Shout – A web IRC client
github.com
github.com
- It stays auto-connected with no configuration.
- It offers a great desktop (web-based) UI, including embedded images, videos, tweets, pastebins, and more.
- It offers fantastic mobile apps for both iOS and Android, for both phone and tablet form factors.
Thanks for all the attention! The repository is here: https://github.com/erming/shout
It's open source, so if you're looking for a cool new project to work on, feel free to help out!
This is a really good looking web IRC client and it works so well with notification sounds, favicon updates etc.
The slots in this case are the amount of connections that can come from the demo server (20 at the moment).
If you ran this software on your own server, you wouldn't run into the problem unless you tried to open 20 simultaneous connections to 1 IRC network.
Can I ask why there is a 20 slot limit? How can one increase this?
I still prefer Convos for all its built-in features and great community.
Otherwise I see 4 open source web-based IRC clients with very few regular contributors each. And the only really good IRC web client will still be IRCCloud (which is proprietary).
After that, it's just a matter of implementing few plugins for popular IRC clients that would improve the UI of their respective clients to support those new features.
If anyone is going to start such endeavour, I'm willing to put some man-hours (starting next month) into writing plugins for ERC and Weechat.el, and maybe help with others.
Pasting code snippets, preserving formatting and getting syntax highlighting. Pasting screenshots, usually of code, error logs, or graphical bugs. Inlining code with `def func_name()` or whatever, making things much more legible. Easily sending files.
Basically all code manipulation at their root.
For actual normal communication, IRC works just as well for me.
Even if the IRC client supports some of theses features, the protocol is just text, you can't add meta/structured data to a message.
Those are all features of the client UI, not of the protocol. All can gracefully degrade to plain text for people whose clients don't support them. Image/video/URL preview is an UI feature over plain link. Tables... those you can either send as ASCII-art tables, or upload to whatever paste service you like (or host) and use the same preview mechanism to show a nice table on your rich UI client.
I find the message history mixed with images in-line a bit annoying.
The person I was replying to inquired as to why uploading the image and pasting a link was a problem. He said:
"Limechat shows image links inline, which is good enough for me anyway. In some ways it is even preferable to tying things to one particular storage solution like s3."
Hipchat:
1) Paste image
Imgur + IRC:
1) New browser tab/window
2) Paste image
3) Click "Start Upload" button
4) Wait for upload + redirect to new page
5) Right click, copy image location
6) Switch back to IRC
7) Paste URL
I use this somewhere from 10-100 times a day, so it adds up. And it doesn't interrupt the flow of whatever I'm working on, just a single action that doesn't require any thought or aiming my cursor to click on things.
Could I write an automated tool that takes my clipboard, uploads it somewhere, and replaces the clipboard with an image URL? Probably. Or I could use a chat platform that supports images.
Edit: Trimmed one step with copy image location instead of actually navigating to the image
Which is exactly that automated tool, but already written by someone else and embedded directly into your UI. It's not that those images magically go over the wire straight from your clipboard without being hosted somewhere (especially that, AFAIR, Hipchat is XMPP-friendly).
I do understand why Hipchat, et al. are getting more and more popular - I was under contract for few weeks at a place that used Hipchat as a company-wide communication tool and I admit it was super-convenient (as long as your machine had enough free RAM to run it, which wasn't always true; for a chat client, this thing is heavy). But the thing is, IRC could easily be on par with Hipchat-like solutions if someone wrote proper UI extensions to popular clients.
Of course until someone actually goes and does that, we'll be stuck with Hipchats.
Oh I'm aware that it's not actually a magical feature, it just feels that way from a user interaction perspective. They all go up on S3 and have a text link, which gets an inline picture just like other image URLs.
From what I've heard about IRC, there's not an actual well-defined spec anywhere that all of the clients follow, which I'm guessing is part of why there aren't a lot of better IRC clients being written. It's easier to come up with your own protocol where you understand how it works and know exactly how every client will behave.
I do wonder why there isn't an open source effort for fancy features on top of XMPP. Not enough interest? Most developers don't do visual work and see IRC as "good enough"? Or it does exist and I've just never heard of it?
> 1) Paste image
Then the history if filled with stupid crap and you can't find any useful stuff in it. Stupid text can be easier to jump over, images spanning the whole page can be more annoying.
