The snippet isn't loaded from their servers when you request it either. Do you think that's a bad idea as well?

I don’t follow. When I search I see google.com in the address bar, so logically that is where the snippets come from.

If I click one of the links, then yes, I absolutely expect the publisher to know I did.

yes.

Can you explain? Every time you search, do you want all of the publishers who appear on the results page to know that you were searching? Or is it that you don't want snippets to appear in the search results and just want a list of links without any evidence for why they might be good matches for your query?

This is a good example of the contradictions within AMP that limit it's usability as a technology. Unless AMP was really intended as a privacy tool this whole time and I just didn't know it, publishers probably "should" know the same amount they would get to know if you visited one of their actual pages, if you are visiting an AMP for the purpose of consuming their content rather than merely searching for it. But if that disrupts the smooth functioning of AMP, well, now we're just exhibiting an architectural shortcoming that's baked into AMPs philosophy of how to serve content.

>Or is it that you don't want snippets to appear in the search results and just want a list of links without any evidence for why they might be good matches for your query?

Is that how you feel about regular search results?

Your comment does not make any sense to me. Are we talking about the same thing? A snippet is a summary of the page served by the search engine. The publisher currently does not have any idea that a particular user has seen a snippet, so publishers do "know the same amount" as they did before with AMP, which is nothing.

> Is that how you feel about regular search results?

Regular search results have snippets.

I'm sorry that you were having difficulty interpreting my comment. It may help if you go back and note that I was making a distinction between using AMP to consume content vs search for it. Publishers normally know when their content is being consumed. Let me know if that makes sense to you.

>Regular search results have snippets.

Right, and I was asking about the search results, not the snippets that accompany them. That is to say, the part with the blue title, green link, and the few lines in black displayed from the page that are displayed ten at a time, not the snippets that accompany them at the top of the page. Unless you were just using 'snippets' as a general term to mean the same thing that I mean by search results, in which case you were just repeating the content of my own question back to me.

> Publishers normally know when their content is being consumed.

They still know when their content is being consumed. They just don't know when their content is being searched for until the user clicks their link, exactly like a snippet. Does that make sense now? My point was that search engines already show cached portions of the page. Read the parent comment of my first "snippet" comment to understand why I was making that point.

If a site is cached locally, it isn’t loaded from ‘their’ servers when you request it. Or maybe it’ll come from a CDN rather than origin.

What’s shown in the URL bar has long been divorced from the HTTP request(s) that are made.

> It blows my mind that anyone thinks that’s an acceptable idea

It blows your mind that different people have difference opinions and values than you?

What is stored in your cache was loaded from their servers when you originally visited. Also, you can clear your cache at any time you choose to re-fetch the original content from the provider.

Also, a provider uses a CDN at their discretion. Giving them the ability to invalidate or update cached records at times of their choosing. Or remove the CDN entirely if they choose to.

This is Google using their weight to be anti-competitive and fall further down the anti-trust rabbit hole.

A provider uses a CDN at their discretion, but it's totally legitimate to have a client-site cache as well. If I've got 20 users going through a squid proxy to get to the internet, that's something the provider has to live with. Not to diminish your core point, which I think is correct, but there are limits on what providers should be able to expect.

I sort of agree with you, but robots.txt is a thing.

I know it seems impossible but nobody has to use Google.

Are there any copyright implications of amp? They're essentially republishing your property.

Does the original publisher get their ad revenue?

> It blows your mind that different people have difference opinions and values than you?

Given context and full knowledge of the issues involved? Yes, absolutely. For this particular issue anyway.

> Or maybe it’ll come from a CDN rather than origin.

That’s not a problem, since the URL and where it points to is still decided by the content owners.