I’m Marco Arment:
a programmer, writer, podcaster, geek, and coffee enthusiast.
A Birth Story
→
•
∞www.marco.org/2014/11/08/a-birth-story
My friend Meaghan O’Connell, one of my favorite writers, describes childbirth in only the way that she can.
It’s long (save it), and I wanted to quote almost every part. Here’s one of the many sections that made me laugh out loud while reading it in bed at 1 AM last night:
I wished for a way to communicate pain more precisely than a scale of 1 to 10. But the scale is subjective, I longed to say. We have no way to know. I hated this. I said 7, 8. I didn’t know. It was the worst pain I’ve ever felt, but I have never had my arm cut off. That was what I always imagine to be the worst pain: having a limb chopped off. I saved 10 for it, out of respect. I wanted to save 9 for the moment the baby tears its way out of my vagina. So what’s left is 8. I wanted to seem brave, so I said 7, but then I worried they wouldn’t understand the immediacy of the situation, so I came back with 8. …
The residents wanted to check me. I was told this wasn’t a teaching hospital, I wanted to say, but don’t. They asked me if it was okay with me if two people checked me. I thought it had to be for accuracy but no, it was for their own benefit. If there’s anything I’d take back from labor it would have been the fact that I let two people fish around in my vagina for their own benefit.
I’d been “checked” before. This is what they call it. They want to “check you.” You means your cervix. You are your cervix. “Check” means stick a hand inside of “you”—your vagina—and measure how open your cervix is. They do this with their fingertips, because that is where we’re at with science in 2014: We use fingertips as a unit of measurement. Then you are pronounced whatever number of fingertips wide the gap in your cervix is. You are your cervix.
In typical Meaghan style, it’s refreshingly humane, surprisingly informative, hilarious, and terrifying.
◆
Panic discontinues Unison gracefully
→
•
∞www.marco.org/2014/11/07/unison
Unison’s end is bittersweet. The market for a Usenet client in 2014 isn’t exactly huge. But if you know Panic, you know we do our very best to never drop things awkwardly — we like to leave our apps in a good place for our (very) valued users.
While we can no longer work on Unison or offer support for it, the good news is it’s also free. This version of the app will be automatically unlocked for all users, no serial needed.
Unison has the ridiculous level of polish and detail you expect from any Panic app, applied to a type of app that’s not only completely out of fashion, but whose typical customers usually couldn’t give less of a damn about UI quality.
I can’t really blame them for ending support — I’m surprised they were able to justify working on it as long as they did. Still, I hope it works for a long time so we can all keep using Usenet for its completely legitimate purpose.
For some bewildering reason that’s likely a reflection of society’s bankrupt standards for journalism, a lot of people read Business Insider.
Whenever one of my products or I am mentioned in it, I get more people coming out of the woodwork and telling me they saw it than from any other press coverage — usually via Facebook, a website that assumes I’d like to read an algorithmically profitable subset of the random writings of people whose random writings society expects that I should care about, but I don’t.
Anyway, somebody wrote this on businessinsider.com, a domain name on the internet, and the aforementioned people are woodworking about it despite it having no relation to me:
Facebook Is Seeing More And More App Developers Go Android-First
Facebook is seeing a trend in Europe of app developers going “Android-first.”
The first line of the article has already proven the headline to be misleading. And those who find the rest of the article between all of the other garbage on the page1 may notice that the headline actually doesn’t reflect the real story at all. This is the bulk of the actual information being reported:
But Codorniou told us Facebook has a team of evangelists encouraging Android developers to use Facebook as a way to build and promote their apps: “As of today, I have four guys from my team in Paris talking to Android developers about the greatness of Parse, Facebook login, app links, app events. It’s a very important bet for us.”
“There is a pattern coming from Eastern Europe. The Russian developers develop on Android first because of a big audience, and it maybe being easier to develop. They liked the fact that they could submit a new version of the app every day. This is a trend that I see and I think it is going to accelerate.”
