8 ms·
I don’t even understand how AMP is legal or how they could possibly argue it gives any benefit to the user. Every single AMP page i’ve used has broken CSS and o
by peanut_worm 6y ago
I don’t even understand how AMP is legal or how they could possibly argue it gives any benefit to the user. Every single AMP page i’ve used has broken CSS and often broken content.
The speed boost is negligible if it even exists and all it serves to do is add an additional stupid pop up I have to click out of to read a web page.
I really wish I could switch to DDG but queries related to anything technical, like biology or programming, usually fail to turn up any relevant results.
- spijdar 6y agoFor a contrasting anecdote, when I'm traveling abroad and sometimes get stuck with 20-50 kb/s transfer speeds, AMP is the difference between waiting an incessant amount of time and getting pretty instantaneous loading. On fast internet? Yeah, negligible. On non US/European internet, though, it's very tangible. Just my experience, though.
- peanut_worm 6y agoWhat websites are you noticing better performance? I usually encounter AMP on reddit and, at least on my phone, it seems to make the page load even slower.
- spijdar 6y agoOn basically anything that uses it. I'll give an example from the google feed on my phone, now. Some article on zdnet about RedHat discontinuing CentOS in favor of stream. If I open the AMP link [0] in Firefox and measure the bandwidth used, it's around 685 kilobytes transferred in 1.32 seconds. The original article transfers almost 3.6 megabytes over 3.89 seconds, with the page taking almost 9 seconds total to load in, vs only 1.43 seconds total on the AMP page. This is on my residential internet on an 8 core, 32 thread POWER9 workstation. My internet isn't great, but it's a heck of a lot better than the 15-40 kb/s I got overseas. There, that full page would take almost 2 whole minutes to load, versus only 20 seconds. Obviously, there's more to this -- the non-AMP page loads its text before the other elements, so it's not like you'd have to wait the entire 2 minutes to begin reading. And beyond the technology, there are other reasons to not use AMP. But suggesting we shouldn't use AMP because it's bad technology or broken isn't really telling the whole story. There are plenty of people who will get much better experiences with AMP than whatever fat pages would otherwise get shoved their way. We shouldn't become dependent on google for this, but we also shouldn't pretend like there isn't a need for AMP, at least for some people. [0] https://www-zdnet-com.cdn.ampproject.org/v/s/www.zdnet.com/google-amp/article/why-red-hat-dumped-centos-for-centos-stream/?amp_gsa=1&_js_v=a6&usqp=mq331AQFKAGwASA%3D#ampshare=https%3A%2F%2Fwww.zdnet.com%2Farticle%2Fwhy-red-hat-dumped-centos-for-centos-stream%2F https://www-zdnet-com.cdn.ampproject.org/v/s/www.zdnet.com/g... [1] https://www.zdnet.com/article/why-red-hat-dumped-centos-for-centos-stream/ https://www.zdnet.com/article/why-red-hat-dumped-centos-for-...
- lern_too_spel 6y agoReddit doesn't have its own AMP cache, so you won't get any prerendering benefit.
- asddubs 6y agoi use old.reddit.com on my phone, since ironically the mobile version is too bloated to ever finish loading before I give up on my cheap mobile internet.
- d3nj4l 6y agoUse i.reddit.com, or add .compact to the end of the link. It's old reddit but with a mobile (and touch) friendly layout.
- Griffinsauce 6y ago"non US/European internet" Do you want to retry that description?
- crazygringo 6y agoI've literally never seen broken CSS or content. I'm assuming you must be running some kind of extension or blocker that is interfering? There are plenty of criticisms of AMP but failing to load content isn't really one of them. Also the speed boost is massive. AMP pages usually show content instantaneously due to precaching, while a new site takes 5-7 seconds to load.
- mrtranscendence 6y agoThat’s not my experience at all, at least on iOS. I notice essentially no performance improvement when accessing AMP vs when I inevitably choose to load the non-AMP site because AMP sucks terribly.
- lern_too_spel 6y agoIt's more likely that mobile Safari sucks terribly. On Android, loading an AMP page from a Google or Bing results page is instant.
- crazygringo 6y agoSo which is your experience? I'm quite curious. Is it that AMP sites also take 5-7 seconds to load? Or that native sites load instantaneously?
