Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We Tried Slack and Regretted It

blog.freecodecamp.com

Post reply on HN
Print view View original

101–110 of 192 posts

Re: We Tried Slack and Regretted It

#101

Post by chx »

michaelq wrote: post #74
derefr wrote: post #24

It's pretty clear that Slack is not, and never will be, built for this use-case. Slack is for teams : small groups where everyone knows one-another by name, can be trusted with one-another's email-addresses and other contact information, can be trusted to only use @everyone triggers for important things, etc. A lot of Slack's features are built to assume this "small group with a shared purpose where everyone can be t…

I authored this blog post. There is a lot of merit to your criticisms of my decision making. I want to point out that communities are increasingly using Slack, and many of them are also in the thousands of users. Slack does nothing to discourage this, aside from posting warnings about archiving messages. The real problem is that they have an undocumented user limit. Like I said, I'm pretty sure we're the first commun…

There's a golang http://blog.gopheracademy.com/gophers-slack-community/ Slack channel. Not 100 000s for sure but there are a lot of people using Go.

Re: We Tried Slack and Regretted It

#102

Post by Lazare »

voltagex_ wrote: post #87
mtrpcic wrote: post #27

Earlier quoted context omitted.

I agree that IRC would have been a good fit, and with modern IRC clients like LimeChat and Slate, it's not the "legacy" experience a lot of people remember. IRC also would solve all of the problems this guy faced: * Free. You can hop onto an existing tech-focused server like Freenode, or if you want your own control, throw it on your own server(s) for pennies. If you do the latter, you have full control over all logg…

The slow demise of IRC is sad. Clients like WeeChat excel on the text side of things, but where's the knockout web, desktop and mobile client? Something with the UI of Gitter would go far. I always feel like IRC is a few protocol extensions and a few clients away from being relevant again. /me goes back to Freenode

I think IRCCloud has a UI significantly better than Gitter. (Then again, I'm not a fan of Gitter's UI, at least as far as their desktop OS X client goes...)

Re: We Tried Slack and Regretted It

#104

Post by ksk »

coldtea wrote: post #61
ksk wrote: post #45

OK sure, I get that Slack wants to make money and doesn't want too many 'free' users. The restrictions themselves are acceptable to me. However, their platform seems super brittle and janky if it can't handle the users in the first place. Seriously? 5000 users, 10,000 messages is a "Use Case" now? Sending out emails without fucking up needs "engineering"? Um. OK. Personally, I would be embarrassed if I put my name on…

> Seriously? 5000 users, 10,000 messages is a "Use Case" now? Slack has millions of users and billions of messages -- working alright. 5000 users were seeing slowdonws for a SINGLE team -- not across Slack. And 10.000 is a limit put by Slack intentionally in place as a freemium boundary. All systems have breaking points, especially if they weren't designed for some specific use. You might as well ask "Twitter can't h…

I said that I would be embarrassed. Your opinion does not change that. You obviously seem to have lower expectations than me.

Re: We Tried Slack and Regretted It

#105

Post by jsmthrowaway »

voltagex_ wrote: post #87
mtrpcic wrote: post #27

Earlier quoted context omitted.

I agree that IRC would have been a good fit, and with modern IRC clients like LimeChat and Slate, it's not the "legacy" experience a lot of people remember. IRC also would solve all of the problems this guy faced: * Free. You can hop onto an existing tech-focused server like Freenode, or if you want your own control, throw it on your own server(s) for pennies. If you do the latter, you have full control over all logg…

The slow demise of IRC is sad. Clients like WeeChat excel on the text side of things, but where's the knockout web, desktop and mobile client? Something with the UI of Gitter would go far. I always feel like IRC is a few protocol extensions and a few clients away from being relevant again. /me goes back to Freenode

I'm happy IRC is dying, because it's a giant attack surface. As a hosting provider sysadmin, dealing with IRC servers was among the absolute worst parts of my job -- mitigating DDoS attacks against them, extensive hacking attempts directed at their infrastructure and later, mine, when they couldn't get in, credit card fraud associated with the accounts, IRC networks designed specifically for command and control of botnets and trafficking in child pornography, FBI raids against said networks we hadn't yet discovered...

There are exceptions, like OFTC, but merely waltzing into the wrong place on EFnet these days is enough to get 10+ Gbit of UDP traffic directed at your IP address. I used to be fairly okay bouncing on a vhost, but now even the 12-year-olds have enough traffic to knock over a 40 Gbit port channel. An increasing number of datacenters are just filtering 6667. I'm on board with that.

Seriously, let IRC relegate itself to the dark corners of the Internet on shitty hosts and stay there, IMO.

Re: We Tried Slack and Regretted It

#106

Post by ksk »

fleitz wrote: post #53
ksk wrote: post #45

OK sure, I get that Slack wants to make money and doesn't want too many 'free' users. The restrictions themselves are acceptable to me. However, their platform seems super brittle and janky if it can't handle the users in the first place. Seriously? 5000 users, 10,000 messages is a "Use Case" now? Sending out emails without fucking up needs "engineering"? Um. OK. Personally, I would be embarrassed if I put my name on…

Have you ever thought that they put those limits in there precisely so that situations totally inappropriate for slack don't occur? I love how programmers are always "embarrassed" about the limitations of billion dollar companies, have you ever though that these limitations might be vital to their success?

