Hacker News new | comments | show | ask | jobs | submit login
iOS 11 Safari will automatically strip AMP links from shared URLs (twitter.com)
534 points by OberstKrueger on Aug 23, 2017 | hide | past | web | favorite | 420 comments



TL of AMP here. Just wanted to clarify that we specifically requested Apple (and other browser vendors) to do this. AMP's policy states that platforms should share the canonical URL of an article whenever technically possible. This browser change makes it technically possible in Safari. We cannot wait for other vendors to implement.

It appears Safari implemented a special case. We'd prefer a more generic solution where browsers would share the canonical link by default, but this works for us.


> We'd prefer a more generic solution where browsers would share the canonical link by default, but this works for us.

We had that solution it was called 'the web'.

I'd prefer a more generic solution too: get rid of AMP.

It really isn't about what 'works for you' it is about what is good for Google versus what is good for the web.


Hear, hear.

If you want to incentivise not delivering multi-megabyte "experiences" to people, I could see a case for feeding page weight into the ranking algorithm. Then people would be less inclined to have tag managers loading tag managers loading a zillion random js trackers.

I have personally saved at least one decent laptop from landfill by installing uBlock Origin on it, and making the web usable again.


Connection count multiplied by number of bytes transferred.

That would get rid of the cruft on the web in a heartbeat.

Every SEO consultant would suddenly have an actual job, playing code golf on CSS and HTML to deliver the content in the minimum number of bytes possible.


Google already wrote AMP which parses page content and presents the user with Google's ideal version of the page. Just include the delta in the ranking. The more a page has to be "cleaned" by AMP the lower it ranks.


That’s... not really what AMP does.

I mean, kinda, but not really. You have to cater to it specifically


Yeah good point. In general something like that could still work though. Whatever characteristics of a page Google is trying to "fix" with AMP could just be added to the ranking algorithm.


Absolutely


Exactly, make it an internal delta signal thing rather than pushing it externally on the entire web. Yes they can because of their position and size, but should they? No they shouldn't.


Clever!


This penalizes data-heavy websites, though, no? For example, I can see how a photojournalism piece with lots of photos would be ranked to hell, whereas a content-free text piece would be ranked very highly.


But on the same subject all pages would be data heavy. Only where a data heavy page would compete with a similar page that was lightweight would the one be ranked above the other.

But your average news page would weigh in roughly the same no matter what the source was. Assuming you'd strip out all the cruft.

For reference, I optimized my website a while ago and ended up with the average page being < 15K.

https://jacquesmattheij.com/the-fastest-blog-in-the-world

Try it, on any connection in normal use today it should load faster than you can blink.

Add a few pictures and it would still work quite comfortably.


> But on the same subject all pages would be data heavy.

Not all of them. For a trivial example, an in-depth article about a topic would be penalized more than a more epidermic one, because it would contain more text, graphs, etc.

> For reference, I optimized my website a while ago and ended up with the average page being < 15K.

I feel you, I absolutely hate that my website uses a bunch of stuff for almost no reason. 80 KB of CSS or 90 KB of font isn't awesome, and I started a project to provide information on how to create lighter websites (https://www.lightentheweb.com/, stalled a bit).

All I'm saying here, though, is that the solution is more nuanced than "heavier = lower". AMP is definitely not a step in the right direction, though. I wonder if Google could give sites that used a specific CSS file and no JS a bump, rather than loading things on their own domain.


> For a trivial example, an in-depth article about a topic would be penalized more than a more epidermic one, because it would contain more text, graphs, etc.

But such trivial examples should be trivial to get right no? It's all about the relative weight of the various factors that determine the ranking and a 'lightweight' article could be recognized as such.


Well, that's my point, it's not trivial to automatically recognize which photos are content and which are style. You want to have many content photos, but having many style photos probably means cruft.

How do you recognize which is which?


Same way almost all browsers offer a Reading view. They still show the article image. Doesn't seem to be that hard of a problem.


How does AMP recognize this right now?


It doesn't, it just forces you to display the pages through Google, stripping everything else, which is much easier to implement than recognizing if a web page's heft is because of legitimate content or ads.


So what difference would it make? After all, this is the result after AMP has done all its cleaning, which would give you a pretty accurate measure for page weight. What remains is the actual content of the page. It's the cruft that should be penalized, not the actual content.


Exactly what my sibling comment said, it's the publisher (presumably a human) who does the cleaning. It's not as easy to get a bot to clean things up and then measure weight, and the various "Readability" views aren't always correct (they're using heuristics).

Sounds like a publisher-assisted "readability view" would solve both problems, though. It would allow the user to view a "lighter" version of the page and search engines could just measure how large the "non-light" version is and penalize accordingly. I'm afraid that would incentivize the publisher to put ads and other crap in the "light" view, though.


AMP doesn't do the cleaning.

You as the publisher/author have to implement an AMP page yourself and then alert Google of its existence.


A photojournalism page would have a low connectivity count. It wouldn't be connecting to various other 3rd party websites


Most useful content on the web is text. Photojournalism is a niche exception and those pages could be linked to from articles.

A lot of the useless pages are 3 stories of scrolling through large images with very little text.

Also, "data" as in tables and small/vector charts isn't all that big.


> Most useful content on the web is text

Tell that to a photographer, designer, artist, illustrator, etc


I would, their requirements are niche. Most people need text, possibly with one or two small photos or illustrations.


Think you may be looking at this only through a developer lens. Bet if you showed anyone under 15 an internet without pictures and video they'd consider it broken.


Weight photos differently to js?


What does the connection count bring anything? All that matter would be the total number of bytes transferred.

I don't think that would be a good metric for relevance though and in most case, the total size isn't an issue. Which is why AMP is good, low file size when it's needed, full size when it's not.


Connections^2 to penalize the worst offenders with 20+ cross site scripts most.


Google already tried this, back in 2010.

https://webmasters.googleblog.com/2010/04/using-site-speed-i...


> I have personally saved at least one decent laptop from landfill by installing uBlock Origin on it, and making the web usable again.

For me, I'd given up on trying to browse the web on my smartphone until Firefox Focus came along. I realize Firefox for Android now allows extensions, but I still used it as a typical browser. I signed into my accounts, I had expectations for a personalized experience, etc. With Focus, it's just a quick visit to a site and then it disappears. I can find the information I want, quickly, yet again.


That's what google would do if they wanted to help the web, and not just faking they do. They used to, now they are just playing pretend for PR reasons.


[flagged]


It won't help you to find what you are looking for but it incentivizes the websites to reduce the weight of their pages and that is ostensibly what AMP tries to achieve.


> I'd prefer a more generic solution too: get rid of AMP

User here. I like AMP. At the end of the day, I want to read an article. Most of the time, Reader view works. But not always. AMP reliably and quickly renders legible web pages. This wouldn't be an issue if 90% of newspaper websites didn't look like they were designed by turnips.


Yes, I get why Google positioned it the exact way they did. It's very clever. But clever doesn't make it right.

If Google really wanted to improve the quality of the web they had excellent tools to achieve that: the weighing of the various elements that go into the ranking algo.

But instead they chose to do an end-run around all standards processes and to attempt to capture mobile content (which is an extremely valuable slice of all web traffic) at the expense of competition and the openness of the web.

And so you are now a very willing pawn in the endgame for the domination of the world wide web, which will eventually result in Google being the sole provider of your content by virtue of the web being consumed more and more through mobile devices.

If you're not concerned I totally understand, but maybe you should be.


> If Google really wanted to improve the quality of the web they had excellent tools to achieve that: the weighing of the various elements that go into the ranking algo

Doesn't Google already do this [1]? There are unintended consequences to increasing speed's weight in search rankings. Principally, it disadvantages smaller players. (Also, the New York Times would consistently rank under Buzzfeed.)

> you are now a very willing pawn in the endgame for the domination of the world wide web

Most people get news from the walled gardens of Facebook and Twitter [2]. The minority of us sharing links to newspaper articles probably also have direct relationships with publishers. For example, I read the New York Times, Washington Post, Economist, Bloomberg and other papers directly on their sites because I subscribe to their newsletters.

> maybe you should be [concerned]

Maybe, but I don't think so. The things I read about in the newspapers AMP gives me faster access to yield better (i.e. more serious and more actionable upon) avenues for my attention. In every context where AMP is a worry journalism's decline seems more important.

[1] https://webmasters.googleblog.com/2010/04/using-site-speed-i...

[2] http://www.niemanlab.org/2015/07/new-pew-data-more-americans...


> Principally, it disadvantages smaller players with smaller IT budgets. The New York Times would consistently consistently under Buzzfeed. A Pyrrhic victory.

