Aside from it's zealous following that the others never enjoyed, Slack is also becoming a true platform. None of the other services had so many companies building integrations on top of them.

> None of the other services had so many companies building integrations on top of them

I disagree. While there is no centralized list of IRC bots (because IRC isn't centralized), there is an uncountable number of IRC bots and libraries for people to make their own.

"companies", not services. Aside from tech companies, IRC was never adopted by wider industries. We've been using it amongst a primarily operations-based company which relied on Email and Lync previously.

It helps that Slack is spending $40M (plus another $40M from their investors) to encourage that.

http://www.entrepreneur.com/article/254078

But WHY did so many people decide to build Slack integration? It's not like HipChat and Campfire didn't have APIs. I say follow the money. My theory is that Slack has influential investors who were extremely successful in selling hype.

I've already seen groups moving on from slack to other services.

I still really do not get why technical people don't just use IRC with a decent client (Hexchat?)

Here's my theory on why most people use this sort of service (see also: Skype, Gmail, etc.) rather than a technically superior, independent, non-client-specific service: no server names to remember.

Want to set up an IRC client? Sure. What's the server and port name? Got firewall rules set up? That's complicated, and easy to get wrong. Your own IMAP email service? Same.

Want to set up Skype? Just sign in here. Want to set up Gmail? Just sign in here. Want to set up Slack? Just sign in here.

No regular person used a chat client until MSN Messenger and Skype came along. Nobody. Now they all use the messenger service built in to Facebook, and they don't even think about it.

It's just way easier, even for technical people.

> No regular person used a chat client until MSN Messenger and Skype came along. Nobody.

AIM?

That was the heavy hitter. jen is still right that simplicity and ease of use are incredibly important. Especially that forcing anyone to work with server IP's, ports, and firewalls leads to instant negative reaction. Gotta make that disappear somehow.

Fortunately, there's a lot of schemes to do that with untrusted 3rd parties. Need more work and polish w/ OSS implementations.

No regular person ever renames files. No regular person ever uses a browser that didn't come with their system. No regular person ever properly navigates an average website without help. Your point is?

> No regular person ever renames files.

But you can still use a file that you haven't renamed.

> No regular person ever uses a browser that didn't come with their system.

But the browser that comes with your OS is a perfectly serviceable browser.

> No regular person ever properly navigates an average website without help.

Nonsense.

> Your point is?

That your average person - let's take my best mate, a pro photographer and perfectly competent human being with few technical skills - can't/won't/doesn't want to use services that require complex set-up.

Whenever she moves to a new Mac and needs her email setting up, you know what she does? She calls me. She has literally no clue what her email server settings should be. If I told her, she wouldn't remember. The detail means nothing to her.

You know what she doesn't ask me to? Sign her in to Skype or Facebook.

That's my point. What was yours?

My point was that the 'average' computer user is easily bewildered by many things, and using an IRC client isn't especially difficult compared to them. Using a link to a ready-made config for a web IRC client is easy and relatively painless and doesn't require any additional magical insight and fairy dust by a group of Silicon Valley dreamers. The use cases for Slack is arguable an Euler diagram circle exactly on top of those for IRC, a slightly different one than those for FB Messenger or Skype. Slack is redundant/bloated because of IRC, Messenger isn't.

Yeah I just think - and this is just off the top of my head, I have nothing but anecdote to back it up - that as soon as you reach that certain level of "config", their eyes glaze over. That level seems to be when you start talking server names.

People just don't understand what that is, let alone how to actually use it. But everyone seems to be able to figure out how to sign in to an app with nothing but their email address and the same password they use for everything else, and so there the line is drawn.

Again, nothing but conjecture. It's just something I've come to realise over the years. (FWIW, I hate Skype etc. Horrible! But, I understand why Jane Public uses it.)

because businesses don't have only technical people, and you don't want to have multiple chat systems, which will eventually bring worst communication between different groups of people. If technical people are on IRC and all the other people are on hipchat, how do you make the 2 groups communicate?

Plus, the fact that someone is technical does not always mean they want to fiddle with getting chat systems working. None of the engineers in my department would be willing to mess with IRC when Hipchat and Slack provide turnkey solutions that just work, the first time you open them, and require no configuration or special treatment.

I would imagine most non-technical people could scrape by with a certain client, even a web one. As someone who's on Freenode, Lord knows they do.

Usability. I don't think Hexchat is "good enough". Slack is vastly superior from a usability perspective in my opinion. I don't have to use commands for the most part, I can customize my interface as much as I "need" to and there's no setup, also there are mobile apps that "just work".

It's almost like what OSX/iOS is to Linux. Sure Slack has a particular way of doing things, but they take care of all the hard, annoying shit you don't feel like doing.

Konversation is really nice and cleaned up, runs on every OS like HexChat / XChat does. Slack has a lot of sugar coating compared to any IRC client I've ever used, and I'm surprised nobody's beefed up an IRC client yet to "compete" with Slack. Some clients have really nice features though.

Because a new IRC client without widespread protocol extensions means everyone currently using IRC will hate people using the new client due to the fact there is no place to hide metadata.