In other words, the staff members at Facebook tasked with promoting Facebook to Android developers have, unsurprisingly, been talking to Android developers. And it turns out that many Android developers prefer to develop for Android first.
If there’s actually a trend toward Android-first, this story isn’t showing any evidence of it.
The likely truth is that there is no noticeably shifting trend of developers choosing Android first because of its market share, or choosing iOS first because of its profit share, because that’s not how developers choose. Most developers with the authority to choose their platform will choose whichever one they use and like best.
If there was really a shift occurring toward Android-first, a significant number of iOS developers would be switching teams, developing for Android first and probably switching to Android as their carry OS. But of every iOS developer I know — and I know a lot — only one is choosing that path.
And the number of successful startups that launch on Android exclusively, or even first, remains far smaller than the number that start on iOS first.2
Maybe someday this will change, but it would require far more talented mobile developers and startup founders to switch to Android themselves. Find that story first, and the apps will follow.
At the time of writing, these distractions include an ad for denture glue, a newsletter sign-up solicitation, a Facebook sharing solicitation, a LinkedIn sharing solicitation, a Twitter sharing solicitation, a Google+ sharing solicitation, a “Print” solicitation despite most web browsers including a print feature, an email sharing solicitation, a huge ad for Dropbox, clipart stolen from Flickr, a Facebook stock ticker, a generic stock ticker labeled “Your Money” that doesn’t actually represent my money as far as I can tell, banners for articles including “A ‘Sexist’ Ad From UK Newspaper The Sun Offering A Date With A Topless Model Has Been Banned” (with a photo of five women wearing bikinis and promoting unattainable body images), “The 6 Types Of Killers Who Use Facebook To Connect With Their Victims”, “Here’s What Happens When You Eat Olive Garden For 7 Weeks Straight”, “6 Scientifically Proven Things Men Can Do To Be More Attractive”, “Scientists Have Figured Out What Makes Women Attractive”, “How To Get A Dancer’s Body”, “Women Are Going Crazy Over These No-Underwear Yoga Pants”, a repeated sharing-solicitation bar at the bottom of the article with every aforementioned option except the already redundant “Print”, links and banners to 31 (!) more garbage articles, and Tynt to taint your copy-and-pasted text. ↩
I know of zero. I’m sure it’s not that bad, but I bet it’s not great. ↩
Apple had actually been pretty clear about what developers can expect with the rollout of WatchKit. This has helped temper my expectations and make me more realistic about what will be possible to build at launch.
◆
South Park’s full episode on freemium games
→
•
∞www.marco.org/2014/11/07/south-park-freemium
Nailed it.
◆
Amazon’s Echo Chamber
→
•
∞www.marco.org/2014/11/07/amazons-echo-chamber
Dustin Curtis on Amazon’s weak consumer-hardware efforts:
There is simply no rational explanation for its products. The only thing I can come up with is this: Amazon continues to make hardware because it doesn’t know that it sucks, and it has a fundamentally flawed understanding of media.
(Via Ben Thompson’s Daily Update, which is great.)
◆
Microsoft Office now mostly free for iOS and Android
→
It seems that Microsoft is finally accepting the reality of Office’s market position on smartphones and tablets. To some degree, Office needs to compete with the free Google Docs and iWork, but for many customers, it’s also competing with the idea of simply not using office apps (or using them far less often).
This is a bold move, but I think it’s the right one: give it away for casual and independent use to protect (and hopefully grow) its ubiquity and relevance from further erosion, and make money from big-business users with big-business needs and big-business budgets.
It’ll be interesting to see if they go further with that strategy. Why not bring the same business model to Office for Windows and Mac?
Statements like Matias Duarte’s justification for not using the iOS share icon in Google’s iOS apps are why I don’t think much of Google:
The share icon Google uses in it’s [sic] properties (and the share icon that Android endorses) is a popular opensource icon and one that we feel well describes the connective nature of sharing. In a sense you could say we believe it’s part of our brand and that Google’s brand is to embrace the open and universal standard.