Well, assuming equal relevance yes. But the NYT is not exactly a small player and could easily up their game in the page speed department. Besides that you could slowly ramp up the weighing to give parties the time to adjust.

> Most people get news from the walled gardens of Facebook and Twitter [2].

And this is what worries Google. Their long term adversary is Facebook (not Twitter, at least, not nearly so much).

> The minority of us sharing links to newspaper articles probably have direct relationships with publishers.

Not necessarily. I use a variety of aggregators.

> For example, I read the New York Times, Washington Post, Economist, Bloomberg and other papers directly on their sites because I subscribe to their daily newsletters.

Ok.

> The things I read about in the newspapers AMP gives me faster access to hand me better (i.e. more serious and more actionable upon) uses for my attention.

Yes, it's a very seductive proposition, and one that given Google's power in the marketplace they just might get away with. See also: Microsoft bundling their browser and tying it in at every level possible in to the OS and other abuses of monopoly power.

> In every case where AMP is a worry, journalism's general decline is more important.

But that's the same kind of argument that people use when India has news about their space program: Why don't they fix hunger first. It's possible to both worry about journalisms decline and monopoly power abuse on the web, especially when it concerns the distribution of news.


> But the NYT is not exactly a small player and could easily up their game in the page speed department. Besides that you could slowly ramp up the weighing to give parties the time to adjust

When I'm looking for news, I want the highest-quality source first. AMP gives everyone the option of going fast now while kinks are worked out back home.

(I should mention that my default search engine is Duck Duck Go. I usually !news (Google News), though, because I like Google's news results better.)

> it's a very seductive proposition, and one that given Google's power in the marketplace they just might get away with

They might. But a lot of things might happen. AMP makes good journalism more accessible. If Google abuses that privilege down the road, it's not like they're legally unassailable.

> It's possible to both worry about journalisms decline and monopoly power abuse on the web, especially when it concerns the distribution of news

Lecturing newspapers about why they should have had speedier websites went nowhere. Seeing that the same content produced more views through AMP got the message across.

I don't think the analogy to India's space program is accurate. Technological development lifts standards of living. There's a solid argument for doing both because each helps the other. Killing AMP and forcing to choose users between crappy websites and efficient walled gardens is counterproductive.


> When I'm looking for news, I want the highest-quality source first.

Well, no, you want the highest quality source first that loads fast. Otherwise you wouldn't care about AMP in the first place!

> AMP gives everyone the option of going fast now while kinks are worked out back home.

So, do you expect that Google will ever give up their stranglehold on the sites that use AMP once they are in?

In my experience such measures tend to become permanent fixtures with their own hooks embedded in the various fabrics making removing them hard to impossible once they have achieve critical mass. We may have already passed that point, in which case you can safely ignore me.

> They might. But a lot of things might happen. If AMP makes good journalism more accessible I'm all for it.

That's your privilege. I have yet to have a day where I could not read more news than I had time for so I'm not familiar with good journalism not being accessible but I concede that this may be your situation.

> If Google abuses that privilege down the road, it's not like they're legally unassailable.

Who will take that up? You? The EU? The publishers?

My guess is that once entrenched it will be impossible to get rid of, a parasitic element in the pipe between content producer and consumer.

> Lecturing newspapers about why they should have speedier websites went nowhere.

True.

> Seeing that the same content produces more views through AMP, however, seems to have gotten the message across.

Well no, I still don't see news websites getting any faster. But they're already ridiculously fast, consider that 30 years ago you had to wait a full day to read the news and now you can read it in 10 seconds. It's great if that can be cut down to 0.01 second and I'm all for doing that and incentivizing those companies to do this but in the end it is their web property and the market should be allowed to sort this out rather than that a search engine monopolist that is already the gateway to almost all pages read on the web becomes the gatekeeper to our news.

> In the meantime, Google's cash flows get to subsidize this IT expenditure for the news industry while consumers get faster news sooner.

No, Google's cash flow will not subsidize this because they now have the eyeballs. Sooner or later there will be a push for monetization.

> Killing AMP and forcing to choose users between crappy websites and efficient walled gardens, however, is counterproductive.

There are other choices besides.


> Well, no, you want the highest quality source first that loads fast

AMP isn't forcing me to read fast newspaper A over slow newspaper B. It's letting me read slow newspaper B fast. I've never clicked on an AMP link for a source I didn't know.

> In my experience such measures tend to become permanent fixtures...Who will take that up? You? The EU?

You made an analogy to Microsoft bundling browsers earlier.

> I still don't see news websites getting any faster

They are [1].

> [websites are] already ridiculously fast

That's your preference. Mine is for them to be faster.

> Google's cash flow will not subsidize this because they now have the eyeballs. Sooner or later there will be a push for monetization

AMP is open source. If we'd gone the penalise-slow-pages route, every newspaper would have had to "roll their own" AMP in house. Even if Google throws a tax on later, readers got fast pages sooner and publishers got free code.

> There are other choices besides [crappy websites and efficient walled gardens]

In theory. In reality, that's the battleground.

[1] http://www.niemanlab.org/2014/09/how-nytimes-com-cut-load-ti...


> AMP is open source. If we'd gone the penalise-slow-pages route, every newspaper would have had to "roll their own" AMP in house.

Note that you have to use the Google CDN provided javascript, the 'open source' bit is as good as meaningless as long as that restriction is in place.


> have had to "roll their own" AMP in house.

That's called HTML. Nothing special.


>visit nytimes.com >10 seconds to load page

Maybe if you're using the Trump definition of fast...


"NYT could easily up their game" smacks of "I could build that app in a weekend." It's not so easy.


> There are unintended consequences to increasing speed's weight in search rankings. Principally, it disadvantages smaller players.

It's that the case with SEO already? The rules are pretty opaque. If e.g. page weight or connections made was used as a metric, it could be the other way around. Small players don't usually have hugely bloated sites, and don't load hundreds of javascript tracking shit - that's what the big players do.


There's a problem with a lot of HN discussion where people who are very tech savvy tend to have a lot of ideas for how things "should" be. But it's an echo chamber and the average user probably won't care about the nuances in development philosophy. This debate reminds of all the shouting people used to do over Spotify ruining ownership of music. There was a lot of talk about ownership philosophy, DRM and the future of digital rights and "openness" but in the end, it didn't matter because the average user didn't care: convenience won out over whatever perceived philosophical downsides where voiced. Devs can moan about the philosophy behind AMP all day long but the reality is the end experience for the user is an article that loads instantly, looks clean and doesn't bombard with pop-overs. They don't care what tech makes the magic happen, just that it happened. And Google is in a unique position to actually make the web more enjoyable to use and this is the method they chose. Is it right? Maybe...Personally I don't think so, but it does work as advertised. I've heard a lot of people complain about the problems AMP is designed to handle, but as of yet I haven't heard any good alternative solutions. Sorry, but your free-range, open-source, made-with-love ad blocking extension won't be the end all you're hoping...Google is actually in a position to fix the problem and why not let them benefit while at it?


> They don't care what tech makes the magic happen, just that it happened.

True. But I do care what tech makes the magic happen and the consolidation of the web into a very small number of silos is a dangerous development for many reasons and AMP is one plank in that consolidation game.

> Google is actually in a position to fix the problem and why not let them benefit while at it?

Because their stated reasons for AMP do not align with the implementation details, AMP grabs way too much control for a simple improvement in speed.


> I do care what tech makes the magic happen

And to give credit where it's due, I'm glad you do. While we disagree on this point, I have a great deal of respect for your comments and found our discussion enjoyable.


Ditto!


> This debate reminds of all the shouting people used to do over Spotify ruining ownership of music. There was a lot of talk about ownership philosophy, DRM and the future of digital rights and "openness" but in the end, it didn't matter because the average user didn't care: convenience won out over whatever perceived philosophical downsides where voiced.

Your analogy may be reasonably apt, but if so, it shouldn't make anyone who makes their living thinking about systems sanguine.

It does turn out many consumers will trade ownership for convenience. Particularly if convenience comes at a fraction of a cost. Meanwhile, there are side effects: the people actually creating the product (recorded music) get paid a vanishing fraction of what they used to. That changes the economics of actually producing recorded music, shifting the ability to do it to people who can get their money elsewhere. Some people like to try and obfuscate that reality with talk of "new business models" and "innovation," but it all boils down to the assertion that people who make recorded music should have to do another job in order to provide cloud record collections like Spotify to a consumer for the cheapest price possible.