I don't see how a giant uninterrupted wall of text going back to the beginning of time is better.
HTTP and TCP are just plane text too. Its what you do with that text that matters.
Somebody just has to find a way to shoehorn it in. I'd propose using a basic system built on top of the basic AK meta-protocol for serial communication.
[1] https://github.com/ircv3/ircv3-specifications/blob/master/sp...
https://github.com/ircv3/ircv3-specifications/blob/master/sp...
Inline images are annoying and I don't like that it uses non-standard protocols.
I also understand that Hipchat is not open source completely, and if so I don't like that.
Notifications are part of a client.
Permissions for large teams are nice, that is a useful feature I would like.
On the plus side, the Slack Android app is pretty good. Much better than any IRC app I've used.
https://github.com/glowing-bear/glowing-bear
https://play.google.com/store/apps/details?id=com.glowing_be...
Why reinvent the wheel when you can let someone else do the heavy lifting? ;)
I'm not trying to take anything away from these excellent projects though. Just saying IRC bouncers aren't any harder to install than these projects
1: hide your IP (moot point if you're hosting your own bounce though. And a lot of IRC networks hide your IP for you anyway)
2: offer a way of keeping your history
3: store your unread messages server side.
4: keep all your clients and states synchronised
I'm not saying there's anything wrong with a web client (in fact I'm tempted to try the one submitted here). But I think there's a lot of misconceptions about bounces too.
In an ideal world I'd use both; have a web client connect to ZNC. So I can log onto the IRC client when a guest on someone's PC (eg cybercafe) but also have the ability to use native clients from the same nick for when I'm sat on my own laptop or on my phone (there's some half decent native Android IRC clients)
(Sync from WeeChat to Glowing Bear and between multiple instances is something we're working towards with the developer of WeeChat)
I don't want to run my bounce in a detached screen session, I want a proper daemon. It's easier to install and manage if it's a proper daemon. And if I'm going to be using a terminal based IRC client (which I often do) then I'd be running Irssi.
I don't want your stupid halfway solution when I already have a proper IRC bounce set up that does all that and more.
I really don't understand this. What features does IRC lack that Hipchat has? AFAIK, the advantage of Hipchat is the branding, the packaging, and the support.
The one thing I see ("image previews") is a feature of the client, not IRC itself. And many clients used to offer that, way back - the reason they turned it off was because it was a great vector to (e.g.) goatse someone. In any GUI-based IRC client, this is dead simple.
Most of what Hipchat seems to provide is branding and support, and I get that that's very valuable, but I don't see why that needs to be viewed as an either/or situation.
Never forget the "Show HN: Dropbox" and its immediate technical rebuttal [0], see where it is now.
Oh and Hipchat seems to do screen sharing, which AFAIK is not possible with IRC.
It is a Bouncer with full PostgreSQL database of the whole log, full search functionality, a shared history between users of the same bouncer, etc.
There are clients as a website (runs on node.js), for Windows, Mac, Linux, iOS and Android.
So, if someone would add the missing features to Quassel, we could make that one the new new HipChat.
There's also a /quote command.
Edit: wait, that's for password protected channels, right? I need to pass a password through on server connection otherwise my ircd boots you off and kills the socket :(
Feature request: Make it work with quassel core. Currently there is https://github.com/magne4000/quassel-webserver but I don't like the UI as much as yours. https://github.com/magne4000/node-libquassel might help you with quassel connectivity.
You can do some amazing scripting stuff and really customize it.
There are a lot of custom versions of mIRC online with crazy color themes and what not.
A suggestion - (I know I can just use my own CSS but,) there are low contrasts in the default theme. I usually have very good vision but for me it's pretty hard to distinguish the green from the white background. Maybe you could edit the default colors a bit?
I'll be forking shortly. Thanks!
However, I added "public" mode, where anyone is allowed to connect. And the demo server is currently running in this mode.
So right now? Nope. I'm not limiting anything. I guess this will be up to the person running the server to decide.
Thanks!
For example, I run a web-irc instance for the people on the Montreal mesh network (which runs mostly on private/unique-local addresses).
Looks pretty neat - even setup to work decently on mobile devices.