I don't really know why I stopped to post about the podcast my pal Timo and me are doing here – probably because of the different languages, now that I think about it. Meh, who cares?! Anyway, it's time for a little update.
While ungeheuerlich eschoofierend is somehow getting zero traction, it is more and more fun for me: We're getting better at identifying and talking about topics, and there is a trend more the 30-minutes episode, which is a comfortable length I feel like.
Inspired by less than or equal's, we redid our website. It is much more welcoming now, I think. The website also sports the podlove subscribe button, an attempt to make subscribing to podcasts less cumbersome.
I am doing the editing for the show, and boy is that fun. Everythin audio was completely new to me, and I found – and still find it – interesting to dive in. It's both the best and the worst: Every little detail is a potential rabbit hole, from the microphone (let me tell ya!) to the bitrate of the audio file you put out into the world. But it also is tremendous fun, and I honestly think ue is a pretty good sounding podcast by now. *humblebrag*
So, if you speak or understand German: the latest episode of ungeheuerlich eschoofierend is about user interfaces – in general, and especially in public toilets. It may be my most siracusa-like performance yet.
Here's an AppleScript that passes variables (episode title and description) to a bash script, opens it in a terminal window and runs it, letting you watch how it generates a podcast-ready mp3 file. I've put it into an Automator workflow that accepts audio files.
Here it is on GitHub.
It should be pretty self-explanatory: Just fill in the path to the script. Which script? I'm glad you asked! It's the script that does the "work", if you will, of generating an mp3 using LAME. I used a fork of Marco's script because it was easy to feed with variables and a bit more elegant. To use it, just fill in your podcast's name, the path to the artwork, and you're good to go!
Getting the POSIX file name of the input file via AppleScript was kind of a hassle, I hope whoever runs into the same problem finds this.
With the method of using AppleScript to pass variables to a bash script, I riffed off of Jason Snell's solution.
Once again, something seems to be happening with podcasts. There were roughly 500 articles and blogposts written about Serial, so people seem more likely to know what podcasts are. It's not quite a gold rush like I've heard it was around 2004, but it's something.
Caused by this, there was more talk about the medium itself, and I found myself thinking more and more about all the ideas and observations that were brought up.
There's a lot to say about the mechanisms of this medium, how it works, and more importantly how it does not -- yet. Podcasting is so much rooted in tech geekdom that not all of the problems are easy to see or grasp for everyone involved. Actually, that could be the cause for many of them. Acknowledging that podcasting can't retrospectively change its roots, let's just take stock for now:
- There's the whole 'white media' thing (hello again, Serial).
- There's the (also white) 'dudes talking about tech for (four) hours' phenomenon, which has multiple layers (quality, topical and conceptual monoculture, as well as 'white media') as Dave Wiskus points out.
- Discovery is a big problem. There are podcast directories, sure, and iTunes is fine for being the place where all the podcasts are. But you likely won't find your new favorite podcast by browsing a genre list.
- There is the underlying technology, both for consuming and for producing podcasts that has multiple intertwined problems
All of these are certainly important and/or interesting in itself. In this post, I am picking what I find to be the most fundamental, and if you look at it chronologically from a user's perspective, literally the first problem: The onboarding experience: Somebody's first podcast episode.
Nobody listens to a Podcast by Accident
But before I look at that specifically, I want to present my theory why podcasts are so different from everything else in today's media landscape. This namely is why nobody finds himself on a podcast's website, clicks ▶ and by that became an avid podcast listener.
One could rant about how it is 2015 and it is audio we're talking about, how it is unbelievable that we haven't solved this yet. But that would miss the point: The problem, when you think about it, is actually pretty complex and hard to solve. Spoken word audio cannot be handled like text or video, or even like music. It exists in a totally different space. In the average day of the average human in our modern world, it requires a very specific state of mind to be able to even begin listening to spoken word audio.
What about Text and Music?
