I saw Moana recently with my wife and in between being delighted by the story, tapping my toes along to the music, and crossing my fingers my toddler would like it when I inevitably buy it, I could not stop starring at the water.
Disney Animation does not go for photo realism with their characters, but materials still need to behave the way your brain expects them to so they stay on the good side of the uncanny valley. What Moana has accomplished with the animation of water is truly staggering. And it is, unsurprisingly, the same folks made the snow so believable in Frozen.
Seriously, it is worth seeing this movie just to marvel at how far computer models for complicated things like fluid dynamics have come.
Postscript: In the time between when I saw the movie (and literally gave myself a todo in the theater to write this blog post) and actually got around to writing it, I came across this fantastic post about the water animation that you should go read.
Two years ago I gave a lightning talk where I used the story of Ignaz Semmelweis as a metaphor for educators denying reality and sticking to their “tried and true” methods, no matter how untrue we now know they are. I even have some useless-without-me-talking slides to prove it.
We now know for example that left and right brain thinking is not supported by any evidence, we know that the belief in learning styles is not just wrong but dangerous, we know that the claim that we only use 10% of our brain is completely unfounded and that if you are using Neuro linguistic programming for education purposes then you might as well be practising astrology. Yet despite these important findings, may of these beliefs still persist in education settings today.
Comcast has once again called my attention to a UX anti-pattern: misnamed credits. To be fair, I’m sure this isn’t unique to Comcast, but they get all the accolades for brining it to light for me.
Recently I’ve been having some issues with my service, and after many attempts to resolve it I was offered a credit for the time I was without usable service.** Imagine my confusion when I recieved an email notifying me of my free “SPEED INCREASE”.
I suppose technically they did try to increase my speed from so-slow-as-to-be-unsuable to the-speed-I-was-paying-for. Even their website and my bill refer to this credit as a SPEED INCREASE. Since this credit was issued from a phone call I have no record of their admission to any problem even occurring, let alone any ownership of the problem. All the documentation I have shows them generously giving me a free speed increase! How nice of them!
And this is not an isolated incident. I looked back at previous emails/bills and all credits from Comcast I could find were framed in this way. I’m no UX expert, or novice even, but this is such an obviously shady attempt to keep any damning documentation out of the hands of customers. I would almost admire them for this clever spin if it wasn’t so infuriating.
**Fun fact: Comcast does not appear to have any guaranteed uptime and maintains full control over if and when they issue credits of any kind. Basically you don’t have any in-writing expectation of service.
The first line of the article is the most revelatory (emphasis added):
Apple and the Maine Department of Education have offered to swap school iPads for MacBooks at no additional cost, after it emerged that students and teachers overwhelmingly favor the use of laptops in class.
Apple is helping them make this change at no additional cost? When does Apple do anything at no additional cost? That’s very telling.
I recently setup the Starbucks app on my phone and I was furious to find out that it supports Apple/Android Pay only to reload a digital giftcard stored in the app instead of just paying for the purchase directly. This is an insidious way of trapping your customers because the minimum reload amount is always a round number, and a purchase is almost never a round number, which means you’ll have money left on your ridiculous digital giftcard. That means the next time you want to buy a cup of coffee you’re going to think “Well I do have a couple of bucks left on my Starbucks giftcard…” and once you get to the point where the amount remaining isn’t enough for your purchase, you’ll have to reload it and the cycle starts all over again.
Mobile games are the most notorious example of this design choice, and they add in the additional manipulative elements of theming and big numbers. “I need 1500 gems to unlock the next level, but I only have 500! Good thing I can buy 2000 gems for just $0.99! Oh wait… it’s 2500 to unlock the level after that… good thing 2000 gems are just $0.99!”
The worst part of psychological trickery like this is that knowing what’s happening doesn’t automatically free you from it’s effects. I totally bought that cup of coffee using the app because God damn it I want those stars.
I love podcasting. I love the openness and approachability of it. And I love the community, tools, and culture that is growing up around that openness. I don’t want to see podcasting turned into the clickbait mess that most blogging has turned into.
From to the NYT article:
Apple’s conservative approach may be creating an opportunity for competitors, as has happened before. Apple’s iTunes software helped popularize online music, only to watch streaming services like Spotify and Pandora create compelling alternatives. Apple pressured the television and film industries to sell, and rent, their content online; then Netflix built subscription streaming into a business worth nearly $40 billion.