The consumer, as you've pointed out, doesn't want to think about this. Whether they think about it or not, it will shape what kind of music gets produced and by who, so perhaps they should think about it, even if the incentives are more long term and non-obvious.

Buuut even assuming they don't care to do that... it doesn't mean that people who do care to think about how that affects the industry don't have every right to "moan about the philosophy."

Same goes for AMP. Solutions like AMP will absolutely have second order effects. Many users don't care to think about them. Doesn't mean they won't be subject to those effects, that those effects are just about philosophy, and it surely doesn't mean that people who are interested in them have any obligation to stay quiet.

And of all the things to actually have this really be something people waste time on. Working out the economics of production and distribution of music in a time when making copies is just short of free is at least actually a tricky problem. The idea that making mobile pages that load fast is a tricky problem that requires a Google engineering solution is utterly ludicrous.


Once again, though, I think your counter to my analogy of Spotify just reiterates the echo chamber effect of the discussions here that like to debate the philosophies without seeing the bigger picture. Yeah, artists may get paid less now, but Spotify and its ilk pretty much removed piracy from the vocabulary of regular consumers. Back in 2004, music piracy was just something you did...Limewire was an essential app almost. Now, piracy is pretty much relegated to the vocal few that just flat out refuse to pay for things on some skewed principle and that's all thanks to streaming services. What's the alternative? $0.99 songs didn't cut it....should they have been priced down to $0.01? Even then, the amount of consumption would have made the cost unwieldy for many. What people "should" think about is irrelevant. Consumers should also think about using a password manager with 2FA support that stores its database locally. They should think about hosting their own email on a domain they control with SSL support and PGP built in. They should think about switching to using apps like Signal with end to end encryption built in by default. They should think about only driving cars built in the 90s before black boxes were mandatory and OBDII ports with digital interfaces were standard. They should think about a lot of things....but once again, not everybody is going to become a "geek" and jump through hoops just so they can get online, listen to must, check their bank, message friends and drive a car to the store. I completely agree that AMP and the like will have consequences if widespread adoption is seen and I don't like the idea of Apple, Google and Facebook basically controlling what articles get the most eyeballs simply because their platforms are being utilized, but in the end, what the consumer adopts is what developers will be forced to develop for. My problem is that a lot of developer-type solutions exist in the Silicon Valley echo chamber and will never see adoption because they don't take into account the big picture....technology is littered with the corpses of great ideas that consumers just didn't "get." It's true that Google doesn't have to be the one to decide how the architecture of the web works, but I think you're fighting a losing battle....it's much easier for Google to simply say "this is how it's going to be" than for a startup to say "we made this great new product and you should adopt it." Fundamentally, I don't think we disagree, but I'm playing my own devil's advocate here because I'd rather see some discussion on moving forward positively than just people yelling at a wall about how much they think Google's approach sucks which will change nothing...Google will do what they want to do and they're big enough to where if they lose a few million hardcore advocates in the process it's no skin off their nose.


> I think your counter to my analogy of Spotify just reiterates the echo chamber effect of the discussions here that like to debate the philosophies without seeing the bigger picture.

This isn't about "philosophy." This is about systems and consequences. The consequences are real to the user whether they're aware of the chain that produced them or not.

> Yeah, artists may get paid less now, but Spotify and its ilk pretty much removed piracy from the vocabulary of regular consumers. Back in 2004, music piracy was just something you did

2004 is a poor point of comparison if what you want is to get the big picture or even just answer the narrow question of whether the choice really is between cut-rate cloud record collection services like Spotify or free-for-all piracy. 2003 is when the iTunes Music Store launches and 2004/2005 is when you really start to see the rise of digital music retail. Between then and the early 2010s you see those services become widely accepted (even embraced by non-technical people) as an alternative to piracy, and you see revenue from those retail services rise just fine without streaming. In fact, I've seen some reports that suggest that by 2012/2013 the profits from digital retail was on its way to the profits from physical formats.

So the choice wasn't necessarily between fiddly inconvenient piracy and butter smooth streaming experience.

And the battle doesn't even necessarily play out on the field of consumer choice, really. I mean, in a perfect world I might well expect enough consumers to recognize how they vote with the wallet will shape the world, that when we value McDonalds the economy produces McJobs, that when your airfare dollars are ultimately decided by the lowest price you'll get a shitty airline experience.... but yeah, people don't. That's actually why it's more important for people who can see the consequences to discuss them, publicly, loudly, maybe even forcefully enough that product owners who can't see past their A/B testing (and may not even have arrived at the points where they understand the limits of that along with the value) might pay attention.

> My problem is that a lot of developer-type solutions exist in the Silicon Valley echo chamber and will never see adoption because they don't take into account the big picture....

We're concerned precisely because we see the big picture. Understanding that Google has the power to shape the landscape regardless of how that effects the value of the landscape as a whole doesn't mitigate the responsibility to talk about it, or imply that there's a missing larger picture.

> Google will do what they want to do and they're big enough to where if they lose a few million hardcore advocates in the process it's no skin off their nose.

If you're right, and we've reached the point where Google essentially no longer has management and engineering talent that cares to be a good steward of the web or no longer has the incentives to understand how AMP isn't, then that's a much heavier indictment of Google than most of AMP's opponents have leveled so far.


If we take two steps forward and one step back, it's not without merit to think about the step taken backwards and see if that loss was necessary or incidental to the steps forward. If it incidental, to see if it could/should be recovered.

And that's even if you can get everyone to agree that weighing the different aspects of a change says it's an overall benefit.


One solution punishes people, the other one rewards people and provides a nice framework where every improvement benefits everyone using it. For example, at I/O, they announced that new changes to AMP was able to speed up rendering of all AMP based by 200%, and that's without any work being done by any of the people using it.


Google is already using the ranking system to promote certain websites over others, this would be just one more factor (and presumably it already is a factor of sorts, but with a different weight).

Even if AMP pages would render 1000 times as fast it wouldn't matter, it would still be an abuse of power. The relationship between a reader and the source of their content should not be mediated by Google. What point is there in HTTPS everywhere campaigns and such if Google MITMs every news article you read? What point is there for a content provider to even have a website if Google will end up serving the content?

This is simply bad, the only upside is the speed gain and that is one that could be sorted out by the market with a gentle nudge from Google rather than by some kind of monopolistic end-game on their part.


It's literally a CDN. So in your mind is Cloudflare also abusing their power by rehosting people's content?


To make matters worse, at least for some people, you can't tell which search result leads to to an AMP page, and which doesn't -- they have removed the little AMP logo.

Urgh.


I like that Google's pushing for simplified pages. This is a good hedge against FB Instant Articles and the like.

What I don't understand is why they cache the page themselves, and make that the requested version.

It would seem like any cached page from a CDN would be just as quick.


Because it changes the relationship in a very important way. It makes Google the point of contact and the supplier of the page a mere 'information provider' rather than that Google is a source of traffic and the website a property with traffic of its own.


The main reason Google caches the pages (when a user comes from Google search) is so that the page can be preloaded by the browser before the user even clicks. This preloading is part of the reason that the experience is so fast.

If Google preloaded the page from a non-google server, there would be a few problems. The most important one is that it would violate the user's privacy. That server would have in it's logs a request from a user that didn't click on the article, and could intuit information about the user's searching behavior from that request.

Less importantly, without the page being served from Google's cache, there is no way to guarantee it's actually a valid AMP page. The only thing Google knows is that it was valid AMP the last time it was crawled, which could be days ago.


> This preloading is part of the reason that the experience is so fast.

And a waste of bandwidth if the user never clicks.


I'd bet users (in the America) who are (a) valuable to advertisers or (b) likely to purchase newspaper subscriptions don't care about bandwidth usage.


That just leaves another 7.8 billion of us...


> If Google preloaded the page from a non-google server, there would be a few problems. The most important one is that it would violate the user's privacy.

That's one problem.

> without the page being served from Google's cache, there is no way to guarantee it's actually a valid AMP page

Yes there is. You can validate pages as AMP on any CDN. You just can't have any benefit unless they're hosted on Google CDN.


The biggest unsaid reason is they now own all the data. The AMP website becomes just another page on google.com - giving them direct access to know who you are and what you're reading with 1st party data rights.


Happy to have a technical discussion, but not sure you are really interested.

TL;dr AMP is 100% build on web tech. Pre-rendering achieves the performance but relies on history.pushState and iframes, which doesn't allow presenting a URL that is not on the same origin as the search page.

We think that the UX trade off in the URL is OK given the performance benefits. Given that it is a trade off we now have a bunch of projects to mitigate those trade offs. More coming soon, including lots of improvements to Safari. My team works directly on WebKit to fix bugs that affect AMP (but also the web as a whole, since AMP is just web tech).


