Jun 6, 2013

Coffee can’t do that: 5 updates in 5 days

If you are just getting going with a startup, and especially if you’re considering developing apps, this post is for you.


We released Warmly, our first Google Play app for Android about a week ago and it’s definitely been a crazy ride. 

For readers who may not be familiar with our work, we’re a small startup team that moves fast and likes to experiment. In the past year we’ve built 20+ online projects and shared them with the world for free. For the most part we go with our gut to get something working that we think is interesting, then listen to what users say and watch what they do.


If a project get a lot of traffic and interest, we rev on it and make it better. If no one gets excited, we move on. The approach keeps us from over-thinking and over-planning a fluffy idea; it also keeps us humble.

It’s common that we get mixed reviews when we put things out there. Some people are super excited and see the curious sparks behind a our quick-turn projects. Others expect everything to be robust from day-one and get upset that it won’t load in IE 6. But Chaos is currently just a couple people weaving together design and development as quickly as possible. We get excited about new perspectives and clever ways to do things, but we’re not Google or Apple or Microsoft. We’re not even Facebook or Twitter. We’re small, and we like it that way.

So when we put Warmly in the app store with a $2 price tag on it, we figured we’d do what we do best; make it exciting to us, then listen to what users think and iterate quickly.


But the feedback quickly escalated as many users had issues beyond what we could test on our small set of phones and tablets. The reviews took a turn for the worse and people we’re upset, throwing 1- and 2-star reviews at us in rapid succession.

One of the first bugs was an issue in waking up the phone from a sleep state (we used the recommended android activity flag, but that didn’t work on some devices and lock-screen configurations). That meant that for a few phones, it wouldn’t wake up and sound the alarm if the phone wasn’t active. Users set the alarm, some even testing it before bed, but the alarm didn’t go off! That totally sucked for everyone, so we refunded users who experienced the issue and feverishly worked to roll out an update. We also responded to the reviews and tried to engage everyone and let them know that we were working hard to make things better.

But the next update had another major problem. Since the update process on Android kills the currently running version of the app, the next morning a few alarms didn’t go off for another reason; the alarm wasn’t set after updating, and the phone auto-updated at night. Again, it bit us hard, and we made sure to include a call in the next (and all future) updates that checked if the alarm was set and re-enabled it as needed.

With the most critical show-stopper bugs behind us, we turned to working on features that users were requesting. Warmly is intended to be both extremely simple and experientially delightful. We purposely didn’t include things like a snooze button, multi-day scheduling, and multiple alarms because it makes the experience complicated and promotes objectively less healthy sleep practices (see this fun video by AsapSCIENCE).

But we’ve decided that the middle-ground of sticking to our design/experience direction while accommodating user desires is where we need to go. For example, one of the more recent updates allows for snoozing, but in a way that makes it secondary (it’s visually light and small). We’ve also started tuning sounds and improving weather accuracy by listening to reviews and talking with users. Additionally we’ve found that being very specific in the app description about what the focus of Warmly is (and isn’t) has helped to set expectations before people press buy.

Which brings us to today. Five days and five updates later, our user feedback has turned 180 degrees. Sure there’s the occasional comment from a user that runs into a snag, but for the most part our users are loving that we’re making improvements so quickly. Even a handful of the negative reviews have been updated by users to reflect their appreciation for pushing updates that make things better. 


In fact, we even made it on the top apps page for Health and Fitness, as well as the Google Play homepage. No ads and no gimmicks, just lots of work, sweat, and discussion with users.

It also seems important to note that waiting another two weeks before putting Warmly in the store probably wouldn’t have changed much. The first critical bug only happened on particular android builds and hardware. The second critical bug was tied to updating the app via the play store (maybe this one would have been caught, but still). Everything else has hinged on user comments and discussion.  We took some heat, but it was worth it.