This is obviously a gross over simplification that gives Apple entirely too much credit, but the message is clear, and I don’t think it’s the one the author intended: openness and competition in these spaces is a really good thing.
And while I’m sure Tim Cook buys all his movies on iTunes, I’d be willing to bet he has a Netflix account too.
It’s no secret that iTunes is the center of the podcast universe, or that everyone hates iTunes. Fortunately there are plenty of other great podcast apps like Pocket Casts or Overcast.fm but most podcatchers still rely on the iTunes Podcast Directory for search and discovery. That means that iTunes subscribers, ratings, and reviews are still valuable, even if no one is actually using iTunes to listen to podcasts. When Google Play Music Podcasts (seriously, what is up with that name?) was announced it was also announced that interested podcasters would have to submit their podcasts to be in the directory which many hoped was Google taking a stand against the tyranny of the iTunes Directory.
But a head to head battle with an incumbent is a suckers bet because users need a reason to switch what they’re doing. Last week I read a tumblr post about NPR One and what it is actually trying to accomplish by playing seemingly random NPR content. From the post:
“Like Pandora or other similar apps, it gradually learns what you’re interested in from what you skip and what you mark “interesting,” and it shifts what it serves you…”
Her larger argument is about the importance of radio in delivering news, but the language about NPR serving you a variety of audio content came back to me when I read this line from the Google announcement today:
“…we’ll connect you with podcasts based on what you’re doing, how you’re feeling and what you’re interested in. Similar to our contextual playlists for music, we want to make it easy to find the right podcast—whether you’re a podcast aficionado or listening for the first time.”
By all accounts these feature of Google Play Music are excellent (I wouldn’t know, I’m the one-song-on-repeat while I’m working type.) It makes sense that Google would take this angle. Google is a search company after all, and podcast discovery is really hard. If they can make quality “radio” out of evergreen podcast episodes, instead of encouraging the traditional episodic chronological listening, they’ll have an incredible advantage over every other podcast player/directory. And if they can use their voice transcription magic to automatically create transcripts for every podcast submitted to their directory they’ll have an edge on every other player in terms of discovery and recall.
As of this writing, I’ve been podcasting for over two years and running a podcast network for over a year. Even before I was making my own show, I was a big consumer of podcasts too.
Podcasting is a lot of work. And most of it happens before and after you get on the mic. Thanks to podcasting suddenly becoming cool (thanks, Serial!) lots of startups and apps and services have started to pop up trying to make podcasting easier in one way or another. Come to think of it the two-year anniversary episode of Flipping Tables covered a bunch of them including one I ended up experimenting with: Briefs.fm.
Briefs.fm’s tag line is “A podcast you’ll actually publish” and they say their differentiator is that
“Episodes are limited to three minutes, published by email, and still sound amazing even when recorded on your phone.”
Limiting shows to 3 minutes is very interesting and a lot of people have (I think correctly) referred to Briefs.fm as twitter for podcasting. Publishing by email is also a neat trick, and removes a lot of friction between talking into a mic and getting your content out there. But it’s the last claim, “still sound amazing even when recorded on your phone”, that piqued my interest the most.
Here is the equipment that goes into getting my voice from my mouth to your ears on Flipping Tables:
$580?! Do I really need all this stuff? The boom arm is definitely a luxury and only affects the sound by making good mic technique easier. And this doesn’t include the Macbook Pro it’s all plugged into, Logic Pro X that I use to mix, or the hosting service to get the episode swiftly delivered to you. Not to mention the time it takes to actually mix and eq the show. Briefs.fm is claiming it can get my show published using the phone I already carry anyway and that it will sound amazing? Challenge accepted!
I got in touch with Ben and launched “Human vs Robot Mixing” (discontinued since the experiment is completed) a show entirely devoted to recording with various devices under various conditions. I wanted to see how well auphonic, the algorithm behind Briefs.fm, could possibly do with the sub-par audio recorded under sub-par conditions using sub-par microphones because audio quality on a podcast matters a lot.