Is someone working on allowing users to opt out of AMP when viewing search results? For users whose internet connections are fast and who cannot tolerate the browser UX being broken (can't copy URL out of address bar, back button doesn't always work properly, web pages often don't have comments, the list goes on and on seemingly forever).

Also, is someone working on a successor to AMP that doesn't break or try to replace the browser's own UI? Maybe something done at a lower level, maybe the HTTPS level? E.g., I could imagine something where the browser is informed that www.google.com is loading/proxying a web page on the user's behalf, and through some kind of verification system, it determines the website has given www.google.com permission to do so, and the browser UI updates itself to show the right address in the URL, etc. Maybe there's a better way to do all that, just one idea off the top of my head that seems infinitely better than AMP from a UX perspective.

Though it'd also be nice if you could get websites to make their actual main sites faster for everyone, maybe using a method like jacquesm suggested. That coupled with the low level proxying might be a nice alternative to what we have now.


> Happy to have a technical discussion, but not sure you are reallty interested.

This is not about technology. And whatever goal the AMP project states it has could have been achieved - and better - in other ways, such as the suggestion elsewhere in this thread to simply penalize page weight. If people wanted a consistent user interface across all websites they would have stuck with Videotext.

> TL;dr AMP is 100% build on web tech.

Yes, so was the search engine that put each result page in an iframe. Only difference was they didn't have a monopoly on search. And it does not make it right.

> Pre-rendering achieves the performance but relies on history.pushState and iframes, which doesn't allow presenting a URL that is not on the same origin as the search page.

Minor technical details, not relevant. If you feel the AMP discussion is going to be swayed by technical bits you're simply out of touch.

> We think that the UX trade off in the URL is OK given the performance benefits.

Who made you the deciders of what the UX of the web should look like? Stick to generating the best search results rather than trying to co-opt the entire web one little bit at the time and leave the UX to the browsers, it would seem you have enough input there already.

> Given that it is a trade off we now have a bunch of projects to mitigate those trade offs.

The only trade-off that will satisfy me is AMP dying off because websites will stop to support it. But as long as Google is strong-arming content providers to use AMP that won't happen. It is no longer fair play as far as Google is concerned. If it ever was.

> My team works directly on WebKit to fix bugs that affects AMP (but also the web as a whole, since AMP is just web tech).

Consider doing something more useful with your talents. For instance, fix the long standing issue with the SERPs that makes it impossible to cut-and-paste URLs pointing to PDFs.


This post feels overly agressive and hyperbolic. There's no need to insult their job or bring out the accusations about the reasoning behind AMP.


I also often find that discussion of motives is murky territory but:

* The post it's replying to admitted motives behind those critical of AMP to the discussion.

* The idea that page weight is a problem best solved by a solution like AMP is hostile enough to the principles the web is based on that it's hard to bar either motive or outright thoughtlessness from the discussion. Defenses on either front are also admissible.


I find it a bit strange to be called out for being 'overly agressive and hyperbolic' when that is in fact exactly how I feel about AMP in the first place, it is an overly agressive move by an extremely large company and my post is just one individual opinion which is - pre-emptively at that - already going to be ignored.

As for the job insult: there are plenty of UI/UX issues that Google could fix tomorrow if they wanted but instead they chose to work on this abomination and on top of that caused a bunch of work for companies well outside the Google ecosystem just to make things work that weren't broken before. That's a very inefficient way to allocate one's resources even at the scale of Google.

Finally, DanG has written a lot about the 'principle of charity', it appears to me that you have done everything that you could in order to take the very least charitable view of my comment and then you attack me on that interpretation.

I care about the web, in a way that I don't care about much else (HN maybe), and any and all attempts to subvert it should - in my opinion, which you are free to ignore - be fought tooth and nail lest we lose what came at a pretty great price. If you're comfortable with walled gardens and with large corporations gobbling up more and more of the open web to put it behind their store fronts then that too is fine with me. But you're not going to shut me up on this subject, nor am I going to moderate my language because you feel I should. Take the tone as a measure of my feelings on the subject.


Well, the post he/she was replying to had a provocative line in it as well, so I think we can excuse the tone a bit.


I think it’s ok. Personally, I like it.


The post is critical and frank but well within the bounds of civil discourse. Pleasant and understanding is not always the appropriate tone (though it's a good default). Sometimes people or companies do harmful things. To respond with patience and respect is to normalize them.


> The post is critical and frank but well within the bounds of civil discourse.

Yeah, this is really within the bounds of "civil discourse":

> Consider doing something more useful with your talents.


If that sentence had been followed by something sarcastic, derogatory, or harmful, then I'd agree. However, since it's followed by "here's an actual problem I'm having that Google could fix", I choose to believe that jacquesm is honestly saying "you've got a lot of bright people there, but AMP is the wrong project to have them spend time on".


> Who made you the deciders of what the UX of the web should look like?

It's not like they're forcing you to click on the AMP results. When Chrome started experimenting with SPDY, one could have asked them "who made you the deciders of what the protocol for HTTP transfer should be?". Yet from the lessons learned from SPDY was born HTTP/2.


When using Google search the only link given is the AMP link.


That's not true. If I search <<trump on afghanistan>> I get an AMP result, a non-AMP result, the AMP carousel, another non-AMP result, another carousel, another non-AMP result...

If I were only interested in results from Reuters (the top result in the previous search, which was AMP), <<trump on afghanistan reuters>> returns 3 "condensed" AMP results and a non-AMP result above the fold, followed by 8 non-AMP results.


Except some people are now unable to tell which results link to AMP pages. They are experimenting with removing the little AMP indicator logo.


I just wanted to say that I appreciate your talking the time to argue with cramforce on these points. (I just realized how ironic that username is for the TL of AMP, which is being crammed/forced onto users. Good thing that "most users love the experience".)

I don't think you're going to get far though as you're obviously just "opinionated" and probably not "really interested" in the "technical discussion".

As long as "We know that most users love the experience" why deign to discuss it with the ones that don't?


I'm not certain about this, but my thinking at the moment is that as soon as you need to ask a browser vendor to change their code to accomodate your design decisions, it looks a lot less like the "open" web, and a lot more like something else.


I appreciate the info!

I'm really curious to see where using AMP as a proxy speeds up the page, versus applying similar optimizations on origin server?

If it's a cached page (served from any normal CDN), and crafted with similar optimizations [1] wouldn't it be just as fast, without needing to do the URL redirection?

Even if it were difficult to convince developers to implement the changes, couldn't many of the optimizations could be rolled into a module like PageSpeed [2]?

[1] - https://medium.com/@cramforce/why-amp-is-fast-7d2ff1f48597

[2] - https://developers.google.com/speed/pagespeed/module/

Edit - Reformatted links; I forgot HN doesn't use Markdown.


Our latest numbers show that the average speed improvement of the AMP cache is roughly 75%. I'm a bit shocked by that myself. But that is where it stands.

Getting similar performance out of on-origin serving is a priority for us, but we aren't there yet. One challenge is to give it the same scale as the AMP cache.


Could you at least make it possible for websites to opt out of the AMP cache? Last time I checked, publishing an AMP page grants Google (and everyone else) an implicit license to incorporate your content on their own web properties. The AMP FAQ say: "Should you desire not to have your document cached, one option is to remove the amp attribute from the HTML tag. This makes the document technically invalid AMP, while not impacting the functionality of the document." But from what I understand, this will also impact the presentation and ranking in search results.


> Happy to have a technical discussion, but not sure you are really interested.

Criticism of AMP certainly includes technical aspects, but most objections I encounter are ethical rather than technical.

Did Google explicitly consider ethical considerations when creating and launching AMP? If so, what did this process look like, and where could we find out more?

If not, isn't it time we expected ethical review to be part of significant changes in web infrastructure, the same way we would with significant changes in physical infrastructure?


Funny how literaly tens of behavioural trackers on sites that follow you everywhere, abuse holes in browser sandboxes to steal private data and try to install malware did not bring quarter of vitrol here on HN. More, bunch of people defended that as "we need to make profit".

Now that someone got rid of that it's suddenly an "ethical issue"?!


It's an ethical issue because I can easily block ads and malware.

I cannot opt out of AMP if I use Google as my primary search engine. It's about market power.


We have somewhat different expectations of Google than malware sites, because we expect it to act as a good net citizen, and a thoughtful custodian of its enormous power to shape the future of our culture.

