A Safari upgrade I like: accountability for resource-hogging pages

Apple is a few days away from shipping its next big update to its desktop operating system, but people running its current and previous macOS releases can already benefit from one of macOS High Sierra’s components.

Yes, I’m writing something nice about Safari for a change.

The browser that I’ve spent much of the past few years cursing at for its weak memory management and general inability to let me run the computer instead of the other way around got a welcome, pre-High Sierra update Tuesday.

The most talked-about feature in Safari 11.0 may have been its ability to automatically silence sites that without invitation play videos with audio on (yes, I know that includes some of my freelance clients), followed by its blocking of cross-site ad tracking. But the option I’m enjoying most at the moment is Safari 11’s ability–stashed in a new “Websites” tab of its preferences window–to open every page at a given site in the minimalist Reader view.

Where ad blockers are often clumsy and random, Reader can be an elegant weapon against sites that demand attention with junky ads and auto-playing media. It might also spare you from a particularly piggy page locking up your Mac with a demand for more memory than the system can allocate.

“Isn’t that the system’s damn job,” you say? Yes, it is. Fortunately, Safari 11 also now seems able to quash a site in the middle of a memory binge, to judge from the banner I saw atop a page advising me that Safari had reloaded it “because it was using significant memory.”

I’m not going to tell the Safari developers to kick back with a nice vacation – since this update, the browser has already forced a reboot when it somehow refused to restart or fully quit–only a week after I’d had to go through the same routine with Google’s Chrome. But at least I don’t feel like this app is conspiring against me.

Advertisements

I tried targeting you all with a Facebook ad. It didn’t work well.

Last Sunday, I finally saw something new on Facebook: an invitation to run an ad campaign on the social network and pay for the whole thing with a $30 coupon. Since other people’s money is one of my preferred payment methods–and since I’d been meaning to see what the Facebook ad mechanism looks like from the inside–I accepted the offer.

I couldn’t choose a post to promote, as the coupon was limited to the post I wrote here about money-losing prompts at ATMs and credit-card readers overseas that had become unexpectedly popular when shared automatically to my public page. But I could pick who would see the ad, as identified in a few different ways. In case you’ve wondered just what Facebook advertisers can know about you, here are the options I saw:

• Target people who like your page, people who like your page as well as those people’s friends, or people you choose through targeting. I picked the last, in the interest of science.

• Reach people at a region or at an address. The default was the District. I could have picked an address, but since I’m not promoting a business at a fixed location I didn’t see the point. But with this post’s travel-centric focus, I should have picked Dulles Airport–right?

• Choose interests (as expressed by people on Facebook in things like Likes). For this post, I selected “Air travel,” “Europe,” “credit cards,” and “personal finance.”

• Pick an age range and a gender.

• Pick a duration and a total budget for the ad campaign.

I could have gone deeper into some of these options, but since I was navigating this dialog on my phone during our daughter’s dance class, I didn’t have unlimited time. I submitted the ad, got an e-mail saying it was under review for compliance with Facebook’s ad standards, and got a second e-mail 16 minutes later saying the ad passed.

Three days later, Facebook sent me a summary. Their $30 had brought the Facebook share of my post here to another 2,516 people, of which 56 had clicked on the link and one had left a comment on my page.

This report also informed me that the ad’s audience was 95% male, which is both confusing and unsettling. Maybe I should have targeted only women, considering that my page’s audience already skews so heavily male? Age-wise, the ad found its biggest audience among the 25-34 demographic. I’m not clear about that either.

What I do know is that my WordAds ad revenue here doesn’t support spending $30 to reel in 56 views, so I doubt I’ll be running this experiment with my own money anytime soon.

Do I really have to use Snapchat?

Snapchat filed for its initial public offering Thursday, which makes it a good time to admit that I completely suck at Snapchat.

I have the app on my phone (I installed it first on my iPad, which should exhibit how confused I am about the whole proposition), but it’s among the least-used apps on that device. And that doesn’t seem likely to change.

Self-portrait using Snapchat's snorkel-and-fish lensFew of the friends from whom I’d want to get real-time messages number among its 158 million daily active users, and even fewer seem to use it actively versus lurking on it. I will check out the occasional Story from a news or entertainment site, but that slightly longer-form medium has yet to become a regular part of my info-diet.

I could use Snapchat as yet another way to connect with readers. But without any clients or readers asking me to do this–and with a surplus of social-media distractions already on my various devices–I’m struggling to see the upside.

The biggest reason for my holding off is, to put it bluntly, is that I’m ancient relative to Snapchat’s millennial demographic. I didn’t get the initial appeal of the app when it was focused on sexting disappearing messages, and I’ve been stuck in a get-off-my-lawn mentality ever since.

