Hacker News Reader

> Element ends up with some small (10%?) marketshare alongside loads of other Matrix vendors out there doing their thing.

What is a good reason for Element to relegate itself to only 10% of its own market and allow other vendors to pick up the remaining 90%? Can you name a product or service which Element is unable or unwilling to provide that competing investment could make returns on? If you can, why not pivot into that yourself? Can you assure parties investing in Matrix that Element is a trustworthy and synergistic counter-party?

> What is a good reason for Element to relegate itself to only 10% of its own market and allow other vendors to pick up the remaining 90%?

In a healthy ecosystem, the primary implementations do not represent most of the clients and servers.

HTTP for example is healthy because there are many http servers, many http clients, they're all pretty compatible and replaceable, and none of them represent the primary implementation with its own off spec quirks that others need to take into account.

The obvious answer is that it would be jack-of-all-trading, and others would focus on one area, do it better, perceived to be better, or just differently.

> What is a good reason for Element to relegate itself to only 10% of its own market and allow other vendors to pick up the remaining 90%?

From a purely capitalist perspective, 10% of a $1TB industry is better than 100% of a $10B company.

A growing ecosystem where the largest participant is 10% of the userbase is one which is immune from antitrust concerns.

It's likely one that is responsive to user needs.

It's everything that capitalists say that they want: a healthy competitive marketplace.

> effectively replacing the PSTN & email industries

I could see a path to Matrix replacing email (and it's the only one of the current candidates I would be happy with doing that), but how could it end up replacing the telephone network? Does it have voice capability?

Matrix does support voice chat. I use it every day (and it's been pretty good.)

Did you measure the latency? Latency should be below 100ms to have a pleasant conversation. Talking over a dedicated low-latency app like Jamulus is much nicer than the enormous latency in video calls.

I have not measured the latency but if both participants are on wifi there is usually no perceptible latency with my participants. (US -> UK calling daily)

If either of us is on mobile, sometimes latency can creep in, but this also happened on whatsapp too. That's the nature of mobile IMO.

Yes. I haven't used it so I can't tell how complete it is. But it exists. https://matrix.org/blog/2018/02/05/3-d-video-calling-with-ma...

Yes there is VoIP capability. And video as well.

_

what are you talking about? Are you confusing Parity with Protocol Labs? They are completely unrelated companies (other than both beginning with the letter P :/)

Ah ok, my mistake, sorry.

Rocketchat also has the open spec.

On 'loads of completely independent implementations': how can those independent developers stop Element from pushing some hypothetical changes to the protocol you would want to introduce against their wishes?

They'd object, either while the MSC was being drafted, or during the FCP (Final Call for Proposal) period. If the Spec Core Team voted to merge something which the community didn't believe was aligned with Matrix, then they'd appeal to the board of the Foundation (i.e. the Guardians), who are legally obligated to ensure that the only things which get merged into the spec are aligned with the stated mission and values of the project (as per the Mission and Values sections of https://matrix.org/foundation). If the Guardians fail to adhere this, they would complain to the UK non-profit regulator, and the foundation would get fined/dissolved. The recourse of the guardians would be to either get the Spec Core Team to change their mind or to replace the Spec Core Team to get back in line with the mission. The process for doing this is laid out in the Articles of Association of the foundation: https://matrix.org/media/2019-06-10%20-%20Matrix.org%20Found...

So, all the safeguarding positions are predominantly filled with Element employees (according to another commenter)? [1]

Consider me not impressed.

[1]: https://news.ycombinator.com/item?id=27970233

No, those are not the safeguarding positions; the Spec Core Team are the folks who approve PRs to the spec. Originally they were 50/50 element and non-element, although it's drifted thanks to members of the team choosing to work at Element so they can work on Matrix fulltime.

The safeguarding positions are the directors of the Foundation (aka Guardians), which is very deliberately majority non-element (3:2).

Why not spend your time improving XMPP rather than complaining about Matrix?

