Perpetual βeta: Weblog

The Trawler: № 11

2017-07-17T21:40:40Z

Welcome one and all to another issue of The Trawler, the Internet’s premier collection of “stuff”.

I haven’t produced a Trawler issue in a while. So this one is something of a link dump. There’s enough goodness here to keep you going for a long time. 😃

User Interface

Scott Jensen affords us a tantalising glimpse of an operating system user interface that employs atmospheric perspective to convey depth rather than the drop shadows that are ubiquitous in UI design today. This concept resonated with me since, for a while now, I’ve been using the HazeOver application on macOS and I have come to appreciate the utility in having a foreground window stand out from its companions through the use of lighting effects.

Web Design

Breaking Out with CSS Grid Layout: Tyler Sticka shows us how to allow certain elements to fill the full viewport width from within a fixed-width container. Handy for beating the constraints of a fixed-width container for hero images, video and the like.

Morten Rand-Hendriksen walks us through Building Production-Ready CSS Grid Layouts Today.

Design Better Data Tables: in which Andrew Coyle illustrates patterns for data table design and visualisation.

Designing the Perfect Date and Time Picker: Vitaly Friedman explores, in depth, the design and implementation details of the perfect date/time picker. It’s interesting to note that, as in most things, one size does not fit all, it all depends on your use case.

A slider is a simple user interface control where you adjust some value by dragging a handle in a groove. In Implementing a slider well, Ilya Birman illustrates what does and does not work, culminating with a reference implementation.

From the Baymard Institute, interesting research and a compelling argument that The ‘Credit Card Number’ Field Must Allow and Auto-Format Spaces (80% Don’t).

Enjoy a hilarious journey through the ten things David Gilbertson learned while making the fastest site in the world.

You must check out this gorgeous, animated rendering of an iPhone 6. Created purely in CSS, not a scrap of JavaScript.

Web Development

Ire Aderinokun gives us the low-down on Asynchronous Functions and how to use them sequentially or in parallel with the versatile await command. Is this the end of callback hell?

JavaScript programmers will all be familiar with this. But what is this? Why is this sometimes a scoped object, then at other times the global object? Arnav Aggarwal explains this in The Simple Rules to ‘this’ in JavaScript. I’m glad we’ve got that cleared up.

While we’re on the subject, Ryan McDermott has put together a guide to producing readable, reusable, and re-factorable software in JavaScript.

Pure Geek

PopSQL looks interesting. I can’t count the number of times each day that I share or receive SQL queries from colleagues via HipChat. The collaborative options in PopSQL will be useful to me and my co-workers. I’m definitely going to give this one a try.

“Every time a PHP developer runs git diff on the command line a kitten dies.” So writes Igor Santos in Quick tip on Git diffs for PHP files. Santos goes on to explain how to make Git’s hunk headers a little more PHP-friendly.

Harry Roberts shares with us the “Little Things I Like to Do with Git.” Steady on old boy.

Whilst we’re fooling around with Git, check out diff-so-fancy which “strives to make your diff’s human readable instead of machine readable. This helps improve code quality and spot defects faster.”

In Shell Scripts Matter, we look at improving our shell scripts with static code analysis, “strict” mode, exit traps, unit testing, logging, debugging, self-documentation and version control.

Wolfgang Lutz has a great article on how to customise the small control strip on the Touch Bar MacBooks. With this documentation I have modified my own Touch Bar to feature media player functions on the small control strip. This is useful since I almost always have iTunes playing while I’m using the computer.

Security, Privacy and OpSec

In How Hackers Steal Your Data (and How to Make it Harder for Them), David Brownman gives us a five-point checklist to help us make our digital lives a little more secure.

A reminder that you should always look before you paste from a website to terminal — oh that naughty CSS.

When it comes to network security, most of the tools to test your network are pretty complex. Nessus isn’t new, but it definitely bucks this trend. It’s incredibly easy to use, works quickly, and can give you a quick rundown of your network’s security at the click of a button.

If you haven’t already, you should upgrade your SSH keys to Ed25519, a reference implementation for EdDSA using twisted Edward curves. While you’re at it, you may as well simplify your life with a SSH configuration file.

[0day] Bypassing Apple’s System Integrity Protection: abusing the local upgrade process to bypass SIP. Genius, as always, from Patrick Wardle.

Utilities

Alisdair McDiarmid has become one of my heroes since I discovered his Kill Sticky browser bookmarklet. I use this countless times every day and have quickly come to question how I kept my sanity without it (assuming, of course, that I am still sane).

Noteworthy

They Live and the Secret History of the Mozilla Logo;
Awesome Falsehood: a curated list of falsehoods programmers believe in;
Create a Waveform Image with ffmpeg;
It’s Not Dark Magic: Pulling Back the Curtains from Your Stylesheets;
Using CSS variables correctly;
Glitch Effect;
IndieWebify.Me: A guide to getting you on the IndieWeb;
WebPonize is a macOS application for converting images into WebP;
Getting started with variable fonts;
Level Up Your CSS Skills with these 20 Pro CSS Tips;
Moving beyond localStorage;
Fabulous macOS Tips & Tricks;
Tricks with xargs;
20 years on, the Tomb Raider story told by the people who were there;
15 Fundamental Laws of the Internet;
How CSS pseudo-classes work, explained with code and lots of diagrams;
More drawings about computers;
Grid layout is a much needed step-change for CSS;
@supports will change your life;
SVG Line Animation for the Uninitiated.