Snapchat’s self-inflicted wounds are part of the story too. This startup had barely been in existence for three years before having a data breach expose partial phone numbers of more than 4.5 million users, after which it accepted a 20-year settlement with the Federal Trade Commission. That’s not the sort of thing that makes me want to give an app access to my phone’s contacts list. It does not help that founder Evan Spiegel hasn’t exactly seemed like the most enlightened founder in tech.

Finally, there’s Snapchat’s cryptic interface, which expects the user to swipe in random directions to see what features might surface. When the clearest explanation of this UI comes as a diagram on page 92 of Thursday’s S-1 filing with the Securities and Exchange Commission, we have a serious failure of discoverability. That, too, does not make me want to spend my time figuring out this app.

As that pre-IPO disclosure to investors itself admits: “These new behaviors, such as swiping and tapping in the Snapchat application, are not always intuitive to users.”

I’m not going to delete the app from my phone or anything. It’s an important part of social media today, and I should stay at least functionally literate in it. But if you were hoping to have that be yet another instant-messaging app you can reach me on… look, don’t I have enough of those to monitor already?

Perhaps I’m wrong. If so, please don’t try to convince me otherwise in a Snapchat chat; leave a comment here instead.

 

Sparring with a 3-million-plus-follower Twitter account

I expected angry feedback to Wednesday’s post about WikiLeaks and its increasing recklessness, but I didn’t know how that would play out. The @WikiLeaks Twitter account has 3.33 million followers and a history of jabbing at critics, and the story of WikiLeaks posting a trove of Democratic National Committee e-mails–with zero attempt to blank out personal data like Social Security numbers–intersected with the angst of Bernie Sanders fans who are themselves not known for social-media silence.

WikiLeaks Twitter interactionThe WikiLeaks account quickly took exception to my post (and supportive tweets) in responses ranging from boastful–“The Hill, Gawker and others published alleged DNC docouments months ago. Only WikiLeaks had impact.”–to dorm-room BS–“Sure. Anyone who exposes the estabishment by telling you the truth is not your friend. We got it.”

Many of those three-million-plus followers then started liking and retweeting those tweets. I’m not used to seeing my notifications fill left-to-right from so many people clicking on the same tweet.

My new interlocutors came from different places. Some were hardcore WikiLeaks defenders. Some backed Donald Trump and so were in favor of anything making Democrats’ lives more difficult. Some were Bernie Sanders fans convinced that the DNC had stolen the election from him, despite the absence of proof.

(Sorry, Bernie fans: The Democratic Party–especially the woefully-mismanaged DNC–is nowhere near organized to pull that off. Also, you might want to think about where your militant confusion of a party bureaucracy’s dislike of your candidate with “rigging an election” might end up taking you.)

I tried to reply to the tweets directed at me but soon lost count, leading to me feeling I was reliving Seinfeld’s “jerk store” episode when I saw rebuttal-worthy material half a day too late.

But I did not have to answer any hateful crap attacking my gender, race, ethnicity or religion. Every time that happens, I think I’m playing this game with a WHT PRVLG cheat code.

After a day of this amusement, it was nice to see Edward Snowden come to the same basic conclusion as me and then get his own moralizing response.

Memo to frequent-traveler programs: Kids shouldn’t need their own e-mail addresses

Our almost-six-year-old is already in multiple marketing databases, and it’s all my fault: Once our daughter couldn’t depart with us for free, we started signing her up for frequent-travel programs. The price of miles and points are already baked into the tickets we buy for her, so we might as well take part–and besides, you’ll never hit million-miler status if you don’t start sometime.

JetBlue River Visual viewBut tending these accounts has been more work than I imagined, because some companies have a hard time grasping that children represent a special group of customers who can’t be expected to have their own e-mail addresses.

At first I thought I’d solved this problem with “sub-addressing”creating a new e-mail address on my existing Gmail account by adding a plus sign and additional text to my username. It’s an Internet standard, and I had no issues creating accounts for our daughter at United Airlines, JetBlue, American Airlines, and Amtrak with a “plus-ed” address.

But when I tried logging into our daughter’s United and JetBlue accounts a week ago and was greeted with various errors, I saw that both airlines had stopped accepting sub-addressed e-mails.

The problem was worse at JetBlue, since your TrueBlue ID is your e-mail address. I had to call and provide our kid’s account number and the no-longer-accepted e-mail address; the rep told me she needed her own e-mail address but then accepted a version of my Gmail account with a dot in the middle of my username. It’s weird to have to go through such a workaround when JetBlue’s site has a separate workflow to create a child account.