(Via Daring Fireball tonight.)
Maybe it’s just my inability to understand anything Matias Duarte ever says, but I see this as Google’s typical bullshit, insulting our intelligence as they push a self-serving corporate branding initiative and sheer arrogance as an inevitable, morally imperative “open standard”.
Why not tell the truth? Google’s apps don’t use the iOS share icons because Google doesn’t respect iOS1 and thinks its standard UI widgets are better, even in their iOS apps, on a platform surrounded by other apps that all use the standard iOS share icon.2 Secondarily, it reinforces their branding and makes the rest of iOS feel just a little bit more alien to people who heavily buy into the Google ecosystem, reducing iOS’ lock-in and making it cognitively easier to switch away.
Google’s use of their Android sharing icon in their iOS apps has nothing to do with “open” nonsense and everything to do with Google asserting that they know better.
Apple shamelessly pulls the same move — see, for instance, every Windows app they’ve ever made — but they don’t patronize us with bullshit justifications.
This is what rubbed me the wrong way about Jeff Atwood’s “Standard Markdown” move, too. He positioned it as “open” and “standard”, but it was really about Jeff not respecting John Gruber’s intelligence or ownership at all — which has been clear for years to anyone who follows Jeff Atwood — and wanting to take control of the Markdown name himself for his own blatantly non-standard desires. ↩
Or a slight but clearly recognizable variant, like Tweetbot’s rounded-corners version. ↩
◆
A Voyage to 2009
•
∞www.marco.org/2014/11/05/kindle-voyage-review
The Kindle Voyage e-ink reader is so unremarkable that I’ll just direct you to Jason Snell’s better review for most details, since it’s not worth writing up a full review here.
I was expecting better after years of Kindles being decontented into flimsier, lower-end devices, but I think it’s clear that Amazon just isn’t willing or able to make a premium, high-quality e-reader.
Rather than approximating buttons, the Voyage’s overly complicated “pressure-based page turn sensors with haptic feedback” are the worst of both worlds: they lack the precision, feedback, and intentionality of buttons, and they take more effort and are smaller than touch targets.
The Kindle software and interface is even worse. It has changed very little since the 2011 Kindle Touch, which itself was mostly just basic touch interaction bolted onto the 2009 Kindle 2’s UI.
And this crisp, new, high-resolution screen is still displaying justified text with very few, mostly bad font choices. Some of these choices, like the default PMN Caecilia font, made sense on the old, low-resolution Kindle screens but need to be reconsidered for this decade. Some of them, like forced justification and forced publisher font overrides, have always been bad ideas.
But Amazon doesn’t care. Nothing about the Voyage’s software feels modern, or even maintained. It feels like it has a staff of one person who’s only allowed to work on it for a few weeks each year.
The ideal Kindle would have hardware page-turn, Home, and Menu buttons and a touch screen for UI navigation, selection, and text entry.
The hidden, error-prone touch zones would be optional and off by default, since the hardware buttons would remove the need. Touching the screen would only be used to interact with what’s being displayed on the screen (which usually isn’t necessary) rather than constantly triggering unintuitive, undiscoverable, error-prone actions.
Text would be rendered with a small selection of high-quality fonts designed for high-resolution screens or printed books, and full justification would be either optional or unsupported. The user’s font and justification preferences would optionally override any fonts specified by publishers.
But Amazon has never made such a device, and it seems like they never will.
I suspect this will be my last Kindle. Amazon doesn’t care about e-ink Kindles anymore. Why should we?
◆
Cheap, huge cloud backups for a Synology NAS
•
∞www.marco.org/2014/11/04/synology-backups
I’ve greatly enjoyed my Synology NAS, the DS1813+,1 for over a year, but I hadn’t found a good cloud-backup solution until now.
Arq to S3 or Glacier
Arq runs on a Mac, not on the Synology itself, and can back up any mounted file shares to Amazon S3 or Glacier.