The risk critics are concerned about here is a future where the internet no longer "belongs" to the people on it anymore, and despite the malicious nature of the abuses you named, they hardly represented that class of threat.


I've heard so much negative feedback about AMP that I'm not sure I have the full story. What link(s) would you recommend someone review if they wanted to build their own informed decision on the merits of this technology?



Today is the first time I've heard of AMP, so I don't have a dog in this hunt. I visited each of your links and followed some links on those pages, and I still have no idea what AMP is. All I see is a promise that pages will load faster. It also seems that if my boss drinks the koolaid and says "We need AMP!" it's going to cost me money in developer time, because the technology is inscrutable to my unskilled content providers. What benefits does it provide that disciplined web design doesn't? Is Google going to penalize my sites if we don't use it? We don't include advertising in our pages, so should we care about AMP?


AMP is mostly a series of guidelines. https://www.ampproject.org/learn/about-how/ If your site follows the guidelines, it will be fast. If you also voluntarily add a bit of JS to your page, Google (and other sites, but mostly Google) will cache your article and deliver it from their own CDN instead of loading it from your site. They won't interfere with your ads if you have them set up right. The trouble is that the URL is now a Google URL and Google puts links to more Google pages at the top of the interface instead of other links to your articles.

So, instead of actually using AMP, you could just follow the AMP guidelines and get most of the speed benefits with almost no downsides. But Google started prioritizing AMP pages in search results, so now there's a downside either way.


https://www.ampproject.org/how-it-works/ is a good primer.

You can beat AMP's performance with hand-tuned optimizations. Its goal is to uplift performance across a significant percentage of web content (instead of just that developed by the few experts).

Is Google going to penalize my sites if we don't use it? No, AMP is not a ranking factor.

We don't include advertising in our pages, so should we care about AMP? Advertising is to AMP like Advertising is to the Web. E-commerce is a big use case where advertising is uncommon.


Hi cramforce,

I understand that AMP is not a ranking factor (except for its inherent speed, perhaps). However, is there not a carousal at the top of SERPs for pages in your AMP Cache?

So while not an organic ranking factor, it would still affect SEO. Is that a fair comment? Or am I missing another perspective?


Not only that but site speed is a ranking factor. This answer definitely skirts around the reality that both we and he know since it has come up many times in the past.

As far as I see it, Google sees this as valuable to their interests, and have made clear what the talking points are and what should and should not be responded to. That's the only reason I can think of for this counter argument to be continually ignored by him.


How is featuring a carousel of AMP sites not penalizing non-AMP sites? You're being purposefully intellectually dishonest, maybe even with yourself.


Users loved AOL keywords. Here we discuss technical merit.


I actually think AOL keywords worked fantastically.


People here are users too. Unique culture, sure - but this audience counts.


I'm a user. I DON'T love it. Why isn't my voice respected?

In fact all the people here are users. Funny how that works.


I really dislike it as well, as I think a sizable portion of the HN does according to these threads that occasionally appear. You're in good company.


I love AMP, throwing my voice in the ring.


I also definitely don’t love AMP.


When will I be able to get into the AMP carousel without having to import the AMP script from Google’s servers?

I’d prefer hosting all JS locally. (Or at least within of EU jurisdiction, as Privacy Shield is likely going to fall in the courts, too, as Safe Harbor did before, and I don't want to end up liable for that (and I don't want to sell out my users to some foreign tracking company))


I'm in the same boat. To me as a user, AMP has worked well. Loads content quickly, with from what I've experienced, very little drawbacks.

So I'm not sure whether AMP is just something controversial as a lot of things from Google tend to be it seems, or if there is some real technical merit as to why it's so bad apparently.


You can do the rendering, even within an iframe, on iOS (where iframes have an amazing quirk of resizing automatically) without messing with scrolling (both vertical velocity and horizontal swipes) and breaking the browser's on-page search (which is extremely annoying); in fact, it is so easy to pull this stuff off on iOS (again: in no small part due to the epic iframe quirk) that it kind of comes off like you just don't care.


> We think that the UX trade off in the URL is OK given the performance benefits.

So basically the argument boils down to "because I have shit internet, everyone else should suffer too".


If you hate AMP you should try and use DuckDuckGo.com these days.

I'm astounded at how much better it is than the last time I tried it.

Switched to it in Chrome a few weeks ago and don't miss Google. Which is shocking to me.


DuckDuckGo feels useless for almost every programming related search queries


I've had a different experience.

It's been successfully bringing me the correct stackoverflow results I would expect when I need to search.


I really like DuckDuckGo, but I'm not super happy to get Taylor Swift results when searching for iOS/Swift related things :(


The fact that DuckDuckGo embed a SO snippet of the search query you entered makes it very useful for me.


In my experience it's getting better every few months. A year ago compared to now feels like it's a different search engine and I find new stuff every day, for example, vim cheat sheet


I also thought that months ago, but it seems much better now.

Plus, I'm really starting to like the bangs, which allow you to search language-specific docs directly without any setup.


Not to be pedantic, but you should switch Chrome too. Firefox 57 is pretty amazing (currently in nightly).


I tried this recently because ideologically I agree. But the small issues build up and make me less productive.

For example, if you use a U2F key, you need an extension for that to work - and that extension doesn't work with FF 57. I honestly don't remember the last time something broke with Chrome.

Or smaller things, like the fact that I can't use the keyboard to manage bookmarks (in FF on macOS, backspace doesn't delete bookmarks).

But the worst is in FF, Ctrl+Shift+N isn't new incognito window, it's restore the previous window. Years of muscle memory don't go away overnight.


Chances are that extension is not yet ported to the WebExtensions API. It should probably be ported eventually, as precisely one of the benefits of WebExtensions is that it makes way more easy to port a Chrome extension to a Firefox one and viceversa. Regarding the shortcuts, I feel you. There's an extension for that as well (Menu Wizard) but it has also not been ported. I guess it's too soon yet.


I'm trailing Brave out at the moment and like it but plugins are important to me.

Firefox is an interesting choice but the Chrome plugin ecosystem seems much more complete for me than Firefox has today.

It'll be interesting to see what happens in the space. I was a Mozilla user, then Firefox and now Chrome user. I have no issue going to better platforms.


That's not pedantic, just unrelated.


Hmm I wonder if it's possible to create a kind of independent benchmark for search engine results (I guess possibly doomed to failure due to its inherent subjective nature).

I've been trying to use DDG more, unfortunately I'm not so impressed. Quite often end up researching for Google.


I find it really strange that people think that. AMP's competitors are Apple News and Facebook Instant Articles. They've done this in the spirit of the open Web - but they get some benefit too. That's why they build most of their services.


It's not "the web" vs AMP, it's AMP vs Facebook. Old school web is dead on mobile.


> TL of AMP here

Hello. Sorry, but I have to express my deep disgust for what you do. "Give us your content and we'll rank you higher' is the opposite of what made the web and Google awesome.

If you care about user experience, downrank pages based on CPU usage, Memory usage, number of tcp/ip connections, number of domains connected, number of ips connected, number of bytes transferred, percentage of time that the UI is non-responsive and so on.

Trying to force the web into your little AMP cage is a terrible idea. It's the Facebook way of doing things. Are you really so afraid that you have to join the crooks? Would you have managed to take such a big part of the pie from Microsoft if you had been afraid of them and played their game?

Don't fight oppression with oppression. Fight it with freedom. Get rid of that AMP crap.


> Don't fight oppression with oppression. Fight it with freedom. Get rid of that AMP crap.

The only problem I have with AMP is that (google search) users can't turn it off. Me - I love AMP. You would love it too if you had to browse the internet on shitty third-world 3G connections(if you're lucky! Otherwise you're stuck on EDGE or GPRS). Not all CDNs are equal when browsing from less developed countries.

AMP is great for the next billion internet users, not so much for the HN crowd who don't mind multi-MB page loads without worrying about timeouts. Google should make AMP opt-out so that those who can afford to turn it off have the choice.


I can't agree with this enough. Even though I have a 100Mbps unlimited plan, doesn't mean everyone can and will have it. I can't say I love AMP, but I definitely support it.

Purely from the reach it can potentially have if everyone does start using is great. Everybody is already optimizing so that they get crawled by Google, better faster and more efficiently. This is the logical next step.

From a ranking perspective, I think it's important that the HN community stops thinking just in terms of geographical metrics. Yes, in the US, UK etc where our offices are it gets blotched. But our next users are not going to be there from these countries. If at all, being a PM for a global product here, tomorrow's users, paying or free will be from probably from a country where AMP users would be the most impressionable today.