Of course, it has lead us to a pretty bifurcated set of reviews so far, but hey, things are looking up. :)


Only time can tell what future users may think of Warmly, but I know that we’ll be doing everything we can to make it awesome. The feedback from reviews (both positive and negative) has enabled us to find edge-cases incredibly quickly, and we’ll continue to listen carefully and push updates as we learn what works and what doesn’t.

$2 is real money (I could have had a coffee), and purchasing something that doesn’t work as well as you had hoped is a bummer. But the app world has great potential to re-frame what buying something really means. It wasn’t $2 for a static program in a box, it was a one-time payment of $2 for an evolving app that keeps getting better everyday. I wish my coffee could do that.


TL;DR: Continually improving your app wins hearts (and purchases). Make it good, get it out there, then be ready to lose sleep chasing down bugs and talking with your users. It’s awesome.

Posted by akumpf
Jun 4, 2013

Warmly alarm clock walkthrough

We’ve received quite a few questions about how Warmly works (especially those who don’t have an Android device), so we put together a realtime walkthrough. Since Warmly turns on slowly over five minutes, the video is a bit long but you can always seek ahead and jump around if you’re in a hurry.

The last 48 hours have been a flurry of activity as we continue tracking down device-specific tweaks and bugs. One of the upsides of publishing an Android app on Google Play is that over 2000 models of phones and tablets get immediate access to it; of course, that also makes for a lot of edge cases to catch!

We’ll keep updating the app as we find new opportunities to make it better. Bear with us as we navigate the ups and downs of app development. :)

As always, we’re curious to hear your feedback. What do you think about Warmly?

Posted by akumpf
Jun 1, 2013

Warmly wakes you up, sans alarm

We all experience it; setting an alarm the night before for something important, only to hit snooze in the morning until we’re running behind. And each time the alarm goes off, it jolts you awake with beeps and boops.

Say hello to Warmly. Warmly takes the alarm out of waking up.

Warmly is a friendlier way to start your day than the traditional alarm clock. Wake up gently with the calm sounds of weather and something you enjoy: like cooking breakfast, brewing coffee, whistling at work, or catching a plane. It’s surprisingly natural and a fun way to get up and do something awesome instead of hitting snooze (in fact, Warmly purposely doesn’t have a snooze button!).


We made the user interface ultra-clean and simple. Just set the time you want to get up and Warmly will plan ahead to make stepping out of bed enjoyable.

A few minutes before your alarm is set to go off, Warmly will get the weather for your location and begin quietly playing ambient sounds. Warmly starts out silent and very gradually turns up the volume so you’ll wake up naturally to the sounds of the day and the activity you selected.

The goal of Warmly is to fade in gradually enough that you don’t wake up to an alarm, but rather that your body naturally ramps out of deep sleep as it slowly becomes aware of the environment.


So does it really work? Well, after trying it out the past week I have to admit it works better than I thought it would. In fact, the first couple days the weather sounds tricked me into thinking birds were outside my window as I opened my eyes (and I helped put those bird sounds into the app!).

I’ve also noticed that I wake up gently a minute or two before the final alarm time, which feels both relaxing and productive. I’m actually getting up slightly before I planned to.


Picking an activity is half the fun. We tried to start with nine commonly heard sounds that are both recognizable and pleasant.

The first three are all sounds around the home: making breakfast, brewing coffee, and waking up to muffled news. The middle three are other places: playing sports outside, meditating in the country, and whistling at the office. Finally, the bottom row are all travel sounds (since that’s often the reason to get out of bed): catching a plane, taking the subway, and driving a car.


There are a TON of alarm clock apps, but we really think this one is different. Warmly isn’t an alarm, it’s a better way to stick to your own goals and get up with a smile on your face. :)

We hope you’ll check out Warmly on Google Play and let us know what you think. 

imageGet it on Google Play

Get a great night’s sleep and wake up ready to do awesome things!