See: Microsoft Comic Chat

Not every new feature has to be a protocol extension though? Some clients just lack some sugar coating client-sided features. As for protocol extensions, you could always just disable them in-client on servers that don't support said extensions if anything. But my focus is on making more cleaned up IRC clients, a lot of them feel like they're stuck in time when I know there's lots of things that could be done to them to both make them user-friendlier for newcomers and make them a little more convenient to use.

It makes less sense for open source. Ember would have to pay $30k a month to archive slack. But, all the Ember devs are on slack and irc is a grave yard.

Imo, having logs makes irc demonstrably better.

Recently I was talking to a co-worker about this and I said I prefer IRC over Slack and HipChat. You can certainly build robots on IRC and make fancy automation. You can also use IRCloud if you prefer someone hosting it for you. I came from a company where IRC is literally the primary communication channel for both business and technical staff (the word starts with the letter M). When we did townhall we all jumped on IRC and we asked questions from there. We use other software for conferencing and video calls. We are okay with that.

Here at my current company we have a big challenge. We use Skype for inter-team communication, Webex for grooming and remote conferencing. We also looked at Lync (aka the stupid lame ass 'Skype for Business'). While Skype allows you to keep history (I can go back to 2 years ago!! how amazing) and create groups, but I have to know everyone first. This is where IRC-like channel is really powerful. Everyone can join some lobby and then we branch off into individual team channels. With Skype I can't. Oh good luck finding the group you forgot to favorite 3 months ago. Also, I can't build nice robot in Skype and make pretty command to automate our infrastructure! There used to be restriction with how many people we are group video chat before we have to get into paid plan (not sure still apply or not). With everyone at different location and with different operating system, colleagues would have problem accessing or using Skype / Webex, espeically our Linux users, and they have to boot up a Windows VM. What a nightmare chore.

What if we have a tool that can do

* text chat

* video chat

* voice chat

* integrate with Jira / ticket system

* mention people and actually MENTION people

* create a large lobby

* integration with active directory(?)

* have a decent API for automation

* cross-platform

* safe and secure

* integration with other third-party services/tools that we have signed up (Box? Dropbox?S3?)

* record things?

* history

wouldn't that be awesome?

I am excited about trying out Hipchat and Slack. The problem is, Hipchat is the only one out of IRCCloud and Slack offering on-premise installation (we do have onsite Atlassian). The downside is we'd managing additional infrastructure, esp if I want to use the video & voice feature. Oh.

If I have to choose I prefer Hipchat or the Campfire style (they are pretty similar in terms of style). Slack - a lot of people like it, but the UI is just so awkward and clumsy and crowded. I don't like it. I don't know. It hasn't changed since launch.

Nice write-up. Made a note of those requirements in case anyone asks for a project or startup idea in collaboration. ;)

Personally I went from IRC to Slack. As much as I like the IRC model (decentralized, open protocol) it just lacked so much that even the best client could not fix. Practically speaking, Slack is actually really good. I enjoy using it. I think the progression you are seeing is just an IRC exodus followed by people looking for an minimum viable solution, which may or may not end with Slack.

reason I love slack: I never see the android client in the top consumers of my battery. I'm being forced/fighting converting to hipchat, and if I let their client run on my phone, it's regularly in the top battery consumers. Oh, and notifications on my phone work great. I'm not sure how no one else can make this happen, but it appears to be the case.

Desktop client or web client + animated gifs = my cpu fan starts spinning.

I do. It's so ridiculous to me that people keep comparing it to IRC. It isn't IRC.

ICQ -> MSN -> Facebook -> Snapchat

Facebook's userbase is still going strong though. Facebook also has a moat, since you cannot just move, because Facebook is where your friends are. With Slack, my company can decide to move and there are no other dependencies on anything external. We just move and that's it.

The moat is the integration points. The real differentiators for Slack are centralized history (better than IRC) and easy web endpoints for integration (better than pretty much anyone else). After a company invests significantly in a myriad of integration tools (both external and homegrown), they have little incentive to move. That's Slack's moat.

Still, they have no real presence in the consumer space and that makes them vulnerable to the first player that can really cater to both businesses and consumers. As it is, consumer vendors (MS/Skype, Facebook, Google) are fumbling in the biz space, but that will not last forever.

Edit: have a look at Slack 3rd-party apps for what i mean by integration. Because it's so easy to build them, there are tons.

When you say integration, I assume that goes beyond the chat bots that we have seen on other platforms since ages?

You forgot Yahoo Messenger and AIM. Before Snapchat it was iMessage, WhatsApp (still extremely popular).

ICQ -> MSN & Yahoo Messenger & AIM -> SMS (sidekick phone!) -> iMessage & Facebook (the age of iPhone) -> WhatsApp -> Snapchat

Hey, 56k was a real thing ;-)

x86 Assembly -> C -> C++ -> JavaScript

edit: Okay, I get why I'm being downvoted. :) But I just wanted to say, apparently in a vague way, that evolutions happen. People don't invest their precious time just because something is "hip". Apparently they solve a problem of their predecessor.

C++ to JS is a pretty huge jump. C++ -> Java -> JS might work better.

Hey you miss PHP, Ruby, Python. The new hotness is called Rust.

Don't forget the apparent JS -> Go

More like LISP -> ML's -> C++, Java, C#, Ocaml -> Rust, Julia, Clojure, and the other better ones. :P