We are reading, looking at our phones, skimming Facebook and Twitter all day. And yes, you might think, we are listening to music all the time! That is correct. But music just goes too well with skimming text and looking at pictures. And with working, depending on what you do.
Let's be honest, few people commute or walk around town with their eyes closed, focusing on the music they're listening to. Spoken word audio though requires you to pay attention and therefore doesn't go well with everything just mentioned.
What about Video?
But what about YouTube? Video adds another dimension, requires much more attention from the user and a higher-bandwith connection than text or music. Much, much more so than a podcast, certainly! Still, people are watching videos on YouTube all day long! Why is it then, that videos are so shareable and go viral all the time, while podcasts aren't at all?
Here, I think, duration is the key element. I have no data to back this up, but let's just agree that the "watched all the way through"-rate of a YouTube video significantly drops with every minute or even second that it is longer (some statisticians may just have died. Also: Sorry, Dr. Drang). Podcast episodes rarely come in under 20 minutes, the attention span equivalent to about 200 funny YouTube videos and 1000 gifs.
Listen to it Later
I think that's why the daily routine of most people is a very hostile environment for opportunities for 'getting into podcasts': When we're most likely to stumble across a link or a reference to one, we're the least likely to end up listening to it. That is why podcasts cannot and do not spread like articles or videos or gifs. They just die out there, man.
So, podcasts have to exist in a different space: Mobile. Hands-free. Commute. Dishes, laundry and the like. That is why podcasts, when they streak the user's attention, need to be easily 'saved for later'. That is hard to pull off.
The real sad part is how hard it is for somebody who even has made that intellectual leap and said "Yeah, on my way home, I'm gonna listen to that episode" to follow through and make the technological one as well.
Technology is supposed to connect people to content, but at the same time, and necessarily so, stands between people and content. In an almost literal and very sad sense, the latter is the case with technology and podcasting.
I alluded to this earlier already: I needed, and you, if you're reading this, probably need to realize that we are nerds. Things we don't even think about, things we at best subconsciously recognize are inconvenient, for others are either hurdles that are very annoying to overcome, or brick walls that are impossible to. In that state of mind, witnessing the following was just mind-boggling.
I recently recommended the How does a farmer work? episode of the Working podcast to my sister. She doesn't really listen to podcasts, but she is a real foodie and I thought she might be intersted in how the interviewee grows certain kinds of apples.
Dead End Episode Website
So, I sent her the permalink to the episode on the slate website and she actually was willing to listen to it, albeit later (understandably, as I pointed out above). She was probably sitting in front of her Windows PC, received the link on her iPhone.
Okay. Piece of cake, right? Hoho hooo hoohohoh, noo no no. Inconceivable! The laughable options I could think of:
- Mobile browser. For a 20 minutes episode that might be almost feasible, but listening to a longer audio track in a mobile browser, pre-loading the episode while on WiFi is at best an enormous hassle.
- Dropping the episode mp3 file in Dropbox doesn't work, either. I felt very clever and tried, but this kind of thing only seems to work with .pdf files on iOS (it also would be a hassle, and required the Dropbox app).
- The Working podcast is hosted on SoundCloud. I suppose the SoundCoud app is a possibility. Still, this requires that donwload and doesn't apply to all podcasts. People's willingness to download one app per podcast might be the only thing lower than to download one dedicated podcast app.
So, we see the link to the episode at best helps in regard of letting my sister decide if she wants to listen to the episode or not. Like a frikkin program guide from the 80s! When it comes to listening through the episode, there's just nothing. The options listed above all were duds. This battle is lost! We have to take a step back, regroup and take a look at the podcast as a whole.
The next best thing would be to grab the RSS link and subscribe to the podcast proper. Reality check: Just Imagine you had to subscribe to all the YouTube feeds you want to see one video of. 'Subscribe' is a big word! I don't blame anybody who shys away from that kind of commitment. But for the sake of this post let's say you feel crazy today and are fine with subscribing to this podcast you haven't even heard a full episode of yet.
