WatchOS 5 adds support for web content rendering

Apple recently announced updates to the core software on all of their hardware platforms with various interesting new features.

One feature that jumped out when looking through it all was support for displaying web content in watchOS 5, it’s important to clarify that they haven’t added a standalone Safari app to watchOS but instead it enables any links sent via Mail or Messages to be accessed and then displayed right on the watch.

To get a quick overview it is worth taking a few minutes to watch the “Designing Web Content for watchOS” video on the WWDC2018 videos site as it gives a good overview.

Here’s a few thoughts and info about key aspects that I picked up from watching the video:

User navigation / interaction

  • You can scroll using the Apple Watch’s digital crown or via pan gestures
  • Double-tap to zoom in / out on the page content
  • Back/Forward navigation is controlled either via an overlay UI brought up via a  firm press on the screen or by swiping back and forward from the edges of the screen.

Web browser feature support

Content is optimised for display on the Apple Watch so certain features are not supported in watchOS 5:

  • Video playback
  • Service workers
  • Web fonts

If the web content being accessed is responsive then it treats the content as being 320px wide, the same width as if on an iPhone SE (iPhone 5 or older width). So text may be smaller but at least it will basically render the smallest breakpoint content, so it doesn’t require any new even smaller breakpoint to be catered for.

This is done by overriding the “initial-scale” value and provides a viewport with the dimensions 320px by 357px and reports a media query size of 320px. So existing responsive content will render on the Apple Watch without requiring any changes – at least from a layout perspective, worth noting the lack of support for Web fonts as this will likely have some rendering impact as it falls back to alternative fonts in the font stack do display.

Optimising content for Apple Watch

Even though responsive content will be rendered quite well by default it is possible to optimise content for display on Apple Watch.

The above image shows the standard responsive content being displayed on the Apple Watch, basically just the same as it would be on an iPhone SE (minus any web fonts of course!).

Responsive Layout on Apple Watch

Using a media query it is possible to modify this layout to display as a single column. There is an example given in the video which obviously won’t apply for all uses, but basically it uses “min-width: 320px” as the baseline for showing the content as two columns, so any content below that would render as a single column. Again, how this works specifically for your layouts will vary, but there will be some methods to use for frameworks like Foundation or Bootstrap etc.

“Disabled-adaptations” meta tag

The important addition to using a media query though is a new meta tag which disables the default adaptations that the Apple Watch makes when rendering content by default:

<meta name=”disabled-adaptations” content=”watch”>

With this meta tag in place the device width will be treated as the real width of Apple Watch’s screen. This again has similarities to how content was handled when the iPhone originally came out, existing content is displayed as best as possible but there are ways to optimise for the device if you want to.

Form controls on watchOS

Making use of HTML5 form control types is really important on watchOS, setting the type attribute to “email”, “tel” etc will bring up a specific, full screen UI to allow interaction.

Additionally making use of labels, placeholder or aria-label attributes enhance the context given when interacting with these controls. Hopefully you’re using these already but here’s another reason to do so.

Safari Reader on watchOS

This is a feature found on iOS and macOS which basically formats pages to show a more readable version of web page content. It’s a little unclear from the video but it sounds like pages that are “text heavy” will get displayed using Reader, although I’m not 100% sure how that would be determined exactly if so. Perhaps this is a way to handle big pages that might have a lot of adverts on it? Reader view is an option that users can choose by firmly pressing on any page to bring up the navigation overlay, so even if content is displayed normally a more readable version can be accessed.

Semantic markup in Reader view on watchOS

Reader view makes good use of semantic markup, using the “article” tag helps the display of content, and attributes like “item-prop” and other semantic tags like “strong”, “em”, “blockquote” etc enhances the display of content in Reader on watchOS.

Open Graph meta tags

Using Open Graph meta tags is something that makes sharing content around the web such as into Facebook, Twitter etc look better by providing specific preview content such as images, titles etc. watchOS makes use of these Open Graph meta tags to make the previews for any shared links look as good as possible.


That’s a quick overview of some aspects of watchOS 5’s support for web content, there’s definitely a few things to consider in there but if you’re building pages using responsive layouts and using semantic HTML then things should work fairly well without having to do anything.

The biggest issue I see initially is the lack of support for web fonts, that seems like it could cause some display issues due to the fallback to alternative fonts in the stack or if web fonts have been used for icons etc.

I’m also interested to know what the impact on battery life on the watch is like when loading and rendering multi-megabyte web pages which are not uncommon these days, I think Reader view is going to be an essential feature for viewing web content on Apple Watch.

How to fix “unable to boot device because it cannot be located on disk” in Xcode 9

I have recently been doing some development using the Xcode 9.2 Beta, but when it came to pushing the app to iTunes Connect I made the switch to the final release version of Xcode 9.2. In the process I ran various clean up tools to remove the beta version and related files etc.

However, after downloading and installing the release version of Xcode 9.2 and trying to test using the Simulator I encountered the error:

unable to boot device because it cannot be located on disk

This error also included a link to a location within simulator files which obviously is no longer there due to the clean up process I had used. I tried a few options but eventually found a series of commands that clean this up and fixed the problem for me. Make sure to quite Xcode and the Simulator app first and then run these commands in Terminal:

sudo killall -9 com.apple.CoreSimulator.CoreSimulatorService
rm -rf ~/Library/Developer/CoreSimulator
rm -rf ~/Library/Developer/XCPGDevices

Hopefully after running them and starting Xcode the Simulator devices should all be functional again. Please note that if you have done any custom devices within your Simulator device list then these will have been deleted and the list reset to the default list of devices.

