Hacker News Reader

Google is really losing its way. AMP always looked like an attempt to appify the Web on Google's platform, and that URL scheme makes it even more worrying.

When your AMP pages appear in Google's Search results, users aren't visiting your site. They are visiting a restricted shell of your site on google.com with functionality that benefits Google, not the website owner (examples: a back button to google.com, left-right swipes that take users off of your site, restricted monetization options, pages that don't display for 8 seconds if you block potentially-dangerous 3rd party scripts, etc.).

Is there no one left at Google who still believes in an independent, decentralized Web? The open WWW has always been under attack, but to see it fundamentally wrecked by Google is depressing.

These points are critical.

The fact that you have to surrender full control of your own site and content to be featured by Google is criminal.

They don't own the web.

They could improve their indexing to handle more of the web, but instead they're trying to shift the cost of indexing to people who publish on the web. The small guy. Not the near trillion dollar company.

You find it weird that you have to kiss Google's ass in order to get special treatment from them?

Not weird, just shitty

The most prominent results should be he most relevant results, not the ones who kiss the most ass

I know the formula is mostly about relevance, but ass kissing shouldn't even be in there

> Google is really losing its way.

You sure about that?

To me, it looks more like Google has just become emboldened enough to start openly abusing their position to gain control of the web. This has always been Google's way, it's just that they dressed it up in prettier PR speak previously or tried to do it more quietly to keep the "we're not evil" wool pulled over the eyes of most of the population.

I agree generally with what you're saying, but I think they once at least tried to not be "evil". I don't think they are trying very hard any more.

They were willing to give RSS/Atom a crippling blow at the launch of Google Plus.

I remember Usenet going downhill immediately after they started pulling it into Google Groups (an unusable discussion platform).

I suspect that it will be the end of the WWW if Google is successful with AMP and the removal of URLs. Something new will have to be created from the ashes, though it may be difficult if Google controls the hardware, software (browser, OS), and potentially the network.

Google is motivated by getting people to click on ads. That's why Google Chrome's URL auto-completion is so bad -- you're supposed to go to Google Search on the way to your destination and click on well-concealed ads. I'm guessing that somewhere in the plan to hide URLs is a scheme to get people to depend more on search (to click on ads) and less on going directly to the destination. With AMP, the search and the destination are both google.com.

> I remember Usenet going downhill immediately after they started pulling it into Google Groups (an unusable discussion platform).

Don't get me started on Google Groups... The UI is so bad it looks like an hackathon-level quality...

> The UI is so bad it looks like an hackathon-level quality.

It's called Material Design.

I'm not against Material Design personally, you have plenty of websites & apps with good Material Design, Google Groups just has shitty UI.

Could be time for grassroots UI/UX to have cool logos and brands, like security vulnerabilities.

Based on those links, it doesn't look like Google is trying to get rid of URLs. It looks more like they're trying to make sure that URLs can still be used to locate content and ensure it hasn't been tampered with, but removing the need for browsers to make a round-trip to the origin server to retrieve it.

If anything this has the potential to help further decentralise the web, by allowing users to retrieve content via whatever protocol works best for them, without losing the benefits of URLs.

At work I develop mapping apps for projects which are shared using a URL in an email. This can be confusing for people as they have no intuitive way of knowing if data is live or not. It is very useful to have URLs that record a moment in time for record purposes. But people also sometimes expect the data to update automatically to the current version. You also, have issues with authentification and user accounts being needed.

Normally when you share a document on email it is very intuitive. People realise that it is not live data, and that they can share it with anyone over email. It would be good to have a web package format that could be sent by email and treated like a normal document. But also benefit from server based resources when neccessary.

> It would be good to have a web package format that could be sent by email and treated like a normal document

I'm setting up a static server to deal with hosting analyses purely because there's no current good solution to this and my approach otherwise is "tell people to download then open this html file" which is merely bad for techies but atrocious for non-techies.

I would absolutely love a format I can deliver some HTML + bits of data in that's treated just like any normal document that can be sent around.

I want to try something like this with a plain HTML file. It would need a way to include data that is not processed by the browser at load time, but selectively processed in JavaScript when needed.

This "hide the url from the user" is exactly what America Online did back in the day to try and keep everyone from leaving their garden. I remember being offended by it and trying to explain to people what that meant and that they should cancel their AOL accounts over it, with little success.

How did Google get this bad? Why are we letting the internet devolve back to the early 90s?

Google is panicking over user behavior changes. Fewer people using the traditional search, more people using voice search, more people using a mobile web browser which they can't control. Someone is plotting the use of paid search ads in 2028 and it doesn't look pretty.

If Google doesn't aggressively pursue things which jack up their revenue, they switch from a growth company to a non-growth company sooner rather than later. Instead of a P/E of 52, maybe thats a P/E of 10 or 20. If that happens too soon, then their ability to attract and pay talent diminishes significantly. Thus, "don't be evil" is replaced with the ends justifying the means.

The only big reprieve we have here is that if Google comes up with some wacky new idea which breaks w3c standards and Apple doesn't go along with it, it probably doesn't get adapted.

To be honest I think that's jumping to conclusions based on an article that's absolutely leaping to conclusions. It's highly unlikely Google's web crawling team conspired with AMP to force the web to convert to AMP

Are you replying to the correct comment? Was that claim made in this thread?

This thread is in response to a section of the article where this claim was made

That claim was the basis of the entire article. Did you read it?

Yep. For those not familiar with the heyday of AOL: "Type in AOL keyword XYZ to buy an XYZ"...where AOL sold the keyword to the highest bidder. They weren't selling the top spot in a list, it was exclusive placement.