The process isn't trivial at all. The user needs to a) find the RSS link on the website and then b) know just to copy the RSS link and then c) drop it into his podcast app of choice that d) needs to be installed or bought/downloaded beforehand.
If the user doesn't copy the feed URL and instead just taps on the RSS feed address, they're at the mercy of the OS: iOS 8 comes with Apple's own podcast app, so in that case they're kind of lucky.
Android's Chrome browser just displays the raw RSS feed. Your average user just freaked out.
Quick aside: This is a thing the Podlove project's Subscribe Button tries to fix. It sits on a podcast's website and presents the appropriate options of subscribing to the podcast to the user. At first I was sceptical, but now I think it's a great idea. Check it out on ungeheuerlich.org!
So, to be clear: I don't think anybody makes it this far without really wanting to listen to the episode, or being really committed and still having somebody explain the whole thing to them. Meaning: Why it is so annoying and that it's worth it.
App Good, All Good?
Anyway: I recommended a podcast app to my sister (Castro, for its slickness – I myself prefer Overcast), so she didn't use Apple's app but still was able to go ahead. Reminder: Interacting with a whole new kind of app is always a big hurdle for most users.
Should be easy now, right? Again, wrong. We wanted to get to a specific episode, remember?
Either by having grabbed the feed URL from the Working website, or by searching for the podcast name within the app (remembering the name and then manually typing it, like it's 2007) my sister subscribed to the feed. Phew!
The average podcast app now would go and download only the most recent episode. While this is perfectly reasonable, this again isn't what my sister wanted.
She had to go in, delete that most current one and then skim through the feed to find the episode we're all doing this for. For shows with huge backlogs, finding an older episode is real hassle.
We're done. Finally. Recommending podcasts almost seems offensive to me at this point.
Sorry, I don't have the silver bullet. But maybe something to make the last 10% less painful: I'm thinking about URL schemes or something like that. Something that not just kicks the user over to the podcast feed, but a specific episode of it. Opened in-app, that would make it very easy for the user to download that specific episode (kudos to Marco here for allowing Overcast users to do so without ever subscribing to the feed).
Maybe a search function within a single podcast's feed, to make huge backlogs more manageable, is feasible as well.
Of course, there's the idea of chapter marks out there, and in some way there are time stamps in Overcast links. Both featuers add to discoverability and such, but chapter marks don't spread that well and Overcast's timestamps were broken for me when I tried them. Besides that, I don't believe any of that alone would have a profound impact.
We also have to keep in mind that every bolted-on convention puts the single best (and only good one, one could argue) thing about the technology at risk: Its openness, it being decentralized and stupidly simple (from a technological point of view): Just an RSS feed with linked audio files and some descriptive text. So beautiful.
Anyway, I hope something happens in this regard. I hope this article helps to identify problems that keep podcasting back. There is so much great content out there, it would be a real bummer if we can't figure out how to make it more accessible for everyone. Thank you.
Thanks a lot to Jason to so kindly mention my article, or rather my feedback email, on the latest episode of Upgrade. Interesting thoughts about how Apple might be in the best position to do some (more) ground work for podcasts.
Make sure to listen to (at least) that intersting bit, starting at minute 08:04! <-- Overcast timestamp link!
Google strives to make every kind of information searchable. A fascinating, and to some somewhat sinister goal. Now, the European Court of Justice has decided that every citizen of the EU has a "Right to be forgotten" that is enforceable against Google (press release, .pdf – En, Ger). Surely, that is for the better, no?
No, it is not. Not at all. Except for one detail: The ECJ ruled that European Privacy laws are applicable to Google's services offered in the EU, even though it's incorporated in California. That makes sense. Ok.
Let's take a closer look at the ECJ's decision.