>I love how programmers are always "embarrassed" about the limitations of billion dollar companies,

In my mind I picture a station wagon outracing a Ferrari. That would be cause for embarrassment. Your quotes seem to serve no purpose as far as I can tell.

>have you ever though that these limitations might be vital to their success?

You have much lower expectations than me. I don't see Slack doing something that is in any shape or form particularly challenging from an engineering standpoint.

Re: We Tried Slack and Regretted It

#107

Post by comex »

fredkbloggs wrote: post #85
slyall wrote: post #57

Earlier quoted context omitted.

I'm afraid not. eg for me the 3 killer features on hipchat have been (a) Ability to scroll back and see what happened when I was offline (b) Ability to post pictures/files (c) search. Now all 3 of these could be implemented on a ICR server client. But they are not simple, not supported with all clients and not something that your marketing guy can easily use. With hipchat/slack/etc they are all out of the box on minu…

(a) Use screen or tmux. Log. (b) Post links. Everyone in IRC has been doing this for 30 years. Works fine. (c) Have your client log, or have a bot log to something with HTTP access. You can then use grep or a fancy web app or anything in between to search. I'm with the parent. Scalable chat is a solved problem; it was solved in the open, via open source and open networks, over 20 years ago. Everything since has been…

> (a) Use screen or tmux. Log.

This is not something "your marketer guy can easily use". And even for techies it has issues like lack of notifications (fixable), lack of non-suck mobile support, and, extremely important for me even if apparently not for some, typing latency (mosh is a partial solution but not good enough for me).

There are a few applications that use a daemon running a custom protocol to fix the scrollback issue: Quassel, Smuxi, weechat remotes, others. This is a decent approach in general, and the one that I use (weechat + Glowing Bear), but I haven't found any such applications with good mobile support or which are very high quality in general.

Maybe IRCCloud is the answer - has a spiffy web client and iOS and Android apps. It's a hosted service and non-free so I am not willing to use it myself, but most people don't care that much about such things, and it's not like Slack is any better on that front.

Re: We Tried Slack and Regretted It

#108

Post by dsacco »

detaro wrote: post #49
fredoliveira wrote: post #25

Earlier quoted context omitted.

This is the real mystery here. I'm a huge fan of what Slack has done and was a big advocate for it when it was starting up. But it feels off that people with high realtime chat needs don't just use IRC, which has been around for ages, is super stable, and deals with this amount of users easily.

Is there any web client you can easily deploy that comes close to Slack in comfort for the users? IRC is cool, but getting thousands of users comfortable with it sounds like quite a challenge.

We run our entire real-time communication infrastructure on an internal IRC server with ZNC. We have literally everyone in the company on it in various rooms, from #backend to #support; engineers and HR alike.

I don't think it takes much to be comfortable with IRC, given the huge mix of people we have using it every day successfully. The IRC clients are really good at making users feel at home - Textual is set up for everyone in the company, although some people prefer to use Weechat.

Re: We Tried Slack and Regretted It

#109

Post by Lazare »

kolev wrote: post #91

I've always wondered why Slack gets so much love from engineers! It's noisy, it's expensive, and did I mention "noisy". It started as a copycat and didn't do much beyond that. The only thing I'm thankful to Slack about is that it's pushing HipChat to innovate. Finally we got multi-account support in there! Yes, I openly do like HipChat better - it's free, it's less noisy, it's coming from Atlassian, it integrates nic…

Having used HipChat and Slack, I prefer Slack as a much better designed product. It's more stable, less buggy, and has more features.

> did I mention "noisy"

You did, but I have no idea what you mean.

> it's coming from Atlassian

One more reason to prefer Slack. I've had a lot of experience with Atlassian tools, and it's almost all been negative. Everything about Atlassian is slow: Their products, their support, and the pace of development. Submitting a bug report to Atlassian is an exercise in futility; if it's critical and affects hundreds of people, you MIGHT get reply within 2 years; you certainly won't get a fix that quickly.

One bug I hit was opened in 2007, got an official response saying "maybe we could look into trying to fix that" in 2009, was officially added to their backlog to fix in 2010, and was announced as officially fixed in 2011. Spoiler: It wasn't fixed, and for all I know, it's still broken.

The worst thing to happen to Hipchat was being purchased by Atlassian; it's when they stopped innovating and started stagnating.

Re: We Tried Slack and Regretted It

#110

Post by neilellis »

I'm very disappointed with the level of comments here - the emotional attachment to Slack is just not neccessary; same for the ad hominem comments. For sure they have a great tool that I love and use but they have been slow to act upon the fact that many communities are using their free option with huge numbers. Rightly or wrongly.

Slack just simply need to decide whether they are supporting this model or are happy to hand it over to Gitter to deal with. I suspect Gitter will be more than happy.

Personally I think they could offer a fixed fee for public rooms. But hey that's up to them, it's their business model to decide.

I think the problem is just the lack of clarity about how they are going to manage what they definitely know is going on (I have asked).

Post reply on HN