At United, I could change her e-mail to a dotted version of my Gmail handle after logging in, since MileagePlus account numbers double as usernames. United’s Twitter account then told me I could have put in my own e-mail for her account from the start. I would not have guessed that, since UA’s account-opening UX assumes you’re a grownup–and the e-mails sent to our kid suggesting she jet off to the likes of Australia, Brazil and Israel don’t exactly speak to the under-10 demographic.

Meanwhile, Amtrak and American Airlines still seem to tolerate plus-ed e-mail addresses. (I can’t speak to Delta, as that airline’s network doesn’t work for us.) But after the last week, I won’t be surprised if our little one gets unexpectedly locked out of either account; I just hope I don’t have to spend too much time on the phone to fix that problem.

Dear Android gesture typing: Enough ashtray with this one autocomplete error

About 95 percent of the time, the “gesture typing” built into Android’s keyboard is one of my favorite parts of Google’s mobile operating system. I trace a fingertip over the letters of the word I have in mind, that item appears in an overlay before I’m halfway through inputting it, I lift my fingertip, and a Scrabble winner like “deoxyribonucleic” pops into a tweet, e-mail or note.

Ashtray forbiddenBut then I have to try to gesture-type the $10 word “already,” and Android will have other ideas. Specifically, cancerous ones: It keeps subbing in “ashtray.”

I could take some comfort in the privacy-preserving thought that Google’s sense of my interests is so off that it thinks I write about cigarettes all the time. (This post probably won’t help!) But, really, I only want to write a common adverb without having to tap keys one letter at a time, like some kind of an animal.

And because I’m in the same abusive relationship with autocorrect as everybody else, I keep hoping that this time, Android will finally realize I have no interest in tobacco products. That’s when it will humor me by inserting “airway” instead of “ashtray.”

I just don’t know what could lead Google to misread me so consistently. The path you trace to gesture-type “already” requires going significantly farther east on the keyboard than the route for “ashtray,” and Google’s own search results suggest the former word is used about 143 times more often than the latter. Who would see any upside from this pattern of error?

As I said, this is annoying because it’s so unusual. The only other common word that Android gesture typing botches halfway as often is “conference.” And there, it’s at least more creative: The software will sometimes choose to read my attempt to enter this noun for an occupational hazard of my job as me gesture-typing “cicatrice.” Yes, real word: It’s a Middle English-derived noun for tissue that forms over a wound and then becomes a scar.

One thing I do know about this mystery: If there’s a tech conference with “already” in its name, I may have to decline all of its invitations to preserve my own sanity.

Mail encryption has gotten less cryptic, but some usability glitches linger

I seriously underestimated you all late last year. In a Dec. 7 post about encryption, I wrote that I hadn’t gotten an encrypted e-mail from a reader in years and said I expected that streak to continue.

PGP keysIt did not. Within a week, a dozen or so readers had sent me messages encrypted with my PGP public key (under subject lines like “Have Faith!” and “Challenge Accepted”), and several others have done the same since. That’s taught me that the crypto user experience has, indeed, gotten pretty good in GPG Suite, the Pretty Good Privacy client of choice in OS X.

But at the same time, some awkward moments remain that remind me the woeful state of things in the late 1990s.

Most of the them involved getting a correspondent’s public key, without which I could not encrypt my reply. When it was attached as a file, dragging and dropping that onto the GPG Keychain app had the expected result, but when it came as a block of text in the decrypted message, I (like other users before me) wasted a few mental processor cycles looking for an import-from-clipboard command when I only had to paste that text into GPG Keychain’s window.

I should have also been able to search keyserver sites for a correspondent’s e-mail address, but those queries kept stalling out at the time. One reader did not appear to have a key listed in those databases at all, while I had to remove a subdomain from another’s e-mail address to get his key to turn up in a search.

One more reader had posted his public key on his own site, but line breaks in that block of text prevented GPG Keychain from recognizing it.

The GPGMail plug-in for OS X Mail is in general a pleasure to use. But its default practice of encrypting all drafts meant that I could no longer start a message on my computer and finish it on my phone–and one e-mail that I’d queued up in the outbox while offline went out encrypted, yielding a confused reply from that editor. I’ve since shut off that default.

It’s quite possible that the upcoming stable release of GPG Suite for OS X El Capitan will smooth over those issues. But that version was supposedly almost ready in late September, and there hasn’t been an update on that open-source project’s news page since. I suppose having to wonder about the status of a crucial software component counts as another crypto-usability glitch.