I don't even work on the web professionally, and even I can figure this stuff out. It's really not that hard.
I don't even work on the web professionally, and even I can figure this stuff out. It's really not that hard.
This is not even wrong. I'm sorry; I never post just to criticize someone's comment. But yours was the top comment on the page and was, in addition to being wrong, condescending to an author who is clearly very knowledgeable (I say that as a web publisher who has monetized with ads for more than a decade).
From the article:
> [The recent DOJ lawsuit against Google] claims that Google falsely told publishers that adopting AMP would enhance load times, even though the company’s employees knew that it only improved the “median of performance” and actually loaded slower than some speed optimization techniques publishers had been using. It alleges that AMP pages brought 40% less revenue to publishers. The complaint states that AMP’s speed benefits “were also at least partly a result of Google’s throttling. Google throttles the load time of non-AMP ads by giving them artificial one-second delays in order to give Google AMP a ‘nice comparative boost.‘”
This is a very big deal and a huge trust buster, and a giant dick move by Google against publishers. It's literally an evil act by a company that said Don't be Evil. The article is spot on that AMP has irreparably damaged publishers' trust.
I provided a link that explains exactly how it works. The reason it's not even wrong is that it's absolutely correct. I say that as somebody who has actually read some AMP docs and understands the value proposal of same origin serving for prerendering.
> This is a very big deal and a huge trust buster
It would be if true, but blindly trusting the prosecutor is a huge patsy move on the author's part. They have no clue what the mechanism is. I provided it.
It has nothing to do with "pre rendering" its built right into the CSS.
The term "rendering" is used
(a) to describe the actual translation of textual-instructive HTML/CSS/Javascript into a graphical representation within a browser window and
(b) to describe stuff being loaded into a page via Javascript, most prominently ads, and more oriented towards load time ("How long until the page has finished loading" = "rendering time").
With CSS, you can define pixel-perfect areas for stuff to be rendered into, which makes the browser not having to decide on the size of the area after it has loaded further material from third-party sources. Doing so reduces "screen flickering", which is generally annoying (and the major reason why I block ads with a vengeance). This can be described as "pre-rendering" when you understand rendering as (a).
This animation is turned off instantly (as far as I can say) by Javascript. However when you block Javascript entirely or block ampproject.org you have to endure the whole 8 seconds. I inject "body {animation: none !important}" into all pages to remedy this.
Or do you just mean that all possible ad creatives are server-side-rendered, so to speak, and kept ready in a google cache?
Edit: oh probably pre rendered by the user browser in the background. In that case any ad could still trigger it‘s impression with a viewport listener, if that‘s what google needed them to do.
With AMP, Google (and Microsoft and the other companies running AMP caches) can verify that the ad doesn't trigger an impression until viewed, which they need to do because they serve it themselves.
https://developers.google.com/authorized-buyers/rtb/amp-ads
> Does it do a speculative auction when you load the search results page?
Yes.
> It so, that is an incredibly unfair advantage for google.
Why would it be? Anybody can participate in the auction.
The one second delay is so that ads that aren’t sandboxed and loaded client side, won’t impact the reading experience. For ads that are loaded server side, they can ensure the quality is high enough that they can be displayed while content is rendering.
Where does it say that? The delay is from the ad spot not even being available to bid on until the user opens the article, some timeout to receive all the bids, and then some time to fetch the ad from the winner's ad server instead of returning the ad content immediately after the winner has been determined.
> For ads that are loaded server side, they can ensure the quality is high enough
For ads that are loaded server side, they can just validate against the spec to ensure that it is safe for the AMP cache to serve inside the article.
Or have you even read the article?
The author made the elementary mistake of blindly trusting the prosecutor without putting on his engineer hat to figure out if the prosecutor's allegations made any sense whatsoever.
>> Or have you even read the article?
> Yes, yes, and yes.
Not so much on that last "yes":
> The author ... without putting on his engineer hat
>>> Sarah Gooding
Putting on her engineer hat.