If an individual wishes for links to be removed from the results for a search for their name, Google has to do so. While there are exceptions for public figures, the rule is that privacy trumps freedom of speech and freedom of information. That's it. There's no balancing by the ECJ, as there ought to be whenever fundamental values like these collide. That is the most fundamental problem.
A very, very disturbing consequence of that lack of balance is that the ECJ separates information itself from its discoverability. I'll give an example: An individual requests the removal of a newspaper article about themself. Google complies with the request. The article, on the newspaper's website, stays online. Technically, the article is still "out there". The point is that nowadays, what Google doesn't find also doesn't exist, as far as the general public is concerned. And the ECJ explicitly knows that.
Just a reminder: We're talking about perfectly correct and lawful content here. Mr. Masing, judge with the German Constitutional Court, hits the nail on the head: He points out (Ger) (I'm paraphrasing) that in a free society, it would suggest itself that a statement covered by freedom of speech can also be distributed freely.
The ECJ seems to disagree on that very point, arguing with the "effective and complete protection of data subjects".
One caveat: As soon as your search query doesn't contain that person's name, the newspaper article will show up in the results. How do you find information about a person without using their name, though? Managing that will become some sort of "skill" or something along those lines.
Trusting the cat to keep the cream
That's not the end of it. Google itself is obliged to make the decision about which request to grant. There's all kinds of problems with this. As just described, the ECJ's decision bases on the very premise that people won't find what Google doesn't list. Hereby, it acknowledges Google's gatekeeper position to the Internet. Instead of carefully working with this though, the ECJ goes on and hands Google the holy grail: The decision over what information is to be "forgotten" and what information is to be preserved.
You don't like Google's power that lies in managing all the information in the world? Better make sure to let them decide which website to remove from search results!
Google's impossible task
I was part of a small group of law students that visited Google's in-house counsel here in Germany, shortly after the ruling. He told us that Google definitely doesn't want the burden of having to make these decisions. Not only because it costs them a ton of money, but also because Google aims to make information available instead of doing the opposite, so the very idea to selectively de-list websites repulses them. But of course they would comply with the law. It was also said that Google, in case of doubt, would risk litigation in the name of freedom of speech.
Whether or not you believe all parts of that statement: They don't want the burden of the process, that's for sure. They are puzzled, and rightfully so, since the ECJ didn't give too much guidance on how to make these decisions.
Google simply isn't capable (and really nobody would be) of making tough decisions like these with a lack of information about the specific case – on a massive scale, let's not forget that.
So, to somehow deal with this appropriately, Google seems to want to assemble some form of "council" to come up with some rules for the decision process. Now, it is understandable that Google wants to show that they're on top of this.
The problem is that any body of people picked by Google will always lack democratic legitimization. If this is to be "proper" co-regulation, there have to be requirements for the regulation process. Coming up with those would be the European legislator's task, not that of a group of people assembled by Google, however qualified they might be. The ECJ also misses the opportunity to lay these requirements out.
Getting privacy advocates into the process also wouldn't work. There is a danger of them becoming "Communication regulation Agencies", as Masing puts it (Ger).
In case you're in doubt that the current model can't work: We've already seen how it works out. Google de-lists specific search results, in compliance with the judgment. Then the Guardian complains about Google not making the right call, playing the censorship card, and legitimately so. Google reverses its decision, also justifiably.
Nobody could blame Google for defaulting to make the decisions so that there's the least complaints, meaning granting all requests except the ones demanding de-listing of a Wikipedia page or a newspaper article. There's some hope that there will be litigation at some point, resulting from Google not complying with a removal request (or defending a specific de-listing, but I honestly think the other way around is much more likely).
That means trusting Google with going to court in the name of the freedom of speech. I'm not entirely convinced that's what the ECJ intended.