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

THE MAGICAL CAPS LOCK KEY

2017-07-16T16:42:42Z

tl;dr

The caps lock key is one of the most useful keys on my keyboard. I use the powerful Karabiner Elements software to empower the key with practical and time-saving functionality.

The Most Hated Key

There is a sizeable contingent of computer users who really dislike the caps lock key. This is understandable. This unjustly forsaken key commonly sits just to the left of the A key on a QWERTY keyboard. A is one of the most frequently hit keys for English writers. Thus one can appreciate that caps lock is often inadvertently triggered. This results in strings of uppercase characters, which the Internet interprets as SHOUTING, something that it frowns upon and which is generally considered to be poor form.

The caps lock key is a mode-switch toggle. You hit it and it puts the keyboard into uppercase mode, hit it again and turns off the uppercase mode. Did you ever wonder why the caps lock key has an LED on it? The LED is there because computer/keyboard manufacturers know how frustrating it is to type for a while, only to discover that you have inadvertently enabled caps-lock.

Some of my readers, if they are of my generation, will have vague memories of the mechanical typewriters of old. On such a machine, in order to enable caps-lock, one would depress the shift key, then the caps lock key to lock the shift operation. One would then depress the caps lock key again to release the lock. This was elegant and, well, logical. It made sense. Sadly, computer keyboards never operated this way (at least, none that I know of). If they did, then the average user might look more favourably on, or least be neutral towards caps lock.

A Basic Fix

Computer users sometimes go to extraordinary lengths to change the behaviour of their caps lock key. It is not uncommon to see this key remapped to control or escape, or even disabled entirely (some users take the nuclear option: physically removing the key or cutting its tracks on the keyboard PCB). Generally, one can remap the key in software. In macOS, for example, one can use the keyboard preference pane to remap caps lock to control, option, command or esc. Or one can choose “No Action” to disable the key entirely.

That’s all well and good. But suppose we want to go further than a basic remap…

Going to the Next Level

Back in 2015 — in an article entitled, “Keyboard Wizard” — I wrote about how I used a software combination of Seil and Karabiner to turn my caps lock key into a so-called hyper key modifier. This opened up a whole keyboard worth of new, configurable, non-conflicting keyboard shortcuts to me and became an invaluable component of my computing.

Then Apple released the macOS Sierra operating system. Unfortunately this broke the APIs or hooks that Karabiner used, effectively removing all of my (by now committed to muscle memory) bespoke keyboard shortcuts. It was a dark time indeed.

The changes introduced by Sierra were so wide in scope that Karabiner was effectively deprecated. The author, Takayama Fumihiko, started over with the entirely new Karabiner Elements. But, for the longest time, Elements supported only simple one-to-one key remapping. The multi-key modifier I had come to rely on was simply unavailable. I’d learn of the odd hack that would restore some of the functionality, but nothing came close to the configuration I’d come to know and love. Then a fork of Elements appeared which restored at least the hyper key functionality. Things were getting better.

Recently, Fumihiko released a new revision of Elements, featuring full support for complex modifications at last, and restoring equilibrium to my computing in the process.

Note: Don’t let the term “complex modifications” put you off, complex in this context simply means not a one-to-one key remap.

Ultimate Caps-lock

So, with all of that background out the way, let me tell you how my caps lock key behaves now:

When depressed on its own it maps to esc (useful on my Touch Bar MacBook);
When used as modifier (depressed in conjunction with another key) it acts as a hyper key;
When depressed in conjunction with the shift key (either left or right) it toggles the caps lock mode.

For myself, this is the perfect set-up. The caps lock key has gone from being the most irritating key on the keyboard, to being the perfect companion to my life as a keyboard jockey.

You should give Karabiner Elements a try. If you’d like to use the configuration I have described here, my karabiner.json file is a available as a Gist.

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

A Little Privacy Please

2017-06-11T16:11:26Z

I’m a long-time user of privacy screens. You know what I mean, those polarising filters that sit atop your screen and reduce the viewing angle from what is effectively 180° to just 60° or so. These filters make it difficult for your neighbours on planes, trains, or in cafés to see what you’re working on or to read the text on your screen.¹

Historically, I have opted for 3M filters as these work well, are widely available and are inexpensive. Unfortunately, the filters I have used in the past are all of the kind that either adhere directly to the screen or that slot into holders that one has to affix to the screen’s bezel. In both cases, the manufacturers tell us that the adhesives are removable and reusable. Sadly, in my experience, they are not.

Therefore, it intrigued me to learn of a new privacy filter by Stark Inc.² that affixes to the MacBook magnetically, thus being easy to remove and reuse.

I ordered one and have been using it daily for a couple of weeks now (in fact, as yet, I haven’t removed it from my MacBook).