I'm criticizing matrix because it is a product of NIH syndrome [1] that should not have existed in the first place.

If you wanted an open federated protocol, you could have spent your time adding to an existing federated network. Everything you do with matrix is easily possible with XMPP. Instead, you want a protocol fully controlled by you. I oppose this goal and spread of such products.

And yes, I do spend all my time improving XMPP.

[1]: https://en.wikipedia.org/wiki/Not_invented_here

XMPP failed for normal users. When I switched to matrix years ago it was because the XMPP ecosystem was totally inconsistent. The best client was on android but no desktop clients integrated properly with it (send files, encryption, roaming) especially across multi OSs. I moved to matrix in the early days and all those problems went way.

What you see as NIH I see as building something from the ground up that works better than what was currently available.

No. Matrix could just build Xmpp clients for all platforms consistently, and it would work out much better than building their own atrocious server. It is a very typical NIH-approach.

Also, what you call a totally inconsistent ecosystem is what real ecosystem looks like. Matrix looks better in this regard precisely because it is not a protocol. It didn't yet deserve to have such problems.

I also don't think it'll ever be able to successfully transition to a real ecosystem because Arathorn here always brushes off concerns about governance of the protocol, which shows that he doesn't have the mature understanding of the nature of the problem.

Unless something has recently changed, XMPP also failed on the server side for normal users:

- Finding a practical provider is nearly impossible, because all of the major ones closed their public XMPP service years ago, and most people don't know what qualities to look for in a small provider, let alone understand why those qualities matter.

- Even if they get help from a friend and manage to find a reliable server that supports the magic combination of XEPs that are needed to make things work, chances are that server is no longer accepting new users.

- Even if they manage to get an account on a featureful[1], trustworthy, stable, free[2] server, many don't last more than a few years, so there is a high chance that the user's new online identity will evaporate before long.

[1](Required features include end-to-end encrypted group chats with offline delivery, among other things.)

[2](Yes, free. That is critically important for mass adoption today.)

I'm not a normal user, in the sense that I have far more technical knowledge than most people who use messaging services. I actively used XMPP for at least a decade; probably closer to two. I tried last year to piece together an XMPP-based solution that my friends and family could use, but even I was unable to solve all the problems I found in a way that would work for normal users.

And then there's the client side, which is also problematic, as soupbowl said.

This has been the state of things for quite a few years. The only people I still see pushing for XMPP are people who have major blind spots. Some of them have experience only with small groups, and have not considered how a whole world of people conversing presents different challenges. Others (usually tech folks) have had a good XMPP provider and comfortable client software for so long that they have forgotten how much knowledge it takes to get set up in the first place. A few are people who run small for-profit XMPP services or develop XMPP software, aka conflicts of interest.

Dear Jabber, I'm sorry. I liked you. I had high hopes for you. I might even still consider you for niche purposes. But as a general purpose messaging network for regular people, you have failed. It would take a lot of effort and resources to make you viable, and I don't see any sign of that coming. The closest you came was when Google and Facebook embraced you, but those days are long gone.

Meanwhile, Matrix is succeeding. The rough spots are being fixed. The bloated clients are getting competition. The missing features are appearing. Even as a work in progress, it is already usable for some of us, and it's consistently moving in the right direction. Importantly, I can tell a regular person where to get it and (often with no help) they can be chatting with me in a few minutes.

It's permissively licensed FOSS. If the primary maintainers go against the community and can't be reasoned with it will be forked. If a fork doesn't end up having enough momentum to maintain it then perhaps the original wasn't actually doing such a bad job after all.

This issue is hardly unique to Matrix.

Not really. You are ignoring the network effect. The vast majority of the matrix network, including the by far the largest node, is using the software developed by Element. If they change something, any competing spec will lose access to federation, and, subsequently, users.

Spot on. When one organization controls servers, protocols and main implementation it creates lock-in just like closed source software.

Signal is a good example.