Does AMP Counter an Existential Threat to Google? | Hacker News Reader

Does AMP Counter an Existential Threat to Google?

highscalability.com

28 pointsmoehm12 comments

Yes it does, but it's one of many counters to it. Google is also crawling deep-linked native mobile content through App Indexing [1]. They own one of the dominant mobile advertising platforms [2]. They also own the dominant mobile operating system [3], and make money directly off selling apps on it [4].

It's a very Larry-Page style strategy: when faced with an existential threat, solve it on every possible level. He did the same thing when faced with the possibility that Microsoft would build search into Internet Explorer: build Google Toolbar, and Google Chrome, and get OEM distribution deals for both, and Chrome Frame so you could use IE inside of Chrome, and Android just in case there was a chance to make the desktop market irrelevant, and Chromebooks as another way to make the desktop market irrelevant. Interestingly, Sundar Pichai was involved in basically all of these, which may be why he's now CEO of Google.

[1] https://developers.google.com/app-indexing/?hl=en

[2] https://www.google.com/admob/

[3] https://www.android.com/

[4] https://play.google.com/store/apps?hl=en

Does anyone know why people write sentences like this?

> Pinterest found AMP pages load four times faster and use eight times less data than traditional mobile-optimized pages.

Does "Eight times less data" seem insane to anyone else? Shouldn't it be "one eighth the data" or "twelve and a half percent of the data"

I feel like I'm taking crazy pills, because I always see things written like the article and it seems to make ~zero~ sense.

Nope, not just you. I suspect it's because most readers are afraid of fractions.

It's because of parallelism. Not the computing kind, but the writing kind. Sentence constructions like "AMP pages load four times faster and use one-eighth the data" or "AMP pages have one fourth the latency and eight times better data compression" read awkwardly; try reading them aloud, vs. the original sentence. You're taught in writing class to use equivalent constructions for all clauses within a sentence. It's much like the stylistic rule in programming languages to not mix camelCase and snake_case in the same file.

I get what you're saying, but FWIW I honestly have no problem with the non-parallel sentence constructions in your example. It's a technical article. I want a high amount of information in a concise and clear package.

> "Does AMP work? Pinterest found AMP pages load four times faster and use eight times less data than traditional mobile-optimized pages."

A better question is: how much faster is it than ad-blocked webpages?

I think it's more about ad-blocking than "making the web faster". Also, it looks like it framejacks web pages so that you just visit sites as a side journey from Google.com, with Google's toolbar on top:

http://searchenginewatch.com/sew/opinion/2430844/google-s-am...

"From the search engine results page, users can click through to a list of AMP-compliant partner sites. Doing so will load that content almost instantaneously, as Google will also pre-render content above the fold for AMP listings. Once a user clicks through, there is a persistent blue bar at the top with a call to action to return to the Google SERP."

Do publishers really want a "go back to Google.com" toolbar on the top of their webpages? I'd rather encourage users to stay on my own sites.

> how much faster is it than ad-blocked webpages?

Poke at it out yourself to get a feel for the answer. Find any recent article on any site of one of these publishers that have enabled AMP pages. You can find a list on https://www.ampproject.org/ or any AMP-enabled site you find. Once you've found an article/web page, view source. Look for a line in the html that reads:

If you don't find that tag, the article is not AMP enabled, so find another article that is.

Now, grab the URL at that href attribute and compare it to the original non-AMP URL you started at. Turn on/off ad-blockers, use a cold cache / hot cache, have fun. Don't forget that this comparison is not including the additional speedup gained from using the AMP Project (Google's) CDN.

I wouldn't consider that a benefit, since I don't want every site I visit to ping Google's servers.

I haven't researched it in depth, but it seems like an attempt to appify the Web even more and bring it under Google's control. A "webpage" just becomes a brief side trip from the Google homepage, with a giant back button to send your traffic back to Google rather than deeper into your site.

The fox (any single, large company with an interest in replacing the WWW with itself) should not be guarding the henhouse (the WWW). I mean that in reference to Chrome, appification, and things like AMP. It isn't good for publishers or for the future of the open Web.

It seems absurd for sites to put code on their pages that adds a giant "go back to Google.com" banner on top of their pages, if that's what it does.

The <link> tag is similar in browsers to a meta tag, it doesn't cause browsers to take any behavior, so this doesn't cause a ping to any server. The AMP HTML version of a page does require a javascript src to make these transformations, but that's cached, so you aren't going to send a ping to request it on every page load.

The AMP version of a page doesn't add any banner or other visual style, the author has full control there. Here's an example, just pulled at random:

Original: http://www.nytimes.com/2015/12/17/theater/broadway-is-gettin...

HTML Mobile: http://mobile.nytimes.com/2015/12/17/theater/broadway-is-get...

AMP: http://mobile.nytimes.com/2015/12/17/theater/broadway-is-get...

I thought a link tag sends a request to get the contents of the href attribute, and then the server sends back the file or a 304 header, if it's cached. That seems to indicate that the server is getting a request from the browser. Also, isn't Google caching AMP pages on their own domains? I've read that Google doesn't send people to your site, but just to a copy on their own site.

When I try the demo for the NY Times, I see a banner there that takes me back to the search results. Did NY Times decide that they want to use that above-the-fold real estate to send people back to Google.com?

What specific code enables or disables this giant back button on an AMP webpage: http://imgur.com/GpBqgnx

AMP is a somewhat-forced restriction on what people can publish on their sites. Providing guidelines on how to make mobile pages faster would be a more reasonable solution. One could argue that no one is "forced" to use AMP, but if AMP pages appear at the top of the SERPs then that is a form of coercion, since fewer people will scroll beneath the fold.

I've been told that google has intraday downloads of their doubleclick logs "on their roadmap". For me in adtech, a lot of the pain I feel in having to add more tagging would be reduced if google had this stuff streaming in realtime. Their lack of this functionality has fed the thirst for other tags (adobe, bluekai, etc.), since we can't build really fast behavioural reactions on their data. They've told me the solution is to use Google Analytics Premium, and upload our behavioural reaction rules into there. Just like everyone else, that is a black box, and I then have to adhere to the functionality they have built. Businesses will always ask for a custom tweak that any black box can't deliver, and that also leads to more tagging. ;)

Truly open up your platforms!!! Currently it feels like a mexican stand-off, where the first company to open up loses control of their data asset, while all others win.

Betteridge's law

Reply on news.ycombinator.com