Here are all six episodes, unprocessed first and auphonic’d second:
Bose Quietcomfort Headphones Inline Mic
Indoors. No ambient noise
Recorded with Quicktime on Mac
Shure SM58 with Cloudlifter and Focusrite 6i6
Indoors near window with loud road noise, tv playing in the background
Recorded with Quicktime on Mac
Nexus 6p Built In
Outdoors near very loud highway road noise
Recorded with Audio Recorder Android app
Nexus 6p Built In
Outdoors near loud road noise, in high wind, mic close to my face
Recorded with Audio Recorder Android app
Zoom H4N Handy Recorder, Built In Stereo Mics
Recorded to mono
Indoors, mostly quiet, squeaky room fan overhead
Recorded with onboard device software
iPad Air 2 Built-in Mic
Recorded with Voice Recorder Pro, settings tuned for voice
#1 Has an obvious reduction in “room noise” and sounds better overall.
#2 doesn’t sound really any different but the levels were brought up which is expected. It didn’t need anything else and the algorithm correctly left it alone.
#3 and #4 are similar except in #4 I am holding the phone much closer to my mouth so auphonic was able to boost my voice while still trying to suppress road noise, though it can’t do much about wind on a mic that has no wind screen. These tests were probably the most taxing on Auphonic.
#5 is so quiet as to be unusable, but auphonic did a great job of bringing the levels up. Couldn’t do much about de-essing though which is a strong indicator that mic technique still matters.
#6 also had the levels brought up nicely without sending it into compression hell. A little bit of room noise still rings in the background (my voice tends to bounce of walls quite a lot) but a little room noise isn’t necessarily a terrible thing.
Across the board Auphonic does a good job of bringing the voice level up, and bringing the ambient noise level down. It also compressed the files in a way that seems pretty optimized for voice, as the processed files are substantially smaller. In the case of the Zoom it went from 18 MB to just 1.1 MB.
Briefs.fm is doing something different by limiting episodes to 3 minutes, and I’m excited to see where that format goes. Even though my show on Briefs is done (I do run a podcast network after all) I’d recommend it to anyone who has ever thought about starting a podcast. Robot mixing isn’t yet ready for a highly edited show, or a show with multiple hosts, but for a solo host limited to three minutes it’s already good enough and getting better.
You still want to have decent mic technique, in a decent setting. If you’re in a bare concrete basement with a washer and dryer running in the background and you hold your mobile at arm’s length and shout at it, there isn’t much anyone, human or robot, can do to make that sound great.
Enable this feature by using the compose image descriptions option in the Twitter app’s accessibility settings. The next time you add an image to a Tweet, each thumbnail in the composer will have an add description button.
…we’ll be turning on the feature for you in coming weeks — look out for a notification in your timeline. We love it and think you will too. If you don’t, send your thoughts our way, and you can easily turn it off in settings.
So a user has to opt-in for accessibility, but opt-out of a “feature” that breaks the core value proposition of the service.
Google, Twitter, and Facebook (and LinkedIn, Github, etc. etc. etc.) are things we all use and it’s really tempting to just “Log in with [SOCIAL SERVICE OF CHOICE]” because of the convenience. And although the privacy settings of all these services change as they add and remove features, change policies, and sometimes randomly just to screw with us, Facebook, one fairly consistent thing is the “Connected Apps” section.
Chances are you logged into some game, some web app, or some other other thing that is now abandoned or you’ve abandoned but it still has access to your account. This is a gaping security hole and one that is trivially easy to fix. Just log into [SOCIAL SERVICE], navigate to their list of stuff that has access to your account and start removing access.
Here are direct links to the page in question for the services you most likely use:
Pro tip: Anything you don’t recognize should almost definitely get the ax. Also any game you “totally plan to get back into.”
I thought I’d share my own stats since I reshare this post every so often. During this most recent purge I disconnected:
Google - 4
Facebook - 2
Twitter - 5
Mostly old games or third-party clients (because sometimes the primary app sucks, Twitter) but there were a few acquired companies in there too for apps I don’t even use anymore, and I definitely don’t want that connection open for no good reason.
I was on a plane recently scrolling through Twitter on my phone (in-flight wifi is slow but it’s still amazing) when I received a text message from a friend asking a question. I replied, we texted back and forth for a few minutes, and that was that. Nothing remarkable.
Then it hit me, I had wifi but not cell service. And I was just sending and receiving old fashioned SMS. How is that possible? Turns out Google added this capability to their Messenger app and didn’t tell anyone. Or at least no one told me. From Google’s documentation:
Tip: You can send texts over Wi-Fi even if you don’t have cell service. Just use Messenger as you normally would.