"All the real magic happened really behind the scenes, utterly invisible to you"

applause from the crowd

So this is how liberty dies... with thunderous applause

What happened to the 'View Source' ethos?!

> What happens if your “web identity” certificate expires or is revoked/blacklisted?

There are no "Web Identity" certificates. Just standard X.509 certs, like there have always been. As for what happens to Web Packages when they expire, that's also exactly what you'd expect: https://wicg.github.io/webpackage/loading.html#trusting-cert...

Stop using chrome, now, do it, it's easy, there are alternatives.

Web packages / wpub look fantastic. Are these things purely at a proposal level or are there demo implementations anywhere? I didn't spot anything obvious but then this is the first I've heard of them.

Thanks for the links--I checked out the AMP event. This quote and after got me thinking:

> From a user perspective, they want to see...where is this content coming from?

It's interesting because it sounds like he's saying: If the bits are coming from Service Y and not literally "Joe's Closet," we want to show people that this Service Y stuff (like Google's AMP URL for Joe's content) is essentially still Joe's Closet.

I'd be more worried for domain registrars in this case than anything. URLs could indeed stop being so meaningful as a faux-ID, because now Joe has much more flexibility and less concern for this or that domain. URL segments to the right of the domain name will be even more of a back-office concern than they are now.

And the big infrastructure cats like Google, who "own" the methods by which published things are effectively spread to audiences, get to now "own" the content, but only if you look at "own" through the old URL-based lens. That's interesting because after this change, what does "my own digital real estate" mean? It effectively means that I own an identity or group of identities, and now vendors get to compete for rights to carry or express my identity. URLs were similar but this can remove a deceptive that's-a-scam-site-even-though-my-name-is-in-the-URL part and allow in a Google-owned-domain part.

And hopefully there really will be more than one entity that can compete. Right now the scary part is that for example if I'm using AMP to publish news, Google gets to express my "everything ADHD news" by default, because they have so much leverage there. (I did catch myself using !gn a lot this week in DDG...mmmph)

Anyway: Remember when you HAD to have a .com to be taken seriously? Wow, "those were the days" comments incoming in the next 5-10 years if this lands well...

Still, I don't know if it's "good" or "bad" so much as another layer through which to work with connectivity.

> What happens if your “web identity” certificate expires or is revoked/blacklisted? Does your web identity certificate require a separate yearly fee from the domain name?

Or maybe more practically: If you have like 100 different hobbies, is a single identity certificate which becomes associated with one of those hobbies going to jack up your chances of getting "found" in some way for the other hobbies? URLs handle this really well.

Also, think about some pretty-good-case benefits: Your domain name or URL problems will no longer get in the way of your "stamp of quality" identity, assuming your identity is looked at well _by some content consumer or group of content consumers_. You have always been a brand just by having a name and an objectively-sense-able set of properties. Now the subjective question is deepened--the quality of what you produce and represent.

RealNames for everyone, Google+ lives!

If pseudonyms are permitted, then how will collisions be handled? Aren’t we right back to DNS namespace collisions and years of case law on domain name trademark disputes?

Organizations that paid millions for domain name real estate will object to their assets becoming worthless overnight. Can they sue Google?

Yep, anon and semi-anon are certainly needed. My guess for a pseudonym mechanism would be borrowing someone else's ID as an umbrella. Going a bit abstract, that's what the Guardian is providing its staff, in that you just see "The Guardian", yet you will probably still be able to find favorite authors there even if they don't use their own special ID certificates.

I'm guessing this won't really (or doesn't _have to_) impede anonymity so much. Some kind soul will set up the "GeoCities / NeoCities of Online Identity" and from there it's just a question of whether you care how you are perceived while using that particular umbrella. Domain names would be the big losers.

But this is just my brand-new mental model for what's going on...

I think ultimately it will have to be social proof that proves which identity a website has. This sounds weird at first but think about it, any type of other identification is going to be text or image based - which is easily spoofed. Certificates can be sort of spoofed too, I mean who looks exactly at that certificate everytime they visit? The certificate itself has a long string in it so the problem really hasn't improved much. For letsencrypt websites the common name actually seems to be some kind of autogenerated string so good luck explaining that to a user (not that he should have to ever check this type of stuff) Google has a social network, they should just integrate URIs aka. "apps" with their social identity. This is also a win&win for them.

Let me explain, if you enter facébook.com then on the url bar it will not show that your friends are on this site as well. The exact gui is irrelevant, the point is that it will be easy to tell whether you are on the right website that you are used to navigating to based on the indicator of amounts of friends and who is using and recommending that site. This will also cultivate a certain awareness that if you are navigating outside your usual url space then you will probably be more careful because it doesn't show that any of your friends are on there yet. Banks might offer a service where their bot "befriends" you on socialnetwork XYZ to get that nice little indicator that you are on the right banking website without anybody else actually having to use it. This system might work for any website actually. If there is a concern that this disincentivizes people from using new websites then I think it will be outweighed by several factors like your desire to establish social proof - meaning that you might be more likely to share the new website you found so you can be safer when visiting it next time because you only have to look at the url indicator and see that friend X that you recommended it to has also added this website to their profile.

> it doesn't show that any of your friends are on there yet.

That wouldn't work in practice. People shouldn't have to share the list of websites they visit with third party entities, and some of us don't use social media. Some people's social media "friends" lists are full of people who they don't know. The future of marketing and phishing would be for bots to befriend unsuspecting people in order to build trust for their sites.