Arq works well, but S3’s pricing can get prohibitive: priced per gigabyte, a 1 TB collection costs $30/month to host. The Reduced Redundancy option brings it down to $24/month, but that’s still significant. On the special bulk Glacier service, 1 TB is only $10/month, but Glacier is extremely clunky — simple operations can take hours or days to complete.
I used Arq with Glacier for months, but I wouldn’t recommend it. Glacier isn’t made for this, and it never lets you forget that. I’d only recommend Arq if you’re willing and able to pay for the more expensive S3 or RRS storage.
But my family’s backup set is about 4 TB (and growing). $96–120/month for RRS or S3 is simply not worth it when there are other options that aren’t priced per gigabyte.
Backblaze via iSCSI
I’ve used Backblaze to back up my Mac for years, and it’s still the option I recommend for that.2 It’s a flat $5/month fee with unlimited storage, but while it supports external drives, it won’t back up network shares. But you can trick it with iSCSI if your NAS supports it, and Synology does.
iSCSI makes networked storage appear as a raw, unformatted, locally mounted disk to your computer that you format and use however you’d like, with all of the benefits and limitations you’d expect from such an arrangement: you can’t share the volume with multiple users and the NAS can’t read the files or do anything dynamic, but it’s much faster, Spotlight works properly, and Backblaze will back it up like any other directly-connected disk.
But iSCSI is not designed for intermittent connections, like Wi-Fi — only use iSCSI over wired Ethernet. More critically, it requires special software: Macs need an $89 or $189 “initiator” kernel extension to use iSCSI disks. Third-party kernel extensions are bad news, and it’s wise to minimize your dependence on them.
I used the GlobalSAN iSCSI initiator for a few months so I could use Backblaze and browse the disk more quickly, but the initiator started causing frequent problems, disconnections, and failures. I eventually removed it and went back to normal network shares that Backblaze won’t back up.
I’ll only try iSCSI again if Apple builds native support into OS X, which seems unlikely — if they ever intended to, they probably already would have.
Symform
Unlike Arq and Backblaze, Symform runs directly on the Synology. It’s a clever idea: copies of your data are split into tiny pieces and backed up on other people’s NASes running Symform. If you contribute space to the pool, you get free backups at a 2:1 ratio.
In theory, it sounds great. In practice, that 2:1 ratio is a big deal — to back up 5 TB, I need to host 10 TB of other people’s data — and I started having doubts of the reality of relying on random people’s NASes for my data integrity. I was tempted to just make a big RAID-0 volume, then realized that my data would likely be relying on a bunch of other geeks’ RAID-0 volumes.
Those concerns aside, Symform’s fatal flaw was the upload speed. I never got Symform to upload faster than about 1.5 Mbit/s — nowhere near my 75 Mbit/s upstream capacity, and fatally slow for terabytes of data. The complexity of splitting files up between so many different people likely requires a bunch of little connections and transfers, and that’s tough to scale for high throughput, especially with the limited RAM and processing power of a NAS.
CrashPlan
CrashPlan can run directly on the Synology or a computer, and if you run it on a computer, it’ll back up network shares (unlike Backblaze). Otherwise, it’s similar to Backblaze — unlimited space, $6/month — but has a clunkier, more resource-intensive Java app.
I’ve tried CrashPlan a number of times over the last few years, both on my computer and the Synology, and I’ve always abandoned it because of the same issue: awful upload speeds, usually well under 2 Mbit/s.
Inspired by this Alter3d post about fixing CrashPlan slowdowns, I decided to give it another try a few days ago. I installed prebuild packages on the Synology using Scott Hanselman’s guide (it used to be much harder).
Based on Alter3d’s findings, I first tried just adjusting CrashPlan’s advanced settings, still using encryption but setting de-duplication and compression to their lowest settings. This worked very well, immediately raising the upload speed to about 20–25 Mbit/s during the day and even faster at night. (I later edited the config file to tweak the values manually like Alter3d did, but it didn’t make any difference.)
So far, it’s working fantastically. I uploaded 1 TB in 3 days. It still isn’t saturating my upstream, but it’s going fast enough that it won’t be a problem.