    You would love it too if you had to browse
    the internet on shitty third-world 3G connections
I did not say I don't like lightweight pages. I say enforce them via ranking down the heavy ones. Not by blackmailing publishers to hand over their content.


This is why I am developing the AMP Browser (http://ampbrowser.com) - it is a godsend in a slow 2G/EDGE network.


If AMP was so innocent Google would give users an easy way to disable it. It seems trivial to do, the fact that it's not in V1 says to me that there's ulterior motives.

There's more reason Google wants you to use AMP than "it makes the web faster!". That's just the sell. They REALLY want you to use it, enough that they don't want you turn it off. Maybe something to do with protecting ad revenue?

I heard that AMP was developed by people connected to the search team. And search team is probably another name for "AdWords cash cow team"


Ulterior motive? Thought it was pretty obvious, shitty mobile page performance pushes people towards using ad blockers, which is obviously bad news for google. Most people don't mind basic, non-intrusive ads so much, which is what AMP retains.


Why not just punish slow sites? It would have been a lot easier than the zillion hours spent building AMP and the caching system.


There was a previous discussion on HN that helped me understand the value of Google being "the bad cop" and making AMP restrictive. Before AMP, you'd have product people and ads people demanding to add all these bullshit ads and trackers to pages. Engineers would often try to push back with the performance argument, but it was a difficult argument to win (ads people could argue, "well, it doesn't really slow down the pages that much, and look at this revenue we get - maybe you just are coding it wrong when you're including the ads code!") With AMP, the discussion is very simple: "Sorry ads guy, I can't add your bullshit ad network JavaScript because it's not possible with AMP. You know Google favors AMP now, don't you, and you wouldn't want to be the guy responsible for a shitty SEO position because we didn't use AMP, do you?"


This hypothesis of "what's in it for google" bends over backwards ascribe altruistic motive, and also relies on the (deeply alarming!) idea that AMP use leads to better search rankings.

It's worth noting that since your comment was made, the idea "AMP use leads to better search rankings" is now directly contradicted by the TL of AMP (cramforce)[1] in a comment this thread [2].

[1] https://news.ycombinator.com/item?id=15086192

[2] https://news.ycombinator.com/item?id=15087277


I should clarify: regardless of whether or not it is used in search "rankings", Google search results clearly consist of an AMP carousel high on the page for many searches, so it's kind of silly to argue that AMP doesn't affect "rankings", as it's patently obvious that it affects where results can show up on a search results page.

Also, FWIW, I don't have any issue with AMP as a technology, but I have a big issue with how AMP cache is used by Google.


That's the irony of all of this. Cramforce is claiming AMP isn't used for rankings but that is disingenuous. AMP pages are clearly featured in the carousel, and Google has gone on record saying page speed is a factor.

So AMP at the very least indirectly impacts rankings (and thus impacts rankings), and it may directly impact the ability to appear in the carousel (which impacts rankings).

Would love for him to refute this, but I suspect that won't happen.


I'm not part of google, but they had been doing that for years, and still nobody was working on speeding up their sites.


Also, if every site slows down, the fastest ones ares still slower than before.


Because that would not achieve the actual goal: to get people to stay on Google's properties, preferably without an ad blocker installed.

Note how over time more and more of the things that Google used to simply point to are served by Google directly. AMP is just another element in that strategy.


Interesting because actually it was AMP that pushed me out of Google properties, to Firefox for Android with uBlock and DuckDuckGo and I'm planning to stay using this setup even though Google's engine is better (minus AMP).


I'm one of the people who severely dislikes AMP and I've always wanted to know: why not give me an option to disable it?


I'm personally neutral on it. In general, product managers hate complicating the product experience with settings and engineers hate maintaining multiple code paths for the same feature.

We did add the link away from AMP, so that users who cared have a per-link opt out. We know that most users love the experience.


I understand you don't want to keep 100 billion options around, that makes perfect sense.

But haven't you seen the level of vitriol that comes up every time AMP is discussed? This isn't a minor thing, a lot of people REALLY dislike it. I have SERIOUSLY considered switching to a different search engine entirely because of this.

The fact that there is now a link to get to the normal page after I've already gotten to the AMP page is not a solution to the issue. As an end-user that feels like I'm complaining that her website is too hard to read because of the noisy background and you're telling me I could just use print preview where the background doesn't show and read it that way.

This feature has significantly change the way I use Google for the worse, and in the time since it was first launched it really hasn't gotten any better for me. It's so bad my OS vendor is starting to incorporate workarounds directly into their code.

If I Google something how do I open five tabs to go look at? I have to ignore every AMP result because I can't open them in a tab. So would Google deems to be the "best" content is now unavailable to me.

I don't have a choice to simply turn it off.

Not everyone will. If you read the comments here there are people who do like it.

But there are clearly a lot of us who HATE it with a passion because it's made your website so much harder to use.

PLEASE give me an option to turn it off.


> This isn't a minor thing, a lot of people REALLY dislike it.

> But there are clearly a lot of us who HATE it with a passion because it's made your website so much harder to use.

You really seem to overestimate the size of HN readership compared to the rest of the internet.

It is a minor thing. We are a niche audience living in a bubble. We're not - by far - the typical or target users. Our usage behavior is - mostly - irrelevant to Google.

The world has over 3 billion internet users. Let's say that HN has 3M unique monthly users, and that only 10% of those comment, and they are split between liking and hating AMP.

That gives you 150k people who hate AMP compared to 3 billion users on the internet.

Why would you build and maintain a feature for 0.005% of your users?


> Why would you build and maintain a feature for 0.005% of your users?

Even taking at face value the assumptions that yield your numbers (not obvious to me that's safe):

150K is still a pretty large audience.

It's an influential audience, to the extent that tech and therefore thinking about tech is influential.

It's an audience that understands the implications of technology choices like this and can articulate them.

Also, I don't understand the assumption that this audience is completely complementary when it comes to understanding how the rest of the population thinks. I expect it's at least as likely a bet that there are people in the rest of the population who sense something is off or some aspect of the experience is degraded but can't understand and articulate the issue at as high a resolution as it is likely that everyone in the rest of the population simple does not care.

And you really don't have to be some kind of programmer or other ubernerd to see some of the relevant issues in play with regards to URLs and walled gardens.


> It's an influential audience

I know HN loves to think they're influential but I'd suggest than less than two handfuls of HN users actually have measurable influence outside of HN.


> 150K is still a pretty large audience.

No, it's tiny. And considering how HN readers usually user adblock, and rarely click on ads, it is a completely irrelevant audience for Google.

But let's say you make $10 out of each user per year. That's $1.5M in yearly revenue. It barely pays for 3 employees at a large tech firm.

> It's an influential audience, to the extent that tech and therefore thinking about tech is influential.

>It's an audience that understands the implications of technology choices like this and can articulate them.

Considering how even HN readers are divided about AMP, I'd say that the net impact of HN audience on AMP is zero. We're not as influential outside the tech bubble as you think.


Barely influential within it either.


You're assuming that the only people who dislike AMP or the people commenting on hacker news.

What if it was 10% of all users? Or maybe 20%?

Where would you draw the line… what percentage of users would have to dislike amp for you to think it was reasonable to add an option?

I'm not sure what number I would choose (ignoring my bias that I don't like it and therefore my number would be smaller than probably reasonable).


> Where would you draw the line… what percentage of users would have to dislike amp for you to think it was reasonable to add an option?

One easy piece of math is (cost of adding and maintaining that feature) vs. [(users who dislike) x (% of users who would quit you) x (profit per user)]

For example, if 10% of all users disliked it (300M users), but only 1% of those would leave your services (3M users) and you made, on average, $10 per user in profit, you'd be putting $30M/year of profit at risk, which is probably enough to hire a few people to build and maintain that feature.


> But haven't you seen the level of vitriol that comes up every time AMP is discussed?

Mostly on HN. And it's clear from this thread that a lot of vitriol comes from deeply ideological positions rather than real technical reasons. That won't be fixed by adding an opt-out.


You're right that would remain no matter what. I don't care that much on the ideological position. My objection is purely from a user experience standpoint.

My assumption is that the idea logical issue is much stronger here on hacker news than among the public at large. Much like views on free software purity.

My guess, for lack of better reasoning, is that the percentage of people on hacker news who hate the experience is roughly the same as those in the "real world". I wouldn't expect that being more technically inclined would make people more likely to like AMP or dislike it (ignoring idea logical issues).


I hope that the option of an 'opt out' or 'disable AMP' option (whether globally or on a per-result basis) is explored in the future. As an iOS user I have found AMP to be a frustrating experience. I'm sure you are including iOS users when you say "We know that most users love the experience" but for me AMP has been annoying.

