> I know this makes me sound old, but I really wish popular open source projects would just stick to IRC as it just works and doesn't require any proprietary client (and isn't trying to turn into video chat, screen sharing, etc.) Or I suppose I wish Slack would make its ecosystem available via IRC.
I'd say you'd be far from the only one. The thing that Slack really brought was an improved onboarding experience, but I have a (probably somewhat vain) hope that protocols will win over corporate controlled walled gardens.
Matrix[1] / Riot[2] are really promising and I really hope we settle on something like that sometime down the road.
We recently did a test run for Mattermost to see if it's a better fit for us than Slack.
This is a note to others who may be considering the same. Specific needs will vary greatly across teams/companies.
We found Mattermost web client to be ok, however phone apps and integration coverage are both poor by comparison to Slack. We also found running, securing, and maintaining our own server to be roughly as or more expensive than using Slack.
I'm hopeful the Mattermost project keeps momentum to improve. Currently though it doesn't quite seem in the same league.
This will likely always be the case with open source chat solutions.
Developing clients is hard. Especially good UIx and complex feature sets across a zillion platforms like Slack does. A lot of their day I bet these days is simply tracking down obscure bugs on certain platform combinations and trying to keep feature parity between everything working properly. I think most of us can stipulate Slack does a decent job there - or at minimum is far better than the competing (open) solutions.
I really wish there was a Slack alternative where I could purchase a really nice unified frontend and simply connect to an open protocol backend such as IRC or Jabber or whatever. This was the original intent behind IRC as well - back in the 90's there were a number of shareware style IRC clients developed for Windows - and you saw some interesting protocol hackery to make some more graphical features work.
I have yet to see open source really truly compete in the "client" arena, save for an exceedingly few notable exceptions. However, it excels at the backend where you will have developer interest and competence - and the long tail of possible bugs (and user competence) is a much more sane problem to tackle.
No, I do not expect this to ever exist for obvious reasons :)
Our focus is helping increase productivity for companies who need total control of their infrastructure. Thousands of companies use Mattermost for that purpose and we're excited about bring them more and more value in each of our monthly releases.
Might still be a good alternative to companies that want Slack, but cannot afford to pay for it? (Assuming they're not finding what they need in the free version.)
We found Mattermost to be more expensive than a paid version of Slack for our team of 16 after factoring in the costs of hosting, maintaining, and securing our own servers. It shifts the engineering cost burden onto your team.
As others point out, Mattermost vs. Slack is more a discussion of self-hosted vs. 3rd party cloud-hosted.
Matrix is really wonderful and I hope it will continue to grow. It's federated, like IRC and unlike Slack. Also there's good support for bridges to other chat protocols / services, so matrix may be the solution to _reducing_ the number of chat clients you need to interact with.
I'd say you'd be far from the only one. The thing that Slack really brought was an improved onboarding experience, but I have a (probably somewhat vain) hope that protocols will win over corporate controlled walled gardens.
Matrix[1] / Riot[2] are really promising and I really hope we settle on something like that sometime down the road.
[1]: http://matrix.org/
[2]: https://about.riot.im/