Hacker News Reader

Comment by cphoover | Hacker News Reader

Thank God AMP is dying...

If you are looking to develop a specification for preloading go through the browser standards process, instead of unloading this proprietary crap on the rest of us.

View on HN

Twitter has become incredibly hostile to users not signed into an account (e.g. many search visitors), so AMP or not isn't going to make a big difference.

Agreed. They've made it almost unusable if you hit it from an embedded browser page.

amp wasn’t a problem. amp prioritized results in search was a problem

and i just saw here they slowed down ads on non-amp pages, that’s not good for business

IMO AMP was still a problem. It’s a version of HTML de facto controlled by Google. There’s window dressing of a foundation or whatever but Google calls the shots on what it does and does not do. To me that made AMP a five alarm fire right from the start.

And it had to be proxied through Google's servers. To hell with it.

I'm curious if anybody else is noticing how G is tracing the same paths AOL once did.

Yes and Google too shall pass.

Would be nice to have a post-mortem on 'how to recognize the outline of an emerging juggernaut'.

Cloudflare, take note.

This is false.

Why, because there's also a Bing AMP proxy? Come on. AMP pages in Google search results are proxied through Google 100% of the time, by design.

In fairness AMP does not have a requirement to use a CDN. It allows the use of a CDN which was another reason why Google pushed it (to do all kind of same-domain tricks), but to keep on topic, I actually don't think Twitter used that Google cache.

If Twitter was serving valid AMP, they didn't have a choice on using the Google cache or not:

"As a publisher, you don't choose an AMP Cache, it's actually the platform that links to your content that chooses the AMP Cache (if any) to use."

"Caching is a core part of the AMP ecosystem. Publishing a valid AMP document automatically opts it into cache delivery."

https://amp.dev/documentation/guides-and-tutorials/learn/amp...

Eh? By your own quote:

> it's actually the platform that links to your content that chooses the AMP Cache (if any) to use

Twitter (the platform that links to AMP content) could choose which AMP cache if any to use.

Oh, I guess maybe the confusion here is that your argument is that as a page publisher you cannot opt out of the Google CDN? I read your original statement as "it is not possible to view AMP content without using the Google CDN"

Right, I'm mainly just talking about Google search results. I probably should have made that more clear up front.

The original article is talking about traffic from Twitter to publisher sites, not traffic from Google to Twitter. Twitter never used AMP for pages on their own site.

In this case Twitter is the platform, not the publisher, and would absolutely have been able to not use the Google cache.

> Now, when using one of Twitter's mobile clients, users will be sent to the amphtml URL in their browser, instead of the link that was shared in the Tweet. Users will load this link directly, not via a page cache. [0]

[0] https://developer.twitter.com/en/docs/twitter-for-websites/a...

Twitter directly linked to AMP pages on the publisher's site. They didn't link to a cache at all.

We're on a discussion page for an article about Twitter's use of AMP in their own app, which was never proxied through anything, much less proxied through Google.

I wonder why someone who works at google bothers to come here to try to defend the worst practices of the company.

I mean, it's never going to go well.

Is it some sort of deranged catharsis in trying to convince yourself?

Google calls the shots on what the web does and does not do via the majority browser - Chrome. I wouldn't consider AMP any more dangerous than Chrome in this argument. There were other problems with AMP.

I'm certainly not a fan of a Chrome monoculture either but AMP crossed all browsers so I see it as more dangerous.

amp was the modern attempt to replicate the turd that was wireless application protocol and I'm glad it's dying.

I'm also glad the community mood shifted about it, no longer than a year ago negative comments about amp where drowned in downvotes

luckily vote count doesn't make ppl right or wrong.

> no longer than a year ago negative comments about amp where drowned in downvotes

This is the strangest case of revisionist history I've ever seen.

No, AMP was a problem. It was a proprietary, ham-fisted subset of the open web entirely forced by Google’s insane search engine dominance. It actively discouraged the creativity, freedom, and self-determination of the web that makes it so special. The AMP boost in search ranking was just the cherry on top.

And that’s not even discussing the fact that Google penalized non-Google ads on AMP.

https://mobile.twitter.com/simonw/status/1451947624513167363

This is exactly right. Imagine a hypothetical where Microsoft created AMP - no one would use it! The technology itself sucked. The only reason some publishers adopted AMP at all was to get/maintain favourable placement on Google, the only search engine that matters.

I'm just hoping that the death of AMP on the web will also kill Google's ambitions for rolling it into Gmail[1].

[1] https://developers.google.com/gmail/ampemail/

The idea of a proven-safe subset of HTML with no custom JS that's suitable for adding interactivity to emails is actually really interesting to me, and I hope it continues. It's been really great to be able to reply to Google Docs comments from within my email client, and the same things that make AMP good for rehosting make it a good foundation for interactive email.

AMP pages being hosted by Google is a pretty big problem as well, though. In my opinion a step in the wrong direction of the centralization of the internet.

Why should it be a browser standard? It’s a web framework. Browsers already support it.

Google did promote some improvements to web standards but they were never tied to AMP, even though they were things AMP would use.

Because without going through the standards process it is a unilateral de-facto standard that has been pushed on the web without community involvement.

Should React be a web standard? How about Typescript?

AMP being built into browsers wouldn't be an improvement.

These technologies are built on top of web standards. It's a different level of the stack.

AMP was built using existing web standards, which means publishers don't have to wait half a decade before Safari lets them use it.

Reply on news.ycombinator.com