The Stark screen is well made (in Germany). It’s thicker and more rigid than the filters I have previously used. It feels like a quality product.

At the top of the filter is the magnetic strip that binds to the magnets(?) or ferromagnetic material in the MacBook’s bezel. There is a cutout in this strip so that the ambient light sensor, camera and camera LED are not obscured. With this magnetic strip, the filter is ridiculously easy to remove and reapply.

So how about those viewing angles? Well, see for yourself.

I took the above photo from a position roughly where I estimated a neighbour’s viewpoint to be. The MacBook is on. Screen brightness is at 50%. Visibility is severely impaired. You can just about make out that there’s something on the screen if you look at the full size image. But you certainly can’t read any text.

I took this second photo head-on to the screen. Here too, the screen brightness is at 50%. Visibility is 100%.

What you can’t see from the second image is the large bay window behind me, with the sun high in the sky. You can’t see it because the Stark filter is anti-glare and anti-reflecting. It’s also scratch resistant.

Best of all, it’s reversible. One side has a glossy finish, the other is matte. You read that right. A matte finish. I can’t tell you how nice it is, when you do as much imaging on your computer as I do, to have a matte screen. How I wish Apple had continued to offer matte screens as an option.

Is this the best privacy screen I’ve used? No, that award has to go to the 3M Privacy Filter. However, the Stark privacy filter, without a doubt, is the most convenient of its ilk. For ease of application and removal, there’s nothing to touch it.

Pros

Limited viewing angle;
Simple to apply and remove;
Reusable;
Matte finish (reversible, glossy on the flip-side)

Cons

Expensive ($57 (US), plus shipping);
Darkens screen slightly (but then, they all do);
Cutout for camera is off-centre, so obscures the camera LED when the glossy side is up;
Magnetic strip is slightly too thick for the laptop to fully close.³

Important when you’re working secretly on the “the next big thing” or checking your embarrassingly negative bank balance. ↩︎
Iron Man, did you make this thing? 😃 ↩︎
This bothers me, but I always travel with my MacBook in a sleeve. One that’s tight enough a fit to keep the laptop as closed as it can be, so it hasn’t been a problem. However, I hope that Stark can address this in their next iteration of this product. ↩︎

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

WTF is Wrong with My Service Worker?

2017-06-11T13:02:00Z

I’ve had a service worker running on this website for the past ten months or so. It maintains a cache of files in the client browser so as to enable the site to run in an offline mode as per the current PWA paradigm.

As far as I’m aware, it’s been running flawlessly since I deployed it. However, last week I noticed that all was not well. The service worker wasn’t updating the caches correctly, or at all in some cases.

I’ve spent the past three days trying to get to the bottom of the problem. So far, to no avail.

Curiously, everything works perfectly locally.

If anyone out there can work out what’s going wrong and point me towards a fix, I would appreciate it.

The service worker source code is available as a Gist.

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

The Perpetual βeta Got a Search Engine

2017-06-11T09:46:38Z

I have used a static site generator — nanoc — to publish this website for some time now. I have enjoyed the benefits of a static website and experienced only a handful of drawbacks. However, one obstacle has bothered me for some time, and that was my inability to add a search engine to the site.

The principle reason why I couldn’t add a search engine was the result of an architectural choice I made early on. And it was this: there should be no server-side scripting. In eliminating a reliance on server scripts I’ve made the Perpetual βeta resistant to hacking (no software bugs to compromise); I have realised significant performance gains (it’s just static files all the way down); and the entire website is dependency free and highly portable (I can zip up the files and move them to another server in minutes).

It is axiomatic that a traditional search engine is reliant on software running on the server. These programs will maintain an index of the content, respond to POST and GET requests from the client, perform the search itself and generate a results page to return to the user. So how could I deploy a search engine, without having this back-end application available?¹

I found my solution in a third-party JavaScript library called LUNR.² This library takes a JSON-encoded corpus — which I have configured nanoc to produce at build time — as the source for indexing and uses a Solr-like scoring formula to generate its results.

What makes this approach so cool is that, being entirely client-side driven, searching is fast. We’re talking immediate results. With no round-trip to the server there is simply no delay in processing a search query. An additional bonus: one can still perform searches when the client is offline,³ which is great for PWA websites like the Perpetual βeta.

I have made the search engine available from the Archives page.⁴ The way I have implemented it is that queries/results happen in-the-page, they do not add to the browser’s history. However, the system appends a query string to the URL, making the results idempotent. Thus one can bookmark, link to and otherwise use the URLs, as you would with any other, and know that you will get the same results page back. For example: here’s a search for “amiga” and one for ”macOS”.

You might wonder if all this utility comes at a cost. Just how big is the search corpus the browser has to download and evaluate? At the time of writing, for this entire website, it’s 226KB. It’s less than 90KB when gzipped, which is equivalent to a standard, non-animated ad unit.