Rise of the Machines

As we move further into a post-Flash(1) web there’s a lot of excitement and hype about HTML5 and CSS3. These new technologies combined with the power of JavaScript frameworks like jQuery offer the potential to create much more dynamic, animated, interactive user experiences on the web.

Many tools are now being developed to make the creation of these experiences much easier (Edge, Sencha, Hype, MotionComposer, Animatable(2) etc). However, there’s a potential danger lurking under the surface with these apps and the many demos and samples showing off these new shiny features: Machine Generated Code.

Made in the image of Flash

Any serious web developer would not consider the HTML code rendered by tools like Apple’s iWeb to be clean or optimal but rather to be bloated and messy. With the rise of ‘HTML5’ and the previously mentioned apps it seems that some normally standards-aware developers are being swept along with the buzz, happy that you can now do all these cool animations that are ‘standards based’. But if you take a look underneath the surface and examine the code being output then there are causes for concern — a high or total dependency on JavaScript to generate HTML content, bloated ‘divvy’ code and no consideration for fallback content or graceful degradation.

The content created might be made using web standards like HTML5, CSS3 and SVG but in reality it fails in exactly the same way that Flash does — poor / non-existant semantic structure and a lack of accessibility(3).

Web apps

Another form of Machine Generated Code that is manifesting on the web is within many popular web apps. Take a look at the code under the hood of Google+ and you’ll see something like this:

Is that code any better than the code that is generated by iWeb? It’s possibly worse. How can a company like Google who has incredibly talented developers end up with code like that? Well, in Google’s case it was written using a templating system that uses Closure, these templates are converted by java on the server into HTML(4). Nothing inherently wrong with that process but the end result is HTML that is machine-friendly but is not really intended to be easily readable by humans.

So it seems that with the ‘rise of the machines’ demonstrated by tools like Edge, Sencha and Hype along with development frameworks like those used by Google that there is a loss of focus on the best practices of clean code, graceful degradation, progressive enhancement, accessibility. Instead the focus is on easy generation of code and content, with Edge and Sencha giving a visual WYSIWYG-ish interface with which to create it, and Google’s Closure / Java dev process enabling code-savvy devs to write in a non-web native language and have this automatically generated or compiled into the final HTML and JavaScript.

(Of course a counter-argument is that clean semantic HTML code doesn’t apply to web apps as they are ‘more complex’ than ‘regular’ websites! That’s something for further discussion).

Change begins at home

So, is there a solution to this? Well, the first solution is to consider these issues yourself. If you’re evaluating tools like Edge, Sencha or evaluating development frameworks then pay close attention to the code they create. If it seems to be falling short then take the time to give feedback to the companies behind them. In the case of Adobe’s Edge tool which is currently an early preview release on Adobe Labs then there is an opportunity to give feedback on the app and influence it at an early stage of development. Don’t just whine about it on Twitter but take the time to give feedback and clearly explain what you think is wrong and how it can be improved.

Muse

Talking of people whining on Twitter, I wrote this article and had it all pretty much ready to publish when Adobe released another product on Adobe Labs, Muse. Although the main focus of this article is on the more interactive, dynamic flash-esque content creation applications that have been appearing I couldn’t really post this article without briefly mentioning Muse (I did mention iWeb and talk about Google+ after all, both of which obviously don’t fit that category of app).

Muse definitely renders Machine Generated Code as it’s sole aim is to enable visual designers to “Design and publish HTML websites without writing code”. I have to admit that the first thing I did when viewing the Muse website (which is itself built with Muse) was to view source, and sadly I was disappointed with what I saw there – lots of nested divs, generally bloated looking content, and despite purportedly being ‘HTML5’ there were no HTML5 structural tags to be seen.

Many WYSIWYG apps have come and gone and have never managed to produce good clean code without requiring the user to intervene and clean up code, this doesn’t mean it’s impossible to create such an app, but it’s definitely a big challenge. As with Adobe Edge I’d recommend taking time to give good, reasoned feedback about what you consider is wrong with the output it produces, don’t just complain about it on Twitter.

Christian Heilmann wrote an interesting article called “Getting rusty – we need new best practices for a different development world” which I think gives an interesting pragmatic perspective on the debate about Machine Generated Code and why we rightly consider clean, semantic code desirable as opposed to that produced by apps like iWeb, Edge, Hype, Muse etc. His final couple of paragraphs sum up the thoughts and feelings that inspired me to write this post:

“Using web standards means first and foremost one thing: delivering a clean, professional job. You don’t write clean markup for the browser, you don’t write it for the end users. You write it for the person who takes over the job from you. Much like you should use good grammar in a CV and not write it in crayon you can not expect to get the respect from people maintaining your code when you leave a mess that “works”.

And this is what we need to try to make new developers understand. It is about pride in delivering a clean job. Not about using the newest technology and chasing the shiny. For ourselves we have to understand that the only one who really cared about our beloved standards and separation of concerns is us – as we think maintainability and not quick deployment and continuous iteration of code. The web is not code – the web is a medium where we use a mix of technologies fit for the purpose on hand to deliver a great experience for the end users.”

Footnotes:

  • 1: By post-Flash I mean as the only way to do interactive animated content on the web.
  • 2: Animatable looks to be one of the most promising apps being developed as the team behind it have kept the whole concept of falling back to non-animated content as core to the way the app works, basing it on actual HTML content rather than just created through JavaScript.
  • 3: I acknowledge that it is possible with recent versions of the Flash player to make Flash content that is more accessible, but generally Flash’s integration with Screen readers like Jaws etc is not great, and you have to specifically apply settings to the Flash content to help it be accessible.
  • 4: This information about Google+ development and other interesting info from this thread.