Embrace AMP or AMP Wins

I’ve written about AMP (Amplified Mobile Pages) on here a few times.  To recap, it’s a standard of sort and coupled set of code.

If you publish a page through AMP, you are limited to a very narrow set of HTML traits and behaviors, and a limited, curated set of JavaScript to provide basic functionality, ad networks, video hosting, metrics, with scripts hosted by the Google owned and operated cdn.ampproject.org. You also implicitly allow intermediaries to cache your content.

If you search Google using a mobile device, links with a little ⚡ icon are AMP links that will be loaded from the Google cache, and by rule (which is verified and enforced) live within the limitations of AMP. You can’t claim AMP conformance and then resort to traditional misbehavior.

The news carousel is populated via AMP links.

Many publishers have gotten on the AMP bandwagon. Even niche blogs have exposed AMP content via a simple plug-in.

AMP is winning.

But it has significant deficiencies, for which it has earned a large number of detractors. There are technical, privacy and web centralization issues that remain critical faults in the initiative.

Anti-AMP advocacy has reached a fevered pitch. And that negative advocacy is accomplishing exactly nothing. It is founded in a denial that is providing a clear road for AMP to achieve world domination.

Because in the end it is a better user experience. Being on a mobile device and seeing the icon ⚡ is an immediate assurance that not only will the page load almost instantly, it won’t have delay-load modal overlays (Subscribe! Like us on Facebook!), it won’t throw you into redirect hell, it won’t have device-choking scripts doing spurious things.

Publishers might be embracing a Pyrrhic victory that undoes them in the end, but right now AMP makes the web a more trustworthy, accessible things for users. It is a better experience, and helps avoid the web becoming a tragedy of the commons, where short-sighted publishers desperate for a short-term metric create such a miserable experience that users stay within gated communities like Facebook or Apple News.

We could do better, but right now everyone has exactly the wrong approach in confronting AMP.

“We don’t need AMP: We have the powerful open web, and publishers can make their pages as quick loading and user-friendly as AMP…”

This is a losing, boorish argument that recurs in every anti-AMP piece. It is akin to saying that the EPA isn’t necessary because industry just needs to be clean instead. But they won’t. AMP isn’t an assurance for the publisher, it’s an assurance to the user.

AMP isn’t a weak, feel-good certification. To publish via AMP you allow caching because that cache host validates and forcefully guarantees to users that your content lives within the confines of AMP. You can’t bait and switch. You can’t agree to the standard and then do just this one small thing. That is the power of AMP. Simply saying “can’t we all just do it voluntarily” misses the point that there are many bad actors who want to ruin the web for all of us.

But the argument that as a subset it therefore isn’t needed — missing the point entirely — is self-defeating because that argument has short circuited any ability to talk about the need that AMP addresses, and how to make a more palatable, truly open and truly beneficial solution.

We need an HMTL Lite. Or HTMLite to be more slogan-y.

The web is remarkably powerful. Too powerful.

We have been hoisted by our own petard, as the saying goes.

With each powerful new innovation in web technologies we enable those bad actors among us who degrade the experience for millions. For classic textual content of the sort that we all consume in volumes, it is destructive to its own long term health. Many parties (including large players like Apple and Facebook) have introduced alternatives that circumvent the web almost entirely.

A subset of HTML and scripting. A request header passed by the browser that demands HTMLite content, with the browser and caching agents enforcing those limits on publishers (rejecting the content wholesale if it breaks the rules).

We need to embrace the theory of AMP while rejecting the centralized control and monitor that it entails.

This isn’t simply NoScript or other hacked solutions, but needs to be a holistic reconsiderations of the basics of what we’re trying to achieve. Our web stack has become enormously powerful, from GL to SVG to audio and video and conferencing and locations and notifications, and that is just gross overkill for what we primarily leverage it for. We need to fork this vision before it becomes a graffiti-coated ghetto that only the brave tread, the userbase corralled off into glittery alternatives.