Bushi.do is now Cloudfuji.com, launches Kandan - beautiful open source chat | Hacker News Reader

Bushi.do is now Cloudfuji.com, launches Kandan - beautiful open source chat

cloudfuji.com

129 pointssgrove60 comments

Please don't require credit card information when you advertise a link on that page as "Launch and use Kandan now". And especially not when that credit card information is required as the second step in an initially single-step seeming sign-up process -- and also, please don't hide all of this behind a big "Launch" button.

Credit card info is only required for the second app launch, or when you invite more than two people. We have to do that to manage our resources, but we definitely didn't want it to be the first thing you see. If you send me your account email , I'll bump you to a verified account manually so you don't hit that :)

Edit: You can send it at sean@cloudfuji.com or just click on the support link in-app, and you'll get a modal form to contact us.

If it's an open source app... where can I download the source to run it on my own server? Don't tell me it's open source, then ask me for a credit card number.

https://github.com/cloudfuji/kandan

http://kandan.me/

The beautiful part of Kandan is courtesy of the design by Sacha Greif. He wrote an ebook about how he went about designing it, and it was discussed here: http://news.ycombinator.org/item?id=3731135

The amazing implementation is from a lot of hard work on Kevin Zettler's (my cofounder) part, who's amazing at polishing apps.

Let us know your thoughts, and switch to using Kandan on Cloudfuji today :)

Purely impersonal criticism, but I'm getting 'beautiful' fatigue from Show HN submissions.

'Beautiful' and 'amazing' in the same paragraph once sounded impressive, but now it's kind of tiresome.

Like I said, this is nothing against you, it's more a general trend that I'm tired of seeing. Not everything on the web needs to be 'beautiful'.

Not everything needs to be beautiful, but we wanted to Kandan to be. We had a prototype of Kandan we called Kogo initially, but it was pretty ugly. We went out of our way to make sure that Kandan was as pleasant to use as possible, visually and functionally.

Agree. It's especially obnoxious if it comes from the author himself. It's the equivalent of having a tagline saying "I'm a smart developer who works hard" - it might be true but it's not your call, it's just a counter-signal.

On second thought, I guess my criticism is more personal than yours, but I really think it's a kind of pollution that has no place in a somewhat clean community such as HN.

EDIT: I do like the look of it, but I think everyone would be better of if you (you as in Cloudfuji, not the specific designer in question) used a neutral language. If someone goes on to say in the comments "That's beautiful." it's worth so much more for everyone involved.

I didn't design Kandan, but I do think it's beautiful. And as I said, that's because of Sacha Greif's amazing work.

Design in open source is - in general - sorely lacking. We want to correct that, and help herald in a new era of apps that are not only powerful, but easy to use and a joy to look at.

Thanks! I personally reserve words like "beautiful" for things like sunsets or mountains, but in any case I'm glad you find the design aesthetically pleasing :)

Reply to the edit - I agree, we'll be less brazen with the next announcement. Thanks for the earnest feedback, we'll use it to iterate on our style.

> Not everything on the web needs to be 'beautiful'.

Is there anything in nature that isn't beautiful? Beauty comes from the optimal intersection of form and function; in nature, this balance is reached through evolution. That's what we should be striving for.

> Is there anything in nature that isn't beautiful?

Yes.

I vote for the fruit of the doll's eyes plant personally, though bleeding tooth fungus is none too pleasant to look at, either.

I blame Steve Jobs: http://www.youtube.com/watch?v=gHZzb42s3y4 http://www.youtube.com/watch?v=Nx7v815bYUw

I came here just to write exactly the same thing. I'd add that in this case, I personally don't find the product beautiful at all. It's not ugly, I don't find it visually offensive, but it's fairly generic and uses too many cliches for me to find it beautiful.

Of course, beauty is mostly subjective. In order to wield it like, say, Apple does sometimes, you have to make something that a strong majority will find unusually beautiful.

Read the book. Good stuff. I really liked the format.

I don't see much beautiful about it.

It's a plain chat layout with too much whitespace for my taste.

You know what's beautiful? A chat that I can theme and customize to match my definition of beautiful.

This one doesn't even meet the minimum requirement of not being trapped in a browser. I want working keyboard shortcuts, I want detachable windows, and I want to copy/paste properly (browser chats give a garbled mess because your fancy CSS layout does not survive the process).

I KNEW IT!

I saw this UI and immediately recalled the ebook. I bought the ebook and reviewed it on HN. I thought the design was beautiful and elegant.

This is GREAT to see it in a live app.

I think Chat is one of those core services, like Email that has standardized around one common standard. People use a wide variety of clients, and very different uses from basic to power user needs. The reason most chat services fail is that even if they have fantastic mobile and web clients, they cannot provide the sheer range of clients needed for the entire range of use cases.

At Chartboost, we've tried pretty much every protocol there is, but we still stick to IRC. It's the only protocol that has clients that suit all of us, from hardcore `irssi` to newbie Adium integration. There's also bouncers and solid local histories and it's easy to hop onto tech channels for some help and it's widely supported with chatbots and libraries for building your own chatbot. (eg. Ruby-IRC)

I agree that IRC is still the chat protocol to use. Though in choosing a chat solutions there are some questions that often arise:

1) persistency (BNCs are still a techie solution, servers)

2) good web interface and

3) file support (other than DCC, unless a bot is present)

Here is where for example Campfire managed to take quite a big chunk out of the startup chat market.

Are there any clients that integrates well with IRC with those 3 qualities? (file support could be solved by a custom file repo with imgur-like links, or a DCC bot for those on irssi)

Quassel solves (1) in a user friendly way. It does not address your other points, though (there's a Linux, Windows, Mac desktop version and an Android app, but not a web interface AFAIK).