I discounted third-party options like Google Site Search as I don’t want advertisements in the results; nor do I want to introduce unnecessary, external dependencies. Additionally, I want complete control of the output. ↩︎
Via Jakub Chodounský. ↩︎
Although the client might not have, in her local cache, all the content that the search engine results might link to. ↩︎
Note: As the search engine is entirely JavaScript driven, it is not available if the user disables JavaScript. ↩︎

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

TLS Adoption is Growing, Fast

2017-05-21T18:44:15Z

One of the Perpetual βeta housekeeping tasks I run (semi) frequently, is a site-wide check of hyperlinks, internal and external, with the Integrity app.

Integrity reports broken links. I then investigate each one and attempt to determine if the linked resource has permanently disappeared — in which case I’ll try to find an alternative resource — or temporarily, which I’ll leave alone.

Amongst the reports that Integrity produces is a list of all redirected links on a site. So, if a resource has moved — and the webmaster’s on top of her game — the target website will have instructions to redirect calls to the old address ‘old-example.htm’ to the new address ‘new-example.htm’. A process that is effectively transparent at the client side.

I usually replace re-directed source links on Perpetual βeta with the destination address the redirect results in, as this removes just that little bit of latency and fractionally eases the load on the target server.

As it happens, I hadn’t run Integrity for some time (six months at a guess). So I recently fired up the app and let it analyse the Perpetual βeta. In the results two things struck me. The first was that there were only half-a-dozen broken links (which is certainly a surprise considering the usual atrophy of the Web). The second was a disproportionately high number of redirects on the Perpetual βeta’s 1,700+ hyperlinks.

Had half the Web relocated? Had I missed a significant Internet event?

Actually, the answer to both questions was, “yes, kinda.”

Almost every single one of these redirects was a protocol change from the plain-text, unsecured HTTP to the encrypted HTTPS.

It’s amazing to see so much of web adopting TLS and so quickly. Just two or three years ago this would have been unthinkable. Webmasters were able to argue against TLS deployment with good reason: cost; expensive processor cycles for encryption/decryption; implementation barriers (including those of getting advertisers and other third-parties on-board); and so on.

Yet here we are, with a Web where now it seems that HTTP is odd man out, rather than its counterpart.

The efforts of parties like Let’s Encrypt — in making TLS effectively free — combined with concerns raised by the Snowden leaks and growing revelations of state-sponsored surveillance; in conjunction with players like Google touting increased search relevance to secure websites; and now vendors revealing plans for, or deployment of, browsers that highlight unsecured websites… all are, I’m sure, factors in the increasing dominance of TLS-secured Web properties.

It’s taken us a long time to get here, but we should all be grateful that we are, and that the rate of TLS adoption continues.

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

Thoughts on the Late 2016 MacBook Pro

2017-02-20T22:09:11Z

When Apple announced the new MacBook Pro, on 27^th October 2016, I placed my order immediately. I took delivery of the machine around three or four weeks later. I’ve waited until now to write about it as I wanted to have some time with the computer first. I wanted to build up some real-world experience with it and get a software update or two into the mix to iron out the bugs that I assumed would there (the first version of anything always has issues).

Mine is a BTO 15″ model, with Touch Bar. It has the 2.9GHz quad-core Intel Core i7 processor with Turbo Boost up to 3.8GHz and the Radeon Pro 460 with 4GB memory. I also upgraded the PCIe-based SSD storage to 1TB. I wanted to max out with the 2TB option but that would have taken the price over £4,000 — and that was just a little too much for me to stomach — particularly with storage being so inexpensive outside of the Apple Reality Distortion Field.

Now that I have some time with the new MBP under my belt, I thought I’d write up my notes. My thoughts on this oh-so-controversial machine.

The Engine Room: CPU, RAM and SSD

Unfortunately I didn’t benchmark my previous MBP, with the jobs I regularly perform. So I am unable to offer a side-by-side comparison. I can say that, subjectively, this MBP feels a little faster. But the difference is more, “oh, that’s finished already” rather than “bloody hell, that was fast!” The speed increase doesn’t take my breath away, but is noticeable with some tasks.

For those who care about such things, the Geekbench scores for this configuration are: 4,379 (single core); 13,760 (multi core) and 47,908 OpenCL. Novabench results in a respectable overall score of 1,217.

Therefore, it’s fair to say that processor and GPU performance are (at the time of writing) competitive with other laptops in this rarefied price range.

With the SSD things are much more newsworthy:

1,944 MB/s write and a massive 2,665 MB/s read speed with a 64GB 5,120 × 2,700 5K RED file. Holy crap. That image says more than anything I could write about this superlative disk sub-system.

Now… let’s talk about RAM. The 16GB upper limit on RAM in this model was widely criticised (myself included). Then I read Jonathan Ździarski’s ‘Can We Put the 16GB “Pro” Myth to Rest?’ and thought, “maybe it’s not so bad.” Having said that, part of me still feels underwhelmed by 16GB. I would have liked to have had at least the option for more RAM. At the same time, I want my laptop to run all day on a single charge. More RAM draws more power and that equals a reduced time on battery. It’s a conundrum.

With the SSD performance of this new MBP we have a lot of leeway with RAM because paging in/out of this SSD is so damn fast. That’s not to excuse the 16GB limit, simply to say that it doesn’t compromise my workflow (and I consider myself to be a “professional” user).

