Overview of Proposal
Rocket.Chat is not awesome.
Formal Proposal(s)
We should consider a new platform.
Consider:
- Gitter
- Already popular in ETH community
- Uses GitHub ID instead of LFID
- Would not need LF to run another server
- Matrix
- Recently selected by Mozilla to replace IRC
- Open Source, would need to pay a SaaS or LF to run the server
- Mattermost
- Would need E20 license for SAML for LFID, $8.50/user/month (open source plan is E10)
Action Items
- Converting a new chat system and preserving links will be a significant lift for the CA team
- Making sure people's old links don't break
- Making sure it works properly with LFID will need to be supported by LFIT not the CA team
- Integration with the new LFx platform again would need to be supported by LFIT and not the CA team
- How to transfer the Data or archive the current data
- Feature by feature comparison and spreadsheet?
Reviewed By
- Angelo De Caro
- Arnaud J Le Hors
- Christopher Ferris
- Dan Middleton
- Gari Singh
- Hart Montgomery
- Mark Wagner
- Nathan George
- Swetha Repakula
- Tracy Kuhrt
- Troy Ronda
9 Comments
Angelo De Caro
Which are the alternatives other than slack?
Silona Bonewald
Personally I would prefer mattermost because we have been running it personally on our home server for 4 yrs now with few problems other than migrating from Slack. But that move is a big ask of my team and for transitioning the community. It will be even harder than our wiki migration at the beginning of the year
https://chat.hyperledger.org/channel/rocketchat-chat?msg=gRiZWjwYBjD93gM3b
Ry Jones
We're trying out gitter - Iroha and Consensys use it; we have a test room for Besu https://gitter.im/Hyperledger/besu
Bob Summerwill
Ry Jones has suggested that I need to contribute to this page, and he is RIGHT.
In the meantime, see the https://chat.hyperledger.org/channel/rocketchat-chat for some "commentary" from me.
I will try to get my thoughts down in the body of this page ahead of this week's TSC meeting.
greg misiorek
have only been familiar with Slack, but whichever tool is chosen we should have the ability to search all the comments at all times, scrolling pages and pages of even recent interactions is suboptimal.
Mark Lewis
There isn't enough information captured in this page for me to understand the issue or what we are trying to achieve by switching to a new platform. I think I've got the gist by trawling through lots of discussion in https://chat.hyperledger.org/channel/rocketchat-chat but it seems like the key requirements for the messaging platform should be captured here so they can be used to assess options.
Ry Jones
I'm OK with withdrawing this proposal. I filed it because people were asking, and my preference is to capture those asks in public
Christopher Ferris
Yes
That said, I think that to Mark's point above, that the requirement is to move off of LFIT provided services. We currently require LFIT to host RocketChat. Gitter might be an effective alternative, but I cannot speak for any pricing, etc. I know that Besu was using it, possibly Grace or Danno could comment on its eficacy and so forth.
Frankly, RocketChat isn't great.
As for migrating content, I would advise against that. Archive what we can as Ry did for Slack back in the day and move on. The search function isn't that great anyway for finding information. StackOverflow is far better for getting questions answered asynch.
Gari Singh
I feel like we made a big mistake trying to leverage a chat system for historical information.
We could use the free version of Slack, for example, if we were not so concerned about history.
Leveraging chat to answer questions for historical purposes is simply flawed. If people want to find some help quickly, then chat is OK ... but people need to understand it is at best "best effort".
If we don't want to push everyone to StackOverflow, then we should probably use some type of forums.