The source code is a treasure trove for those looking for an example of a larger Backbone.js application with a Rails backend and Faye for push notifications to the browser.

Backbone code - https://github.com/cloudfuji/kandan/tree/master/app/assets/j...

Faye (on the Backbone side) - https://github.com/cloudfuji/kandan/blob/master/app/assets/j...

Faye (on the Ruby side) - https://github.com/cloudfuji/kandan/blob/master/app/models/a...

Thanks - Akash Manhor (https://twitter.com/hashnuke) made both of those decisions early on in Kandan's life. There have been challenges around both, but overall it's worked out very well - in fact, the faye interface opens up some really interesting possibilities. We'll have some examples of this soon.

We will be talking about Kandan and other exciting open source projects at our meetup

http://www.meetup.com/San-Francisco-New-Wave-Open-Source-App...

I'd like to thank our friends at singly.com for generously offering to host. Come join us for a drink and meet some friends.

This is great, I admit the design is pleasing, but I gotta ask the obvious question here which is... what is wrong with Campfire or HipChat or the multitude other group chat apps that have sprung up in recent years? (not counting that Kandan is opensource, which is pretty cool)

The fact that it's open source is very important to us, and that (along with it having a wonderful design) will serve as the kernel that'll move Kandan in an amazing direction. With things like WebRTC and javascript audio/video decoders and the open source design, I think we'll see Kandan turn into a really amazing tool.

I wanted some of these features in Campfire for years, but I couldn't just fork Campfire and give the features back. Now I , and others, can.

Most people don't want to run their own services, but they want to be able to fix bugs and add features when they feel like it. Cloudfuji's model lets everyone do that, and app authors to make sustainable revenue.

And thanks for bringing that up, it's an important question!

Just a note: the name comes from the Japanese word 「歓談」(also pronounced "kandan"), meaning "chat."

it carries a thoroughly positive connotation with it, though, as suggested by the 「歓」。

Yes, it can also be translated as "pleasant talk."

Would love to see Etherpad integrated into this or the other way around! Awesome :)

An etherpad running on on Cloudfuji, or integrated into Kandan?

We're dying for a Sharejs-based etherpad app on Cloudfuji, but all things in due time :)

From the source of the app (https://github.com/cloudfuji/kandan), it seems that it doesn't check for authorization -for now-, so be aware that anyone can signup to your own private instance if they know your url and do nasty things like deleting channels.

Good point - we wanted to keep it as open as possible, and as easy to get started as possible. Hence the (literally) copy/paste instructions to get running on somewhere like Heroku, with no need to provision other services.

But obscurity isn't security. We'd love to hear what people think about how this should be handled - https://github.com/cloudfuji/kandan/issues

Quite nice.

But you should seriously improve your registration field, I could register with an one character long password.

even an empty password is possible!

Talk about a nice user experience! :)

In all seriousness, fixed. Thanks!

This app does have a really great design, especially when a lot of open source apps are lacking in this department. I'm really impressed with Sacha's work, and have really enjoyed his "Step by Step UI Design" ebook, which details his design process for Kandan.

Look forward to contributing sometime!

What's going on with the design of http://cloudfuji.com/apps/new?app=kandan ?

Left aligned slideshow of different sized compressed images and no text. You took the time to make the client look pretty, 10 more minutes won't hurt.

Really dig the design on Kandan but I wish it could be used as a front for XMPP/Jabber also.

This is an exciting integration that we've considered and will be exploring more in the future.

May I just put forth a preference for the backend to be based on ejabberd? Perhaps the folks at Process One would be interested in collaboration, as they are a significant contributor to the project.

I personally would love to see it be a self-contained irc/xmpp/jabber/faye server, but all in good time. We want to make open-source beautiful, easy to access, and easily extensible. So we're releasing what we have now, with plenty of momentum towards things like providing an XMPP backend!

I am happy to see IRC here. Way underused communication method! Certainly are things that can be improved with the protocol and services, but few people understand how useful it can be for business aswell as personal chat-needs.

Congratulations Sean! This looks great, and I'm looking forward to trying it out.

I tried to change url of my application, but it doesnt seem to be working. There is just no button to submit the form with application settings, I tried to sumbit it in other way, but it made no visible effect

I'll take a look at it right away - can you give me the app name, or just email me at sean@cloudfuji.com and I'll get it sorted.

http://concise-winner-5927.cloudfujiapp.com/

Use promo-code hn-preview on http://cloudfuji.com/promo for a 50% discount on the team package - 3 apps and 6 teammates.

But make sure to give us your awesome feedback :)

Congratulations Sean and team. Sacha, nice work on the design!

Thanks, glad you like it!

Is drag&drop the only way to upload things? Would be nice to have a <input type="file"> because I have no place to drag stuff from.

Currently yes. I've added a ticket https://github.com/cloudfuji/kandan/issues/20 to see how we'll fix that up.

Is there a reason you don't have a place to drag from?

ChromeOS, possibly?

Ah very interesting possibility. I have no idea how Chrome OS works, or if it's something we should worry much about right now.

Because I use my terminal as file manager, and urxvt doesn't support drag&drop yet.

So you have to "install" this app to chat? How is this better than IRC? Personally, I think grove.io has much more features.

looks great, can't wait to try it

small nitpik: the github link links to creating the app on cloudfuji

The github link is broken for me as well. Here is the direct link to the github project for anyone else looking for it https://github.com/cloudfuji/kandan

Fixed!

Love it, however I wish it wasn't AGPL as I really can't use that in my startup.

Reply on news.ycombinator.com