now has forces acting to shape it (for better or, often, for worse) that email never had to contend with until it was so widespread no one entity or collection of entities could wreck it for everyone (despite the best efforts of some such entities!).

well, the new frontier is "dark mail" or more generally "dark net" where interoperability again is the key [as the only way to grow. It is like natural evolution - any non-interoperability efforts will just not grow, and only an interoperability based effort will produce the critical mass]. It is organically decentralized and by its nature a user is the owner of its data even though actual [meaningless to 3rd party] bits may be stored by one or spread around multiple distributed providers ("dark money" style as if you BT wallet is encrypted and replicated though the BT network itself). Content mining and "user is the product", and thus FB and Google as we know it, would be for some time problematic at best and it will be some time before segmentation era of the new net dawns and new Serge/Larry/Zuckenberg appear :)

But it isn't an empty frontier that the thinkers/developers are entering alone. The existing problems both commercial and legal (and not-quite-so-legal-but-state-funded-so-good-luck-fighting-that-one) will be so close behind to stop/exploit/both as to make the exploration very inconvenient. I'm not saying it can't happen, but it won't happen organically over a period of time like email did.

Less and less of a counter example. If you leave out work mailboxes, most people these days don't send their email any more: they ask Google, Microsoft, or their ISP to do it from them, from their web browser.

Some people are worried about Facebook taking over the internet? First of all, I'm worried about the web taking over the internet. And becoming more and more centralized in the process.

Alas, any truly distributed scheme will face the lack of symmetric bandwidth. Which is why MegaUpload ever existed in the first place: if we could upload as fast as we download, distributed networks such as Bittorent would be too damn efficient. (On average, we're ultimately limited by the smallest of the two. Upload minus download is exactly zero over a closed peer to peer network.)

Less and less of a counter example. If you leave out work mailboxes, most people these days don't send their email any more: they ask Google, Microsoft, or their ISP to do it from them, from their web browser.

Federated is not the same as peer-to-peer. While there are big providers, email is still open to anybody with a non-dynamic IP address, and the fact that almost everyone uses third-party servers doesn't change that.

My point is, there are fewer and fewer email providers, and they are getting bigger and bigger. If for some reason 95% of the population had a gmail account, Google could be tempted to shut down all SMTP communications. Technically that would remove them from the federation. Practically, that would remove everyone else from the federation (they are on the bigger half on the network, after all).

Another, more realistic possibility is that a consortium arises to fight spam, and the result is that email that doesn't come from big players or registered entities will simply be ignored, on the grounds that most mail from little domains would be spam (which by the way is probably true). The next step will be about refraining themselves to send email that look like spam, never mind the freedom of speech issues that will ensue.

And voilà, we have a world without spam. Too bad I can't send email to most of my friends any more (I control my own domain name).

Now Facebook is even worse. It already took over. You could try and build a federation, but unless you can force Facebook to communicate with it, it will still be a mere competitor, which starts at the wrong side of the network effect. Just like any Facebook competitor, actually.

... and VoIP?

Is there much VoIP federation going on? My impression is that VoIP providers still interop through the PSTN.

I have studied XMPP and the whole protocol stack, compared to everything else I could find out there (in a paper I'm working on). However, the conclusion is that XMPP is not worth the an investment although it looks really fascinating at first and second thought. It's pretty much doable with simpler web technology, webstandards also help to keep out the madness that XMPP calls standards..

The biggest selling points of XMPP are that everbody already has XMPP. The counter argument that weakens this position is that everybody already has email, an IP and a phone number too.