Farhan's Personal and Professional Blog

Let IRC Die

For the sake of progress, let IRC die!

IRC was great, but its continued usage is a relic of an older internet, when storage was expensive and servers were expensive. At this stage, communities should migrate off of IRC to slack-like solutions like Mattermost or Matrix.

IRC served as a workhorse for communication, but it has failed to adapt to modern demands. In particular:

  • Offline sending and retrieval
  • Referenceable conversations (ie, URLs to a conversation)
  • Editing messages
  • Formatted text (including media)
  • True decentralization

Many of these problems can be remediated at the client-level, but not without a cost.

For example, consider the missed message problem. You could solve this problem by using an IRC bouncer or getting a persistent VPC. But in both cases, you are required to use another machine, which requires power usage and IP space. Now realize that this solution will be replicated for each individual user of IRC.

Another solution is message history. While there are mitigations of this, such as servers that implement rudamentory message history also works, it is not a native IRC solution, resulting in a clunky experience, such as the timestamp printed as part of the message. Additionally, most modern IRC clients, such as irssi, weechat or quasel, do not natively support commands to retrieve more history and require the user to use /QUOTE. Again, a clunky experience.

While pointing out these problems, many have suggested XMPP. While XMPP is a good alternative, it is still plagued by many of the same legacy problems as IRC. Looking at the market, we can already see that industry has selected solutions such as Slack or Teams. Open source versions of this are already available and used by tech communities, and the Open Source technical community would be wise to do the same!