Google Voice and Hangouts have enabled SMS over wifi for a while but Voice definitely seems to be dead or dying, and Hangouts actually encourages you to stop using it for SMS. So while this is nothing new or revolutionary, it’s nice to see it as a foregone conclusion that SMS would work over data, instead of being a “feature”.
I had a new “how hard could it be” idea and when I cracked open the Inspector in Chrome I was met with a ton of DOM elements I was not expecting plus a bizarrely specific class name on almost all of them. Parsing and extracting the data I needed was going to be harder than I thought. Even with jQuery wizardry.
When I went into Slack to complain about my hopes of a quick fix being dashed the response I got was a link to the Github documentation for the library being used on that page. It’s obvious in hindsight that the class name I was seeing was generated by a library. The person who sent me the documentation wasn’t already familiar with the library, they just noticed an anomaly and a pattern, and turned to Google to quickly confirm their suspicions.
While Apple is fighting the FBI in court over encryption, Amazon quietly disabled the option to use encryption to protect data on its Android-powered devices.
To be fair, they actually did this last Fall but the current climate, and slow roll out to of the update to older devices, is making this into big news as it should be.
For a company that thrives on providing such convenience to their customers this is the wrong direction to go. If users weren’t turning encryption on Amazon should turn it on by default for everyone not disabled it completely.
You can read any number of blogs and think pieces about the benefits of sharing what you’re working on and learning. I work in technology in the education sector, and between those two groups of people there are a seemingly unlimited number of voices telling me to share more of what I’m working on, learning, and most importantly, struggling with.
Most people seem to have no problem going on twitter or Facebook and complaining about a problem they’re having, but a blog post tends to command a little more thought, reflection, and often a solution.
The United States government has demanded that Apple take an unprecedented step which threatens the security of our customers. We oppose this order, which has implications far beyond the legal case at hand.
No matter what you think of Apple, their products, their services, etc. they are undeniably doing the right thing for everyone by making this encryption battle very public, and very pro-encryption.
Ask any security expert and they will tell you there is no such thing as a backdoor “only good guys can use.”
Google? Amazon? Microsoft? The ball is in your court. I know you all hate each other but this is a cause worth siding with Tim Cook and Apple on.
I am on conference calls a lot. Remote meetings are a big part of my day job, and that means dialing the same 1-800 number and the same two or three extensions several times per week. Dialing numbers by hand like an animal is an unsustainable practice and luckily some smart engineers agree with me!
That’s right, you can autodial an extension. Pause between number and extension included. iOS supports this as well as Android, but my screenshots are from Android. The process is similar though.
The number you want to dial
Insert that tiny pause to let the line connect
Add the extension you want dialed for you
Then save this whole mess as a contact and from now on you can connect to your conference line with one tap.
You can even add multiple pauses or a “wait” (which prompts before it proceeds) for more complicated dialing schemes! If you find yourself calling the same conference line even just once per month, I highly encourage this simple hack.
If Twitter were to provide an algorithmic timeline that could allow me to see the most important to(sic) content in 5 minutes or less, I’m very certain that I’d return to using the service more.
Then you don’t want to use Twitter.
Twitter that isn’t real time and reverse chronological order isn’t Twitter. There are already a ton of services that aggregate tweets, summarize breaking news events, and recap live events. Turning Twitter into one of those services won’t save it, it will kill it. A steak and a hamburger are both made of beef, but no one would argue that they’re the same thing. Grinding Twitter up and smashing it back together would represent a fundamental shift to an already crowded space, and an exit from a space that Twitter completely owns.
And do we honestly think that they’d not allow for a pure, “unadulterated” timeline as an option? That’s ridiculous. It just won’t be the default anymore, and that’s a good thing.
Facebook has a similar option; and it resets every. single. time. you. log. in. Twitter would likely take a similar position because advertisers prefer the algorithmic sort.
Oh, and wanting to use Twitter for it’s declared and well loved primary function doesn’t make you a “power user” it just makes you a “twitter user.”
Rules for federal aid eligibility require “regular and substantive interaction” between students and instructors in distance education programs. That requirement does not apply to correspondence courses. Students typically initiate contact with their instructors in those courses, which often are self-paced.
I wonder what will happen once we finally nail adaptive learning and the “instructor” is software?