Tag Archives: foodml

The web without HTML

On Twitter, Alex Fajkowski responded to my blog post about tvOS with this:

“They have access to the open web—NSURLSession exists. HTML rendering is inappropriate for the watch and tv.”

I disagree with both of those sentences. Maybe Alex didn’t read my full post, because I wrote that web services are not enough. HTML and links and URLs are equally important parts of the open web. NSURLSession gets you web services but nothing else.

(As an aside, HTML turns out to be a pretty useful format for styling text, too. Why wouldn’t you want to use it for iTunes movie descriptions on the TV, for example? That seems completely appropriate.)

Think about the full scope of the internet. What percentage of content is available via web services — that is, structured data that can be parsed and displayed with a custom, native UI — compared to all the traditional, HTML-based web sites? You’ll find that there is an almost unimaginably large number of that latter kind of web site, and the only way to access and display that content is with an HTML renderer.

Now imagine a world with only native apps. You’d need custom apps and web services for different kinds of content, just as we have native Twitter or Instagram apps today, but we’d need these for many thousands of categories: tvOS with TVML, recipe or cooking apps with FOODML, and so on. Eventually, having so many formats would get unmanageable. We’d need to invent a general purpose format that could accommodate many app formats, and (surprise!) that general format would look a lot like HTML. Why break old content and essentially reboot the web, when we already have a capable format in HTML?

Of course, there’s no immediate risk of getting to that hypothetical native-only future. But when a company with the size and influence of Apple has 4 major platforms and only 2 of them have access to the open web, that should give us pause. Let’s reflect on how this plays out, so we can get back on track if the web does become marginalized.

John Gruber commented that these new devices don’t need the web at all, comparing it to the original Mac shipping without a command line interface. I realized while reading his closing paragraph that my own blog post had been poorly titled, and so the whole point too easily misunderstood. John wrote:

“Or it could be that Apple has decided never to open WebKit to developers on Apple TV. Either way, it won’t affect Apple TV’s success, and everything will be OK.”

Apple TV’s success doesn’t change my argument. My Apple TV dev kit arrives on Friday, I’m going to build an app for it, and I can’t wait to watch Apple’s latest platform take off. When I wrote that the Apple TV “needs” the web I didn’t mean that it would be crippled and unsuccessful without it. I simply meant that the web should be there in some form, even if limited.

(It doesn’t even have to be Safari. There just needs to be enough web technologies to make some part of the open web possible. Again, that means web services, HTML, and links.)

Yesterday, John Gruber also wrote about web apps and native apps, and what each should focus on:

“Native apps can’t out-web the web, and web apps should embrace that.”

That’s good advice. There are plenty of important tasks for the web community that should be top priorities, such as encouraging a return to independent publishing and trying to fix the lack of redundancy. The web will always be playing catch-up with native apps for user experience, but the web will always be ahead as a distributed, open publishing platform. And that is such an important feature, it should be available on as many devices as possible.