Force Touch Trackpad

The most noticeable aspect of the MacBook Pro’s trackpad is its size. It’s big. My iPhone 6s Plus fits comfortably inside its perimeter. Yet, it doesn’t get in the way. Palm rejection works well (but not perfectly) and I haven’t had any significant issues related to the trackpad’s dimensions.

My previous MacBook had a mechanical trackpad, this one has a Force Touch variant which also features Apple’s “Taptic Engine”. This is my first experience with this technology and I find it impressive. I did not expect it to feel as convincing as it does. The entire trackpad surface is clickable, right into the corners. I could not, through touch, distinguish between a click on my old MacBook and one on the new. It’s impressive stuff. As with all Apple trackpads, it blows the equivalent non-Apple hardware of the water. It’s accurate, responsive, lag-free and the surface texture is nothing short of perfect.

Keyboard

The keyboard has suffered so much criticism that I have to wonder: who of those who condemn it, have actually used it for a sustained period?

I have already written tens of thousands of words on this keyboard and I have found it to be easy to live with. The travel is short, that much is true. But the keys are larger than those of my previous model and they are almost perfectly spaced. Each key is slightly concave, which means my thick fingers more naturally find their centres.¹ Best of all is the distinct click each key makes when pressed. I have spent hundreds of pounds on mechanical keyboards in the past to get such a click. I love the positive feedback from this keyboard and find myself typing more assuredly on it than I did on its predecessor, or the wired, full-size Apple keyboard I often used with it.

As with all MacBooks, the keyboard offers variable backlighting. This is user configurable and responsive to ambient light levels. Unlike my previous MacBook, light leakage on this new keyboard is virtually non-existent. This makes the keyboard look just that little bit more premium than before.

The fly in the ointment, albeit a minor one, is the new arrangement of the arrow cluster. The ← and → keys are larger than the previous incarnation, so their height now matches that of the ↑ and ↓ keys combined. It looks prettier but, for one who routinely navigated those keys by touch, aided by their dimensions, it has taken some getting used to.

Oh, and the esc is missing. More on that below. 😀

Touch Bar

The headline feature of the late 2016 MacBooks is undoubtedly the all new Touch Bar. According to Apple:

“The Touch Bar — located above the keyboard on supported MacBook Pro models — is a Retina display and input device that provides dynamic interface controls for interacting with content on the main screen. These controls offer quick access to system-level and app-specific functionality based on the current context.”

The first thing I noticed about the Touch Bar was that some of the application defaults are weird. Take Safari for example: by default, a scrubber featuring microscopic previews of that browser’s open tabs, occupies the largest portion of the Touch Bar. Now, given that I’ll often work with twenty or more open tabs (and have the eyesight of a man closing in on fifty), this scrubber is almost entirely useless. It may look good in demos and photo-shoots, but that’s about it’s only use.

Fortunately the Touch Bar is highly configurable, even for the macOS stock apps. My customised Safari Touch Bar, for example, is much better suited to my workflow than the eye-candy default.

Left: escape, previous page & next page

Centre: new tab, show tabs, show sidebar, show favourites bar

Right: expand control strip, media & volume controls

Therefore, when switching to a Touch Bar MacBook, you should anticipate spending some time configuring it to your particular requirements. The Touch Bar becomes so much more useful when you’ve tweaked it.

It’s worth pointing out that compatibility with the Touch Bar is not universal, yet is growing rapidly. The MAS informs me at least once a week that one app or another has an upgrade that includes support for the Touch Bar. But some older apps, not actively maintained, will likely never offer such an update. However, with a little effort, one can coerce even these incompatible apps to play nice with Apple’s new shiny.

The Touch Bar is a first generation product and it shows. That’s not to say it doesn’t work, it does. What I mean to say is that it needs more work. It’s way too easy to accidentally activate something (the number of times I’ve inadvertently muted audio is beyond counting). I contend that haptic feedback is an absolute must have for the next generation Touch Bar, and the user should be able to choose, in System Preferences, whether it should respond to touch, or to a press (and I’d choose press, every time).

So what about the esc key? You know, the one that’s been on keyboards forever. There was much angst in the user community when it became obvious that Apple would replace the physical esc key with a soft key on the Touch Bar. But, guess what? There is no justification for the apparent panic. The esc is still there, in its usual place — top left. Yes, it’s on the Touch Bar. No, that isn’t a problem. I’m a programmer and a self-described keyboard jockey. I use the esc key… a lot. I can honestly say that, the number of times its replacement with a Touch Bar soft key has caused me a problem, is exactly zero. It’s a non-issue. The esc key lives. It’s still omni-present — at least when I need it — and, additional bonus, its relative size is larger than before since its contact area is larger than the visible surface area.

Touch ID

One of the gems of the late 2016 MacBook is Touch ID. Familiar to iPhone users the world over, Touch ID is a biometric authentication system, with fingerprints as its access key. As with the iPhones, it works flawlessly and is a wonderful, time (and frustration) saving system. When tied in with 1Password, as on my system, it’s perfect for secure authentication with the myriad websites we all interact with on a daily basis.