Posted by akumpf
Apr 25, 2013

Rebuilding BandFu — Day 7

The day after we launched BandFu, we sat down to prioritize what to make better. We had just been through a major restructuring for the front end and knew we needed to dive back into the music creation side. Making music in BandFu was not quite where we wanted it to be, but we launched anyways…



For us, the answer is two-fold: perspective and predicament.

After several internal drafts of collaborative music tools, we knew we were making a tool for ourselves, not our audience—not a good thing. The simplest, fastest way for our small team of two to get feedback and gain insight about what really needed to be built is to release early. The risk, aside from maybe hurt feelings and diminished egos, is practically nothing.

It was immediately clear that that people were using the product differently than the design intended. No one was making full length tracks, pulling their friends into their sessions, or taking full advantage of the sampling/FX engine running under BandFu. Most notably: almost nobody was pressing ‘record’.

What they were doing instead was sketching. A short jam here, a little experimentation there. Build a synth, play a bit, move on. Clearer still, we realized this almost instantly because that’s how we used BandFu ourselves! We just hadn’t stopped to think about it in the push to create a richer, more extensive tool. And it is this perspective that is defining the next iteration of BandFu.

A quick side note: visitors also tend to gloss over some of the things you’re most proud of. For instance, we were (and still are) pretty excited about our password-less account creation! But so often these things that go unnoticed on launch. Try logging in on bandfu, we dare you ;)

The other benefit is the predicament we are now in. My apple dictionary defines predicament as “a difficult, unpleasant, or embarrassing situation.” As a professional procrastinator our predicament is our call to action. We now have a thing out in the world that needs to improve and grow. It has to be nourished or it will die. It’s not so much that we’re embarrassed of BandFu, but that it isn’t yet what we want it to be. To be content we have to iterate, refine, and rebuild.



After looking at how BandFu was being used, we took a step back: what needs to be better, what needs to be different, what needs to disappear. 

We decided on two driving factors:

  1. The core of BandFu is an instrument, and should be played as such. This means more dynamic input options, and playing a larger variety of sounds at once.
  2. The BandFu environment is your audio notebook. A place to quickly sketch and play with musical ideas (with your friends if you want). Share your creative process, or take them into professional software to expand and improve on them.

This means we’re stripping the current creation space for parts and reconstructing it into a new, better version of itself.

In particular, the record button is going away because BandFu will always be recording. This leaves you free to jam away at will, and jump instantly back to save the parts you like after you play them. The creation space itself will move away from a timeline focus, and become more focused on building a custom instrument for you to play with. 


How all these things will actually be implemented is what’s in progress now. So bear with us as we go through a pretty rapid design and development sprint. If the idea of a live space to collaboratively play, share, and save your musical ideas sounds exciting to you, let us know in the comments. Our goal is that professional and aspiring musicians alike will be able to use and play creatively in BandFu.


What do you think of in-progress posts like this? We’re happy to share more often, just let us know in the comments! Or get in touch with us at hello@chaoscollective.org

Curious about BandFu’s humble beginnings, take a peek:image

Posted by tinymonstr
Apr 17, 2013

Say Hello to BandFu, an Experimental Audio Playground


The WebAudio API was introduced to enable developers to begin creating real-world audio tools in the browser. But so far most of what’s been created is either a sub-optimal port of traditional desktop audio software or fun experimental audio generators and visualizations that feel more like a one-note hack than an expressive musical instrument.

So we created BandFu, an online space that enables new musical instruments and compositions, all from within the browser. While BandFu still has a long way to go, it’s already bringing together some aspects of music creation that we’ve all been looking for; things like realtime collaborative editing, custom synth and sample generation, a dynamic effects pipeline, analog control of timing and parameters, and exporting sounds back out for use in other workflows. All this together is intended to bring a playful approach to the creative process of making new and interesting musical artifacts.