A side note: This is supposed to be about privacy. Well, I have to be honest: I’m a law student, and not since yesterday. But "privacy" for me was more the kind of thing that allowed you to go to Deutsche Telekom and request all of the information they have about you, like Malte Spitz did. Search results are nothing more than links to publicly available websites. "Privacy" wouldn't have come to my mind in a thousand years. And, at least before the ECJ’s decision, many others wouldn’t have thought of it either. Now, of course, nobody can afford to even express this thought.
The decision's severe consequences come from the inapt application of "full effect" privacy principles to an information and communication context.
Inconsistencies and backfiring
On top of that, there are several weirdnesses and inconsistencies.
Normally, you're redirected to a Google page with a TLD according to the country you're located in, for example google.de in Germany. On Google's main page though, you can switch to google.com (link in the bottom right corner) and hereby get to unaltered search results. Yes, it is that easy.
On one hand that is good, because that means all of the stuff I described above is easy to circumvent. On the other hand: The fact that wrong regulation is easy to circumvent doesn't make it right or even less dangerous. Even more dangerous actually, because it makes the slippery slope that this is seem more harmless.
Alluding to Google's position as a gatekeeper: Any search engine operator with meaningful market share will be obliged to do the same. Remember, due to the massive scale, this requires significant infrastructure and therefore raises the bar of entry into this market even further. Maybe Microsoft will get annoyed with all of the EU's regulation and just leave the search engine business altogether. Who knows? Anyway, this is another way the ECJ further reinforces Google's position.
Another point, seemingly small but even more weird, is that publisher-run and for-pay databases will keep all the links (Ger) that are de-listed from Google search results. This, accidentally, makes freedom of information some kind of a media privilege.
Jeopardizing freedom of speech
There's no other way to put this: This is really, really concerning. Freedom of information and freedom of speech are fundamental values, especially in the information society we live in today. All of this for prohibiting you from finding anything not entirely positive about your new neighbor online?
We simply can't afford undifferentiated, trigger-happy and blanket-style approaches that endanger these values, be it in the fight against child pornography, copyright infringement or for the protection of minors (meaning porn filters).
All of this is not to say that I don't see the problem with information living online, possibly forever. But I honestly think that's a societal, not a technical one. Managing an online persona will only become an ever more important skill, and we have to acknowledge that. Making people compete for being the best private "censor" can't be the way. I, for one, don't want to live in that kind of society.
Mr. Masing, the judge of the German Constitutional Court heavily criticizes the judgment in an essay (Ger), letting me look forward to a decision of the German Constitutional Court concerned with any of the relevant aspects. The majority of law-savvy bloggers (Ger) also have many complaints.
The Electronic Frontier Foundation's take is also quite interesting.
I'm anxious to see how this will play out.
My pal Timo and me are doing a podcast again!
After a first attempt two years ago that lasted for an unbelievable amount of four episodes, we're not getting ahead of ourselves this time. However, the name stays the same: ungeheuerlich eschoofierend, and we're still talking in German, but we're not rambling about whatever comes to our minds anymore (at least not all the time). Instead, there's one big topic per episode. For now, we're on a 2-week-schedule.
Find our very nice site here: ungeheuerlich.org – or just get the RSS Feed for the podcast app of your choice.
The first episode was about Craft Beer, the next one's about Religion and will be published next Wednesday. You should consider liking ungeheuerlich eschoofierend on Facebook. Cheers!
Say what you will about Russel Brand, I do like his attitude. He's authentic and is saying something meaningful in an interview. He totally gets that he's not the leader of a revolution, but also that he doesn't have to act dumb because he's an actor.
Check out the short piece on gawker about the interview.
Weirdly fitting words by Mark Sample:
"Low point" is the term for when the worst part of a disaster has come to pass. Our disasters increasingly have no low point.
After the low point of a disaster is reached, things begin to get better. When there is no clear low point, society endures chronic trauma.
Disasters with no clear low point: global warming, mass extinction, colony collapse disorder, ocean acidification, Fukushima.
I smiled all through the second half of this video. The excitement about one product and the craft it takes to get there is really something beautiful.