One of my concerns, before receiving the computer, was how configurable — or otherwise — Touch ID might be. I was absolutely certain that I didn’t want to login to my computer with my fingerprint (I prefer my trusty, high-entropy password for that), but I knew that, once logged-in, I’d appreciate the convenience of Touch ID. Therefore I hoped that I could configure the system to suit my preferences, otherwise I’d probably not use Touch ID at all.

I needn’t have worried. Touch ID has a range of options. It provides for a traditional login along with Apple Pay, iTunes and App Store purchases. And, as I mentioned above, third party apps (like 1Password) can use it too.

There are, however, some inconsistencies in its implementation. Some System Preference panes accept Touch ID authentication while others require a traditional password. This is mildly irritating, but I suspect a further software update or two will bring these remaining components in line.

All in all, Touch ID is a positive addition and brings much to the macOS user experience. It speaks volumes that when I use my non-Touch ID Mac in the office, I really notice that it’s missing.

Screen

In addition to the new features, Apple upgraded the display of the late 2016 MacBook. According to Apple, the screen is brighter (up to 500 nits), has 25% more colours than sRGB (wide colour) and a 67% higher contrast ratio than that of the previous generation MBP.

I can’t really see any difference between this and the late 2013 MacBook Pro I have upgraded from. But then, as I wrote earlier, my eyes aren’t what the once were due to presbyopia. I’ll say this, I have no complaints at all about the display.

There’s something else: Apple have increased the native resolution/scaling in comparison to my previous MBP. It’s now practical to work with two half-width, adjacent windows on this laptop’s display. Something that never seemed workable on its predecessor.² This fact alone makes for some serious productivity wins.

Left: Typora; right: Safari

Thunderbolt 3 / USB-C

Critics have voiced their disgust at the replacement of all legacy ports with the new Thunderbolt 3 / USB-C ports on this generation of MBP. Back in October, I wrote about my enthusiasm for the new port. After these first months of use, I am happy to report that my zeal hasn’t waned. I can truthfully declare, I am more of an advocate for Thunderbolt 3 / USB-C now than ever before.

Let’s look at the benefits of this mighty port once again:

Daisy-chain-able;
Reversible connector;
Universal (plug anything into any port);
Bi-directional power on any port;
Diminutive form factor;
Two streams of DisplayPort 1.2 video bandwidth per port;
Up to 100W of power;
Networking up to 10 Gb/s
Thunderbolt 3 throughput up to 40 Gb/s
Backwards compatibility with Thunderbolt 1 and 2 (20 Gb/s), along with USB 2.0, 3.0 (5 Gb/s) and 3.1 (Gen1/Gen2) @ 10 Gb/s

Make no mistake, this is a game-changing, killer interface.

So, what about #donglelife? Well, for myself, it’s been a non-issue thus far. I bought half-a-dozen Nonda MI22SGRN USB 3 → USB-C adapters, which I leave permanently attached to the cables of the devices I connect to my computer.

I have a single dongle attached, a Belkin USB-C to Gigabit Ethernet Adapter (I prefer wired networking where possible). I also used an Ethernet dongle with my previous MBP, so there’s no change here.

I never read SDHC cards via the integrated reader on my previous MBP as I used that slot to extend the storage of the machine. So I have always used either an external reader, or plugged my camera directly into the laptop. In this case, I did end up buying a new card reader. I had to. My old reader plugged directly into the computer and despite the diminutive Nonda adapter, it ended up obstructing the adjacent USB-C port. For good measure, I also threw a Nonda adapter on to my camera’s USB cable. So I have all bases covered.

To charge my Apple TV remote, I bought an Apple USB-C to Lightning cable. To connect my iPhone to the MBP I use an Anker PowerLine+ Lightning Cable with yet another Nonda adapter. My external HDDs: you guessed it, all are on their existing cables with Nonda adapters.

As for external displays. I haven’t as yet used one with this computer. So I’ll address that as and when the need arises.

A final thought on the ports: Apple should have illuminated them (like they did with the Mac Pro). USB-C boasts one of the smallest I/O ports and, in low light, with a space grey computer, they are just about invisible. Also, there should be a larger gap between them. The Nonda has a small footprint, even so, I can’t plug two in side-by-side… and that’s a shame.

Battery

The battery in the new MBP has had a roller-coaster ride in both the mainstream press and the blogosphere. Some have commended it, while others have criticised it. Consumer Reports published their much-discussed review which, for the first time, failed to give a new MBP a “Recommended” status, due to their experiences with MBP battery life:

“The MacBook Pro battery life results were highly inconsistent from one trial to the next.”

“For instance, in a series of three consecutive tests, the 13-inch model with the Touch Bar ran for 16 hours in the first trial, 12.75 hours in the second, and just 3.75 hours in the third. The 13-inch model without the Touch Bar worked for 19.5 hours in one trial but only 4.5 hours in the next. And the numbers for the 15-inch laptop ranged from 18.5 down to 8 hours.”

“Those were just a few of the results; we tested battery life on these laptops repeatedly.”

Apple subsequently released a software update which apparently fixed a bug in the Safari browser that was responsible for the battery issues. Consumer Reports then changed their position on the new MBP to “Recommended.”