- ninkendo 6y agoThere’s a really easy answer here: you’re all talking past each other. AMP isn’t just a standard for “fast webpages”. It’s also a technique google uses to embed other pages in their search results, such that: - Google rehosts the result on google.com (with the same HTTP2 connection you already have open to their server) - Sites are eagerly loaded in the background even if you don’t click them (making them appear instantaneous) There’s a difference between “I was sent a link to a site that uses AMP and I opened it” (which may be slow) and “I came across a site while searching on google” (which is nearly always fast.) AMP-the-standard gives google the guarantees necessary to enable hosting your site inside the google search results. This is why it’s perceived as fast. If you’re connecting to the original site itself, outside the context of a google search (or some equivalent... I know Facebook does similar re-hosting) you’re not going to see as much of a speed gain. (Side note: if AMP-the-standard was all that it was and the speed gains only came from having less bloated JS and requiring all on-screen elements to have known sizes to prevent pop-in, I’d love it. It’s what google does to it in its search results that I hate.)
- nr2x 6y agoI'm curious what types of programming questions you can't get via DDG? I use it exclusively and very rarely have to look to Google (as in <1x month). Really the only time I revert to Google is if the issue is very fresh (eg some new iOS bug from the past 24hr). Also, you can prepend search with "!so" on DDG and go straight to Stack Overflow.
- coldpie 6y agoIt's pretty bad at stuff like symbols. Searching "c ++ operator" in DDG returns a bunch of "operators in C" results like [1]. On Google, the first two results are pages about increment and decrement operators. [1] https://en.wikipedia.org/wiki/Operators_in_C_and_C%2B%2B https://en.wikipedia.org/wiki/Operators_in_C_and_C%2B%2B
- nr2x 6y agoThanks, that's a helpful example, may be most of my programming questions don't involve C++. ;-)
- kovac 6y agoI can see how that can be confusing :p On Google I got the increment operator as the top result (though not specifically for C language). On DDG I got an equivalent result for fifth result. Second search result on Google which is the increment operator for C++ in C++ reference was probably the best result? Personally for me, when I search something like this I tend to lookup the language reference directly and find the operator I want there since it's a very specific query with a very specific answer. But yeah, ddg, ecosia both fall short on this query.
- OJFord 6y agoI've used it exclusively for years without once looking to Google, which may well be better (I honestly wouldn't know) but DDG isn't a problem.
- stetrain 6y agoWhy wouldn't it be legal? Every site that uses it has to specifically add it and opt-in. Sites are willingly implementing AMP pages and telling Google etc. that the page exists and can be cached/served as an alternative to the original page link.
- tristan957 6y agoGoogle was pushing sites that used AMP higher in search results. They are currently facing lawsuits in court in both the US and EU I believe for anti-competitive practices, AMP being one of them. * Google no longer treats AMP pages as special as of a few weeks ago.
- stetrain 6y agoYes, the favoritism in search results for sites using your own tech could definitely be an antitrust issue.
- mkl 6y agoIt looks like Google still treats AMP pages as special, and will for a few more months, as announced a few weeks ago: https://developers.google.com/search/blog/2020/11/timing-for-page-experience https://developers.google.com/search/blog/2020/11/timing-for... https://searchengineland.com/amp-wont-be-required-for-googles-top-stories-section-335276 https://searchengineland.com/amp-wont-be-required-for-google... https://themarkup.org/google-the-giant/2020/11/19/as-antitrust-pressure-mounts-google-to-pull-back-benefit-to-news-sites-that-adopted-its-preferred-mobile-technology https://themarkup.org/google-the-giant/2020/11/19/as-antitru...
- creato 6y agoI get that there are philosophical issues with AMP and respect and agree with many of those, but I don't understand how people claim with a straight face that it's worse from an end user perspective. Even setting aside load speed, it fixes my top two pet peeves: - Content moving around for a few seconds after the page loads (and of course right when I go to click on something). - No auto-playing audio.
- anonymoushn 6y agoIt's way worse. A prominent American news outlet's AMP pages have Google ads obscuring the text of the article. In the last week AMP added an additional click before you can get the actual URL of the article to be able to read the text. The additional click takes the form of a fake sharing menu in the browser. It's sort of understandable that they would want to add this given that Android's built-in sharing menu takes multiple seconds to load on Google's flagship phones, but the fake menu lacks "copy to clipboard" so it cannot be used to go to the actual article.