A newer DS1815+ is now available, although the differences are minor. ↩
Disclosure: Backblaze sponsors our podcast regularly. For whatever it’s worth, I used them for years before that. ↩
◆
Low post
→
•
∞www.marco.org/2014/11/02/low-post
I love when Dr. Drang is Dr. Drang.
◆
Accidental Tech Podcast: DeLorean + McLaren
→
•
∞www.marco.org/2014/11/02/atp89
A good one this week: Casey’s new-parenthood, Tim Cook’s coming out, iMac minutia, FILM CRIT HULK, and discovering f/1.4.
◆
•
∞www.marco.org/2014/11/01/short-form-blogging
Too much of my writing in the last few years has gone exclusively into Twitter. I need to find a better balance.
Fortunately, I’m not the only one thinking along these lines. These two pieces are great:
Jason Snell: Bigger than a Tweet
Gina Trapani: Short-form blogging
I don’t think avoiding Twitter is pragmatic if your audience is there, but it’s also unwise to dump all of your writing into bite-size pieces that are almost immediately skimmed over, forgotten, and lost to the vast depth of the mostly unsearchable, practically inaccessible Twitter archive.
Twitter is a complementary medium to blogging, but it’s not a replacement.
I’m glad to hear from people like Manton who are exploring this problem as well. Some simple shifts and modernizations in tools, formats, and styles can go a long way.
By knocking down a few walls and moving some furniture around, blogging is preparing for a comeback, and we’ll all be better off for it.
◆
Thank you, Tim Cook
→
•
∞www.marco.org/2014/11/01/thank-you-tim-cook
Casey Newton:
It is one thing for the media to whisper to one another, or to post on their blogs, that the CEO of America’s most valuable company is a gay man. And it is a quite another for the man himself to step up to the microphone, with confidence and grace, and tell us himself. We knew Cook was gay; what we didn’t know is how he felt about it. Or, at a time when being gay is still very much a political act, what he planned to do with it.
Now we know.
◆
Congratulations are in order
→
•
∞www.marco.org/2014/10/30/declan
…to my good friends, Casey and Erin Liss, whom I’m ridiculously happy for.
Watts Martin distills the motivation behind GamerGate:
Those people form the core not only of GamerGate, but the larger culture war that bandies about terms like political correctness and social justice warrior.
When you learn to construct stories, you realize everyone is the hero of their own tale. No one sees themselves as the bad guy. Being confronted by someone saying this thing that you like is hurtful to others for reasons that you entirely missed is uncomfortable. We all know that racists and sexists are bad! If I like this thing that you say is sexist, aren’t you saying that I’m a sexist? Aren’t you saying I’m the bad guy? It’s all too easy to say no, you’re the bad guy, and to fashion a narrative that justifies going to war, literally or figuratively.
I only learned of the “social justice warrior” term a few weeks ago from GamerGate, and it still baffles me that it’s used as a discussion-ending insult.
Avoiding the unnecessary alienation or exclusion of others and getting offended by threats and harassment shouldn’t be controversial. That it is controversial, and that being accused of having such principles is widely considered an insult, shows just how deeply damaged our culture is, intractably infected by jingoism, xenophobia, and widespread acceptability and glorification of violence.
◆
“I’m Proud to be Gay”
→
•
∞www.marco.org/2014/10/30/tim-cook-comes-out
Tim Cook:
If hearing that the CEO of Apple is gay can help someone struggling to come to terms with who he or she is, or bring comfort to anyone who feels alone, or inspire people to insist on their equality, then it’s worth the trade-off with my own privacy.
I look forward to the day when a public figure being gay is no big deal. But that day has not yet come, and I’m glad Tim is helping our society get there.
◆
More on Apple post-rejecting PCalc’s widget
→
•
∞www.marco.org/2014/10/29/jcenters-pcalc
Josh Centers:
Arbitrary decisions like this harm Apple’s relationship with developers. If Apple