The battery in the new MBP is ~25% smaller — at 76 watt-hours — than its predecessor (99.5 watt-hours). The Laws of Thermodynamics tell us that this will affect the run-time on battery power, unless we can reduce our energy demand. The new MBP draws less power than earlier generations, as hardware components have become more energy-efficient. When we combine this with improvements in software power management, the new MBP can achieve similar or better run-times than its predecessors — under similar operating conditions and with comparable workloads. Put the battery system under stress — with a high power draw — and the consequences of a reduction in physical battery size will become manifest.

My own experience with battery life on the new MBP reflects this. I’ve run for more than seven hours³ on a single charge with low-energy workloads: web-browsing (with Safari), coding (with Visual Studio Code), writing (with Typora). But then I also have some high-energy tasks that I regularly perform: image processing, machine learning, parallelised number crunching. When I’m running with all eight processor threads maxed-out, I tend to ensure that I have mains connectivity, because I simply wouldn’t expect the battery to provide power for long under such a heavy load.

All eight processor threads are running flat-out here, during parallelised processing of a 1.5m row CSV data-set

Speakers

In describing the speakers of the new MBP, Apple tells us that they offer 58% more volume; up to 2.5× louder bass; 3× more peak power and 2× dynamic range, when compared to the previous model. These are mighty claims. All laptop users know that inboard speakers are always a compromise, limited as they are by the physical dimensions of the laptop chassis and the lack of space within.

However, with the new MBP, Apple have raised the bar as to the sound quality you should expect of a laptop. You’re not going to be throwing parties with the MBP as the music source, but you can certainly enjoy music direct from the machine, without using external speakers or headphones.

Stealth Mac

The late 2016 MacBook Pro is a beautiful machine. The 15″ model that I have, seems impossibly thin and light. I know, from Apple’s published metrics, that the computer isn’t that much thinner or lighter than the machine it replaces. But, subjectively, it feels a world away.

In space grey, the computer is discreet. It no longer screams “MacBook!” It no longer draws attention to itself. Unlike other social media commentators, I was happy to see Apple do away with both the illuminated Apple logo on the lid, and the attention-grabbing startup chime.

It is now possible to open and boot the MBP in, say a library, without anybody noticing. It’s almost like having a stealth Mac and I, for one, am happy that Apple made these small changes.

Conclusion

The late 2016 MacBook Pro is a decent laptop. It’s not the fastest machine in its class, but is a solid performer. It’s well built (as far as I can tell) and looks gorgeous. Touch Bar, Touch ID and USB-C / Thunderbolt 3 are respectable new features. The keyboard, trackpad and display are all best-in-class and worthy upgrades from my previous model.

On the flip-side, the computer is massively overpriced. The BTO model I specified came in at an eye-watering £3,329.00. Add in an AppleCare Protection Plan — and you better, since this computer has no user-serviceable or upgradeable components — at £329.00 and, let’s say £50.00 for adapters/cables/dongles and your total expenditure sits at around £3,700.

That’s a fair chunk of money. I could rationalise such a large expenditure only because I make my living via my computer and the machine occupies a large slice of my spare time too. But I feel a little cheated. I feel like Apple sacrificed too much for the sake of making a thin computer. I would have gladly taken a thicker MBP if it had offered a 99.5 watt-hour battery like its predecessor. Additionally, I would have appreciated an option to upgrade the RAM to 32 or even 64GB (yes, even at Apple’s exorbitant prices).

In general, I expect to get three years of service out of each computer I buy. With the new MBP I know that in twelve months — maybe even fewer — the specifications of this machine are not going to hold up as well as they have with previous iterations.

I’ll close with this: I’m a loyal Apple customer and have been for more than two decades. This is the first major purchase I’ve made, from Apple, that has left me feeling a little buyer’s remorse. If this computer had cost me £2,700 rather than £3,700 then I’d be raving about what a great machine it is. As it is, it feels like a little “fuck you, suckers” from Apple and I can’t help but feel a little disappointed.

Feedback? Discuss on Hacker News.

I think this is new, I’m almost sure the keys on my previous MacBook were flat. ↩︎
Even on my previous MBP, I was always able to work with two adjacent panes in my code editor and iTerm, but they were about the only two apps that I was able to realistically work with in this manner. ↩︎
Keep in mind that, with the maxed-out processor and GPU, my MBP will draw more power than a stock machine. ↩︎

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

macOS: Fixing an Auto-hide Dock Bug

2017-01-15T20:17:58Z

When Apple announced the late 2016 MacBook Pro I got my order in immediately. Around four weeks later I took delivery of a sexy space-gray super-computer and got stuck into migrating from my previous machine.

Mine is a built-to-order MacBook, upgraded with the fastest processor, 1TB SSD and the 4MB AMD Radeon Pro 460.

Like others, I suffered the odd graphical glitch with the 460. For the most part they were minor annoyances and I figured they were driver related and that Apple would fix them with a software update. Sure enough macOS Sierra 10.12.2 addressed all the issues I’d experienced, bar one…

When I used the Dock — set to auto-hide — to switch to a full-screen app in another Space, the app switch would take effect but the Dock would freeze in the new Space (i.e. auto-hiding wouldn’t close it).