I wrote a comment earlier about my experience with AMP from my perspective as purely a user. Please excuse some of the rantiness of it. Like i said, I can get a little frustrated with AMP at times.

https://news.ycombinator.com/item?id=15087504


> We did add the link away from AMP, so that users who cared have a per-link opt out. We know that most users love the experience.

You know that most users love the experience, or you know that most users take the path of least resistance and don't click away using a UI whose purpose isn't particularly clear yet?


> We know that most users love the experience.

How exactly do you know that?


I can't go into details, but in general we closely track analytics for our product (like most companies with websites would) and run AB experiments to see how changes impact user behavior.


I really question drawing a conclusion from analytics saying people are using it versus actually liking it. I know some folks at work have used it and they regard it as broken enough to call in a ticket about their device being broke.


Dude just said he can't go into details, and you question the method he didn't explain?


Analytics is a pretty well known idea. I question the base theory that it will answer this specific question.

Also, given that the indicator for an AMP page is a greyscale, hard to see icon and words, people who don't want to use AMP might hit it because they are in a hurry and the text they are looking at is black and noticeable.


And they are experimenting with removing the AMP page indicator in search results.


As in it won't go to an AMP page ever or as in it will go to an AMP page without any indication?


The latter.


I don't think usage implies some sort of favorable opinion. They don't really have a choice. Many people may not even realize what is happening.


If you can't go into details then you're likely doing something illegal.

Which means even less reason to trust you.


You misunderstand the issue.

Google search basically is a function where you input a search string and it returns a list of URL's matching the search criteria. That's all I need and want it to do. AMP breaks the core behaviour of Google Search, it no longer provides a list of URL's that match my criteria. Basically most search results aren't actual search results anymore.


There's a happy medium between a product with too many settings and code paths and a product that has none. We're not asking for AMP to run in a plethora of different ways, we're asking for an on/off switch.


As long as it has AMP Google Search is completely unusable to me. It needs to die TODAY.


Your response feels like a lame excuse, and I think is insulting to the HN crowd.


Better still: why not give you an option to enable it and default to normal.


Because the majority of people enjoy it? And the common user wants fast load and doesn't care about the philosophy of data retention and web ownership?


Reddit links via Google search results were broken for months on my iPhone because of AMP. How many other users had the same issue and no idea what was going on?


That's reddit's fault and isn't anything inherently wrong with AMP itself. Just because someone uses the technology wrong doesn't mean we should hate on the technology. If I use AngularJS to make a Neo Nazi site, should we all boycott that library?


How do you know this?

It's equally possible the majority of people just don't care. That doesn't mean they like AMP or would op-in to it given the choice.


> And the common user wants fast load and doesn't care about the philosophy of data retention and web ownership?

While this is most certainly true, it doesn't mean that data retention and web ownership aren't values worth fighting for.


At some point you have to realize you're an outlier. The vast majority or users appear to tolerate and/or enjoy AMP just fine.


The fact that the vast majority of users wants something does not make it right any more than letting your kids decide what's for dinner every day is right. At some level what people want in the short term is overruled by the longer term consequences.

This whole discussion reminds me of the people ranting against RMS who many years later come to the conclusion that he probably was right all along only by then the horse has bolted.

The long term consequences of AMP are to consolidate Google's power and to make Google even more of an intermediary deciding who gets to see what. It's an end run around the kind of openness that is a cornerstone of the web and I care deeply about that openness and will oppose vocally any attempt to hijack the web for fig-leaf reasons.

Note how Google is slowly boiling the frog by doing all those things that they said they'd 'never do' and how their 'don't be evil' motto gets trotted out time and again to show us how they are good little web citizens when in fact they ram their unilateral solutions down our throats in the interest of their bottom line while pretending it is all about the 'user experience'. Accidentally gaining control over the content delivery of the pages containing advertising tags of competing ad networks sure must look like a nice and juicy bonus.


See… I really don't care about the ideology here. If people want to get in bed with someone who doesn't have their best interest at heart that's their choice.

Borders and Toys'R'Us both decided to partner with Amazon for a while. I'm not sure why they thought that was a good idea, but that was their choice.

What I care about is the user experience. That's where I'd be happy with an option to simply disable it on my devices.

I think it's kind of unfortunate that it's not possible for people to discuss AMP without both issues coming up and confusing things. But that's life.


Thankfully, We don't rank RFCs or make technical choices based on what users appear to tolerate and/or enjoy. Nor do we set ethical laws or guidelines based on what citizens tolerate. Opinions of people trained/educated in a specialized domain will always carry more weight than that of a random person.

But forget all that, on a simpler level, if you do nothing most things stay the same. If you're happy with the status-quo, you're welcome to be silent. Others prefer vocal protest.


I would be VERY interested in some sort of study that shows that the vast majority (or even A majority) of users enjoy AMP.

If the majority of users merely tolerate AMP... I wouldn't exactly call that a win for Google.


Have you heard of a company saying their users hate their products? :)

A Google engineer saying their users love their products is probably expected. They were probably fed this by some marketing/sales droid.


Im in the camp that also hates AMP with a passion. It feels weird not liking much of what Google has been doing lately but AMP for me gets in the way of me sharing articles with my friends and family. It also gets in the way of me sharing a clip of the text so the people I'm sharing with including myself have the proper link and some context to get them to read further.


> sharing a clip of the text so the people I'm sharing with including myself have the proper link and some context to get them to read further.

How do you achieve this with a non-AMP website? It sounds great.


i really like AMP, and i wish google would give people an option to disable it just so we could have a discussion about AMP on HN that wasn't completely taken over by people whining about how AMP is an evil plot to destroy the web because it scrolls funny on iPhones.


AMP is Google's new closed web, like AOL. So the opt-out is to find alternatives. If you don't have any alternatives, then Google can do whatever they want to you.


There is another open source alternative called Mobile Instant Pages (MIP) https://www.mipengine.org/


Please, please let us have an option to only see real links in search results... In my experience, AMP has been so annoying that I just want to disable it entirely for my Google account - why wouldn't this be an option?


This might not be a great workaround, but Firefox doesn't "do" AMP. My brother switched to FF on his phone just to avoid AMP pages. As a bonus you can use uBlock Origin on mobile Firefox.


Please get rid if AMP as soon as possible, I had to switch to DDG on my phone just to avoid it. I didn’t buy a high-end smartphone with fast 4G to browse a crippled version of the web.

Congratulations, you re-invented WAP and turned back the clock on the mobile web 20 years.


Really? Because I can't count the number of times that I have attempted to share an open web URL from a Google product or website only to have the AMP URL turn up when I paste.


Yea. How about Chrome?


UPDATE: We confirmed that this actually implemented as a generic solution. (Sharing <link rel=canonical>), so there is nothing AMP specific here.


Brace for angry comments. (I'm neutral on AMP personally, but it has a really bad rep here.)


I’m neutral on most technology until I get a call that the ‘web is broken’ then I care. AMP made the jump for me this week.


What about a feature in Safari that disable Amp completely? That would be neat, seeing as Google don't seem to offer it as an option...


Just to be clear, are you "the" or "one of many" Tech Lead(s) at Google for AMP?


So…

In terms of the AMP open source project governance rules I am THE tech lead. https://github.com/ampproject/amphtml/blob/master/GOVERNANCE...

I also manage a team at Google that works full time on the AMP open source projects and act as an overarching tech lead, but there are several sub teams that have specialized tech leads (e.g. for UI, ads, analytics, caching, etc.).

Is that what you wanted to know? :)


Yes, thanks for the reply; also commented in part because thought the use of "TL" without spelling it out might need clarification too.


It did. Honestly I had no idea what it stood for. Thanks.


What's the deal with removing the AMP indicator logo in search results?


Could you definitively state whether or not AMP pages are ranked higher within searches or will be in the future? If so, please stop. Thanks.


AMP is not a ranking factor in Google Search. More official statement here https://twitter.com/johnmu/status/824185977098960897



Not op, but: There is an official statement from Google that they are not higher ranked within of the results, but AMP links can get indirect ranking boosts, for example Google lists the top 4 or 5 relevant AMP links before the actual results in a carousel.

This is identical to the ranking boost given to Google Shopping (for which Google was just fined 3 billion EUR), so you can expect the same result.


> This is identical to the ranking boost given to Google Shopping (for which Google was just fined 3 billion EUR), so you can expect the same result.

Isn't it different because (presumably) the AMP links aren't owned by Google?