As a brief overview, BandFu allows audio to be directly recorded from the user’s microphone, modified by numerous effects, and added to a timeline. The recorded sample can then be easily looped to create patterns and sample blocks for complex musical arrangements. And while the intention of BandFu’s collaboration is to enable working with others, the same tools and functionality allow you to collaborate with yourself, going back in time and layering new tracks over your previous recordings. And with the ability to create loops (even infinite loops are welcome!), you can essentially be playing sounds in the future moment.

The dynamic effects pipeline is also a key part of BandFu’s capabilities. By enabling sound-altering components to be arbitrarily pieced together, creators can easily transform and bend sounds at their will. Sure, you can plug in your guitar or stand-alone synth and jam, but you can just as easily record your own sound, like saying “booooop”, add some sweet effects, and map it across the computer’s keyboard to build a custom expressive instrument without using any extra hardware. It’s an analog approach built on the power of digital tools.


BandFu has also given us a chance to start developing a framework for Asynchronous Realtime Collaboration (or ARC for short). In a nutshell, ARC lets us conceptually bend time so that a user can experience something as it was in the past (or how it will be in the future). It’s not Sci-Fi smoke and mirrors, it’s just a reframing of what now is, and by doing so we can get around some of the limits of synchronicity that often come with tightly timed collaborative spaces. As we debug ARC to be more robust and capable, we hope to share it as open source soon so that other collaborative tools can leverage the framework.

The app currently works best in Chrome since it relies heavily on the rapidly still-evolving WebAudio spec.

If you’re curious, fire up your browser and give it a try. Might we suggest this funny track about trains as a starting point?

Let us know what you think, or directly file a bug or suggestion for BandFu here.

Posted by akumpf
Feb 11, 2013

An Open Letter to the Speed of Light


Dear Speed of Light,

While it is clear that you are praised in many circles as being incredibly fast, it has come to our attention that you are, in fact, too slow. Much too slow.

As earth dwellers we share a common desire to work together on projects synchronously. As a simple example, imagine musicians performing live together from different locations around the world at the same time — this needs to happen, but it won’t if we solely rely on you. Why? Well unfortunately, Speed of Light, you need to be about 100 times faster for that to become a reality.

Read More

Posted by akumpf
Jan 11, 2013

WebNinja Tutorial: Asynchronous For Loops

For loops are simple, primitive building blocks in code. And they’re everywhere for good reason; we often need to do things repeatedly.

But when javascript took a sharp turn toward making everything asynchronous (particularly Node.JS), the way we were taught to use for loops just doesn’t apply for many common tasks.

Read More

Posted by akumpf
Jan 3, 2013

Maybe one day, Ubuntu. One day you’ll get it right.

I don’t want webblications that make my desktop OS better, I want a desktop OS that makes my Internet better.

A subtle difference, maybe. But an important one.

And 2014? Seriously, what the blazes? It’s only just now barely 2013! Gah, I can’t think about stuff a year in advance. Get that dev image out pronto. 

Posted by tinymonstr
Dec 21, 2012

Open Source in a Closed Economy


We’ve built a lot of things in a short time. And we’re still building even crazier, awesomer things. But we’re faced with a constant challenge: How do we keep the spirit of open source and public sharing while being a viable business?

Read More

Posted by akumpf
Dec 14, 2012

1-Day Hack: DOF-Changeable Images!

Have you ever been curious to try out some of the new photography gadgets that let you play with the focus of a shot after you’ve taken it? They’re called DOF-Changeable images, and the web has been a-buzz about them for a while now. Unfortunately, the $400 pricetag (like the sweet looking Lytro) makes it a bit out-of-range to buy one just for tinkering.

So we created a similar effect using a basic digital SLR camera and shooting video (getting depth-of-field information by capturing it over time instead of over space as Lytro does). See for yourself by clicking around on the images we captured and put up on the DOF page!

Posted by akumpf
Page 1 of 4

« Getting Older

Or, RSS addicts follow along here.