If I were to say this wound me up, you can consider that an understatement. Anyway, I spent some time exploring the timings for the auto-hide and I have finally found a combination that eliminates the issue:

defaults write com.apple.dock autohide-delay -int 0;
defaults write com.apple.dock autohide-time-modifier -float 1.0;
killall Dock

Now, my Dock works perfectly.

Image Credit: MesserWoland based on the graphics by FoeNyx. Own work, copyleft: Multi-license with GFDL and Creative Commons CC-BY-SA-2.5 and older versions (2.0 and 1.0)

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

macOS: Closing the Gap with Moom

2017-01-15T18:24:23Z

I use Many Tricks’ Moom application to handle window management on macOS. Moom provides for complete control over a window’s size and position, based upon a user-defined grid. It’s also controllable entirely from the keyboard, which suits my needs perfectly.

One of Moom’s tricks is “move and zoom to full screen” which works the same way as macOS’s maximise control (option+click a window’s full-screen widget) except that it can be keyboard driven.

However, Moom’s “move and zoom to full screen” shares a niggling “feature” with macOS’s maximise: it leaves a small gap between the window and screen edge — on the Dock side — when one uses an automatically hidden Dock, as I do. It looks like this:

So, not exactly a maximised window then. This has always irritated me, until today. By chance I learned that Moom has a hidden switch, controlled via the command-line, that instructs the software whether or not to ignore this Dock clearance zone:

defaults write com.manytricks.Moom "Ignore Dock" -bool YES

Boom! That irritating gap is no more. Now “move and zoom to full screen” does exactly what I expect and positioning smaller windows to the Dock edge also ignores the gap.

It’s a small thing really, but it makes all the difference in the world.

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.

Adding Survivability to macOS Applications

2017-01-15T16:19:09Z

There are a handful of utility programs that I have come to depend on in my day-to-day computing with macOS. These include:

Flycut: a simple yet effective clipboard manager (a fork of JumpCut, for those of you with memories as long as mine).¹
Contexts: a drop-in replacement for the command+tab switcher. I use this, as opposed to macOS’ built-in switcher, as it lists all the windows for each app (even across Spaces). This is helpful when you run with multiple app windows, as I do.
Moom: a best-of-breed window manager. Because, again, I tend to have a lot of windows open.
QBlocker: prevents you from accidentally quitting applications by requiring you hold command+Q to quit (with a user-configurable delay).
Snappy: a screen-capture application. macOS has great screen capture tools out of the box but, Snappy has a handy feature in that a snapshot, taken with the app, will float atop all open windows (even across Spaces), for as long as needed. This reduces the amount of command+tab switching I need to perform and is a fantastic time-saver.

I want these apps to start when I login and I want them to remain active through the duration of my computing session. Guaranteed availability is important because these apps all share one common ingredient: they each make all of their principle functionality accessible to the keyboard interface. I have committed their keyboard shortcuts to muscle memory. So, if one of these apps crashes and I’m unaware of it, I have this jarring context shift when I issue a keyboard shortcut that isn’t then handled, or the system handles it differently to my expectations. One of the most consequential is that which occurs when QBlocker crashes. In which case my next hold of command+Q quits multiple applications rather than one (due to key repeat).

Sadly, applications do crash. So, in the case of essential apps, wouldn’t it be nice to have the computer restart a crashed program without any user interaction? We can have the machine do just that, with launchd — macOS’ unified service-management framework.

We start by, counter-intuitively, removing the apps from System Preferences → Users & Groups → Login Items (or by unticking their launch on login preference if they have one).

Next, we determine the path to each app’s executable by right-clicking the app’s icon and selecting Show Package Contents from the context menu. Click into Contents → MacOS to locate the name of the executable. For example: Snappy’s executable, on my machine, is at /Applications/SnappyApp.app/Contents/MacOS/SnappyApp

Open a terminal window and enter: sudo nano /LaunchAgents/com.keeprunning.snappy.plist then, paste the following into the nano editor:



 version="1.0">

  KeepAlive
  
    SuccessfulExit
    
  
  Label
  com.keeprunning.snappy
  ProgramArguments
  
    /Applications/SnappyApp.app/Contents/MacOS/SnappyApp
  
  RunAtLoad

Followed by control+O, return to save and control+X to exit.

Finally, we register and load our new plist:

launchctl bootstrap gui/`stat -f %` /Library/LaunchAgents/com.keeprunning.snappy.plist

Wash, rinse, repeat for each application that we require to be “always on” and we’re good to go.

The key commands of our plist file are:

KeepAlive → SuccessfulExit → false: this tells launchd to keep our application running unless it has a successful exit (that is, we intentionally quit it)
RunAtLoad → true: instructs launchd to ensure our application runs as soon as possible during the boot/login sequence.

Fault tolerance for the win. 😎

Via: TJ Luoma.

Why, in 2017, a clipboard manager is not built-in to macOS is beyond me. ↩︎

If you liked this post you should visit the Perpetual βeta website, where you will find further essays by the author.

Formatting messed up in your feed reader? You should view this article in your web-browser.