The fine was specifically for anti-competitive behavior, i.e. consistently surfacing their own links over those of shopping competitors, which at least with AMP links isn't necessarily the behavior, though there is likely some overlap there.


It is actually.

Google only gives the carousel position benefit to sites using the AMP js from Google's CDN.

If you use any other CDN, you don't get the benefit.


Where's Google's upside in all this? This (maybe) improves their search product, but it doesn't directly make them money.

From a legal standpoint, this behavior is clearly different from their last slap on the wrist.


Google's upside is that the amp website on google is fast while everwhere else it's slower.

Google doesn't want the web to be fast or lean, it incentivizes you to create a fast version to be shown from google while the slow version is shown everywhere else.

Google: fast, everyone else: slow.

And amp positioning and ui display is absolutely a ranking boost.


> everwhere else it's slower.

No, everywhere else it's the normal speed your servers can deliver it to clients. Nothing about AMP is slowing down your normal web page.

The positioning and display are a ranking boost for whoever chooses to implement AMP pages - they are not explicitly targeted at Google services, which is the behavior the EU fined them for.

Anyone who thinks the EU is going to fine Google for AMP is actively avoiding reality at this point.


> No, everywhere else it's the normal speed

Faster and slower are relative terms.

> Nothing about AMP is slowing down your normal web page

Nothing about AMP is making the web faster, which it says it's doing on the project homepage. It's making a fork of the web to be shown on Google that is 'faster'.


I realize you hate AMP and rational discourse is unlikely to change your mind, but here's what their homepage actually says:

> The AMP Project is an open-source initiative aiming to make the web better for all. The project enables the creation of websites and ads that are consistently fast, beautiful and high-performing across devices and distribution platforms.


I think amp in spirit, practice and implementation is an attack on the open web.

I understand how it works and Your arguments so far aren't changing my mind.

Thanks for posting the homepage text, for posterity.



AMP supports dozens of ad networks, including AdSense. It gives no particular benefit to Google's ads, and it doesn't require any ads.


Except for all the ad networks that don't get approved...


Given the range of ad networks they support, it still gives no particular benefit to Google.


More minor, but AMP results get the little lightning bolt icon too.

I believe there is research showing that little adornments like that convey trust and increase the CTR vs other listings that don't have them.


Your product is at best a turd burger at worst, an undercooked turd burger.

The easiest way to do the "more generic solution" is to remove that turd burger from search results. Interwebz fixed.


OK, Google asked Apple for input on amp. When will Google ask the EFF?


Octember 27th.


You keep saying users "love the experience". Can you provide data to back that up?


>We'd prefer a more generic solution where browsers would share the canonical link by default

So, like regular webpage visits?


Is google working on opt-in or opt-out functionality, or is that something that might happen in the foreseeable future? I think a move like that would go a long way in the goodwill department.


How would this be implemented without special cases? Why should a platform care about AMP's policy?


Yeah, I don't see this either: amp as implemented is a mitm attack that delivers content in an opaque manner, as it's not a proxy at the http(s)/socks level - how would the user agent know that "amp.google.com/url=actual.url.example.com" should be rewritten, but "random.example.com/article?someParam=old.url.example.com&article=42" should not be rewritten?

Now Google has forever(?) broken links in search results, in order to count click-throughs - so I suppose all browsers should special case that too...

I can't for the life of me figure out why they married an ok html/css/js subset (amp pages) with an absurdly invasive proxy technology.

I hear they say "speed", and "75%" - but if that's going from 130ms to 50ms complete page render does it really matter?


The browser could read the canonical meta tag in the article HTML and just share that.


Right. But then you could provide a different canonical url on any site, altering which url is shared? (hence, I don't see how you could reasonably do this without "special casing").


Good point.


> "we specifically requested Apple"

"we fucked up again web to wap level and now everyone needs to jump along because sharing is royally screwed"


Why can't a site appear on the search carousel if it uses amp.js but is delivered from a fast, non-Google CDN?


> TL of AMP here

I really wish some interdimensional telepathic Fart would Goodbye Moonmen this TL into abandoning AMP.


What does "TL" stand for?


Tech Lead


Please get rid of AMP. You've ruined the internet.


I hope the next step is a way to strip AMP links from all URLs, backfilling the "Disable AMP" setting that Google ought to have provided.

AMP has always worked poorly on iOS: it has different scrolling, it breaks reader mode, and it breaks status bar autohide and jump-to-top. Perhaps Apple would be less hostile to AMP if the implementation were better.


AMP has always worked poorly on iOS...

This never made sense to me.

You can have natural scrolling, an accurate reader view, a working back button and other highly performant features in a mobile-optimized HTML and CSS standard like AMP. In fact, you practically have to go out of your way to break these things in the way that AMP does.

Why does AMP have to suck so much?


A significant portion of the time (more than 1/2) AMP URLs fail to load, forcing me to "Request Desktop Site" in order to load it.

When they do load, they fuckify the URL, so I can't copy/paste it normally. I know I can go in through the page's menus and get the real URL, but it makes it harder for no reason.

I consider the URL part of the page's UI, and commonly manually edit it. This makes it harder.

So far as I can see, there is no advantage to AMP pages, versus a fast and simple HTML page, with no external resource requests.

Sites can serve that themselves, rather than using Google.

In short, it breaks features I use, and doesn't provide advantages.


This is why I hate AMP.

People who like it can't seem to understand that it has a crappy user experience for some people and always assume you don't like it for "philosophical reasons".

I really don't care about AMP from that POV, I just wish it was a secondary link instead of the stealth replacement of where I wanted to go with a page that usually doesn't load and messes up my back button.


I’m with you. I expect in the coming days I will have to switch some of our elder’s search engines to DuckDuckGo to avoid problems. I just hate technology that changes the user experience in a broken way. Browsers are supposed to act in a certain way. Don’t break that. Breaking UI is why we had a hate on for Win 8. It really annoyed people with the changes.


> You can have natural scrolling, an accurate reader view, a working back button and other highly performant features in a mobile-optimized HTML and CSS standard like AMP. In fact, you practically have to go out of your way to break these things in the way that AMP does.

> Why does AMP have to suck so much?

Because it violates every assumption on how browsers should work. I would prefer a slow browser I understand to a fast one I can't control.


Don't forget that many pages won't even load if you have some adblockers loaded. My typical usage also requires a "reload without content blockers" to get anything at all to show up on AMP pages. Crazy idea, I know, but maybe Google could provide even a way to get to a non-AMP version from a SERP. Even better, let me disable AMP completely, but forcing it on everyone like this and refusing to back down just shows how they value their users, AKA their sources of revenue.


I thought the point of AMP was to force ads, by stopping ad blockers from working.


I'm not sure if it has changed recently, but I just tried it and the browser back button is working fine (it goes back to the search results).

My only real complaint about AMP is the scrolling - it doesn't feel natural and breaks shrinking the browser UI (I have an iPhone 5S, so screen real-estate is a premium).


Natural scrolling comes to safari in iOS 11, so every site will behave like amp.


Do you have a source with more information? Because that will drive me so mad I won't update iOS if it's true.



It also breaks "back" IIRC. They expect you to use the "x" on the page to navigate back, not.. you know, the back button built into the browser. That and the scrolling are what got me to switch my iphone to DDG.


Not breaking the back button should be #1 on every webpages priority list. If you break my back button I can almost guarantee I will stop using your site because in my mind the only sites that do that are spam/scams.


This is a general Google trend it seems. Google Images does it too: if you scroll left after viewing an image, you get the previous result instead of the page you were on. Really frustrating.


Ah, yes, that is incredibly annoying behaviour. I was getting annoyed at this just yesterday! That would be fine on iOS if it only happened when you flicked the image, but scrolling from the side edges of the screen in Safari should always navigate forward and back.

They must have actually had to do work to override the expected behaviour and break the user experience, which I just don't get.


Google blogs do this as well, showing you the "previous" blog post when you swipe left. Again, massive usability issue.

I don't understand why Google has to break these well-established UI conventions.


Don't even get me started about this Blogger thing. The triggers are so sensitive that half the time I switch articles when I'm just trying to scroll down.


It also breaks saving URLs in evernote mobile!

Down with AMP!


Is it just my memory or do you have to manually thumb your way to the top of a very long page in order to press said X button?


Maybe they can put it as a big round button in the bottom-right so it doesn't break with Material Design too.


It also frequently crashes the browser when zooming in on something on the page.


> it has different scrolling

Actually, AMP pages had the most consistent scrolling with the rest of the system. It was 'regular' web pages that had the different scrolling

iOS 11 fixes that. All pages now scroll at the same 'speed' as the rest of the system.

More

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: