Slashdot Mirror
'Why I Decided To Disable AMP On My Site' (alexkras.com)
Web developer Alex Kras on Monday listed a number of reasons why he dislikes Google's AMP project, and why he pulled support for it on his website. From his post: Back in the day we used to have WAP pages -- specific web pages that were presented only to mobile devices. Opting into AMP, for publishers, is kind of like going back to those days. Instead of using responsive design (making sure that one version of the site works well on all devices) publishers are forced to maintain two versions of each page -- their regular version for larger devices and mobile phones that don't use Google and the AMP version. The benefit of AMP is that it imposes tough restrictions on content, making it load fast. The issue with this approach is that AMP becomes a subset of the original content. For example, user comments are often removed. I also find the way images load in AMP to be buggy. AMP tries to load an image only when it becomes visible to the user, rendering a white square instead of the image. In my experience I've seen it fail fairly regularly, leaving the article with an empty white square instead of the image. [...] It's up to publishers to decide if they want to add AMP support on their site. Users, however, don't have an option to turn AMP off. It would be nice if Google provided a user level setting to turn results rendered as AMP off. Unfortunately, even if they were to add this option, it wouldn't help much when Twitter of Facebook would decide to server AMP. Further reading: Kill Google AMP before it KILLS the web - The Register, The Problem With Google AMP, 2 Billion Pages On Web Now Use Google's AMP, Pages Now Load Twice As Fast. John Gruber on open web: Fuck Facebook.
Now if only every other web developer in the world followed suit, and also abandoned the numerous other methods of forcing browsers into crappy "mobile site" ghettos, instead of designing their pages properly, the world would be a better place.
The vast majority of websites become crippled when I browse their mobile version, and I am talking about those "responsive designs" (which the summary seems to indicate they are the "good option"), not just AMP. Even on my 5" phone - i.e. a prime target for "mobile web" content - I usually have to switch to the regular website to retain functionality that I consider essential (but the designers apparently do not). I don't mind having to pan & zoom a bit when everything I need is right there on the page - the only difference is that I use landscape mode.
And that includes slashdot...
Violence is the last refuge of the incompetent. Polar Scope Align for iOS
No day goes by without someone complaining about AMP. This should tell you all you need to know about this technology.
Don't Most mobile browsers allow this now? I know of a LOT of pages with crappy mobile versions. Youtube being one of them. I just go into my page settings and select "Request Desktop Site" and TADA! All is well...
so what if it doesn't compact to remove 1/2 of the info I want to read. Pinch and zoom bitches ...
by TheSpoom (715771) Uncaring Linux user here. I have nothing to add to this but please continue. *munches popcorn*
motherfuckingwebsite.com
And in the same way they block indexing by search engines, Facebook forbids The Internet Archive from saving copies of posts.
That's not bad thing though; nothing of value is lost and the storage space could be used for more valuable stuff.
Ezekiel 23:20
designing their pages properly
Then what's the "proper" way to accommodate the fat fingers of users of touch screen devices without needing excessive scrolling for users with more precise pointing devices? Adding padding around links improves usability on touch screens but increases the scrolling for mouse or trackpad users. And controlling the padding with a CSS media query breakpoint based on viewport width doesn't help because a touch-operated iPad held in landscape orientation has more pixels than a trackpad-operated netbook.
My biggest gripe with AMP is that it breaks same-page-searching. If you have a specific phrase you're searching for on a very long article, it just doesn't work well because all the search results still seem to be rendered underneath the article, and same-page searching seems to go through each of those first, and sometimes still can't find text after that on the AMP page. When I'm looking something up on mobile, I often just want to find a something quick, not read an entire article.
I don't whether AMP is responsible, but I visited a site about a week ago with progressive content and image load as I paged down. This is annoying, but nothing new.
Since I wanted to CTRL-F to search within the page, I spent 5 s manually pressing PG-DOWN to fully load the page.
Imagine my horror when I discovered that most of the top of the page—previously loaded already—had now disappeared from my document, and was doing progressive load on the way back up.
That wasn't just irritating. That was outright /etc/hosts-level hellban territory.
Please, for the love of God, look upon my 16 DIMM slots ye Mighty frugal HTTP server, and load the whole damn document all at once, SVP.
The point of AMP is to restrict what you can do so that pages load faster. And as a side effect, Google gets to make the rules.
But AMP would never have existed if webmasters were a bit more reasonable. We are talking about sites weighting several megabytes for the equivalent of a single blog post, with scripts creatively breaking browsers. With a 2.3MB / 66 object webpage, the author is definitively guilty of this.
How about, instead of complaining about AMP, take the core of the technology, which is actually quite good, remove the Google bits, and make your site even faster and lighter.
The DEVICE knows if it is using a touchscreen or a mouse. It is therefore up to the device to render properly, and the web designer should simply not get in the way of this. I.e. don't think you can force a particular layout so don't even try.
The answer was in the article. Read it sometime.
And yet that's not what the article is about Troll....
Please, for the love of God, look upon my 16 DIMM slots ye Mighty frugal HTTP server, and load the whole damn document all at once, SVP.
Even a PC with double digit GB of RAM can be connected to a satellite or cellular upstream connection whose ISP charges $5 to $10 per GB. Though a non-AMP page like this still loads fast because it's so simple, I imagine people aren't going to be happy to pay the ISP to load images that won't be viewed.
@media
APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
It's unclear to which part of the article you are referring. I searched inside the document "I decided to disable AMP on my site" by Alex Kras for the words "touch" and "finger", and neither word was there. I searched for the word "link", and none of the results mentioned adapting link size based on whether the browser uses touch or mouse. Could you quote a sentence from the relevant portion?
And controlling the padding with a CSS media query breakpoint based on viewport width doesn't help because a touch-operated iPad held in landscape orientation has more pixels than a trackpad-operated netbook.
@media
Could you be more specific as to which CSS media query expression you refer?
Expressions using min-width and max-width specified in pixels are a good starting point, as those are compared to the Viewport size reported by the browser. Where you go from there depends entirely on your design needs
All iPad models present a 1024x768 viewport in landscape mode, regardless of how many actual pixels there are. If your content is so crammed together that you can't tap a link with a fat finger on a 7" screen, you need to work on your design a bit.
The shittiest Chromebook I can find presents a 1280x720 viewport. Despite having physically fewer pixels, it presents a larger horizontal viewport dimension than any model of iPad so, yes, you can use media queries just fine. You just have to understand how they work.
For more information on viewport sizes, see here: http://viewportsizes.com/
And if you really can't make it work, consider also using min-width and max-width expressions based on physical size in addition to pixels, so you're measuring physical and virtual capabilities of the device. Even Windows can work accurately with those now.
APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
My smartphone has enough resolution and the ability to pan / scan / zoom over all of your content.
Then what's the proper way to serve the majority of users, who prefer to view documents that are already at a reasonable zoom level when they first load? Search engines behave similarly: Google is known to penalize documents that aren't already at a reasonable zoom level when they first load.
All iPad models present a 1024x768 viewport in landscape mode
So did my first couple PCs (an Acer TravelMate laptop and a Dell Dimension desktop), which had a 1024x768 pixel display and a mouse. My netbook (a Dell Inspiron mini 1012) also has a 1024-pixel-wide viewport because its display is 1024x600 pixels, but it has a trackpad. A full-size PC is likely to have a 960-pixel-wide viewport and a mouse when the user "snaps" a browser window to half of a 1920x1080 pixel screen.
The shittiest Chromebook I can find presents a 1280x720 viewport.
According to the site you linked, the Nexus 10 tablet in landscape orientation also presents as having a 1280-pixel-wide viewport.
Fine-point styli for touch-operated devices still under warranty, such as the Apple Pencil and Samsung S Pen, aren't widespread among most websites' audiences. And a stylus for capacitive touch screens is nearly as imprecise as the finger it's designed to emulate.
When I browse on my phone, I just zoom in if a link is too hard to tap.
You are also aware of the zoom feature. Not all sites' audiences are as technical as that of Slashdot. In addition, the mobile view of Google Search penalizes sites that initially load with text too small or links too close together.
Because its not supposed to be up to the page what the size of links are. Its supposed to be up to the browser. That was the original intent of HTML, of the Web itself.
br. Disabling AMP doesnt solve the problem that browsers arent reformatting automatically, nor can a web page developer solve that problem. The browser authors need to step up and stop trying to be the next PDF.
"His name was James Damore."
... Gruber has written specifically about AMP. https://daringfireball.net/lin...
If you are a publisher and your web pages don't load fast, the sane solution is to fix your fucking website so that pages load fast, not to throw your hands up in the air and implement AMP.
He has written more about it in the past -- links are in that piece.
Dear Slashdot: next time you want to mess with the site, add a rich-text editor for comments.
Without advertisements, how should a site's operator pay its writers, its server operators, and its bandwidth bill? Paywalls don't work for sites that rely on traffic from search engines. What's the third way, besides ads and paywalls, to fund a site that is larger in scope than a hobby?
And if you really can't make it work, consider also using min-width and max-width expressions based on physical size in addition to pixels, so you're measuring physical and virtual capabilities of the device. Even Windows can work accurately with those now.
APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
Hell, you know what?
@media (pointer:coarse)
Enjoy.
APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
And if you really can't make it work, consider also using min-width and max-width expressions based on physical size in addition to pixels, so you're measuring physical and virtual capabilities of the device. Even Windows can work accurately with those now.
I'm failing to understand how high DPI implies pointer precision, unless you're recommending targeting specific models of Apple kit. Laptops and Android tablets vary so much in DPI that the high end of tablets, which lack a trackpad, is likely to overlap the low end of laptops, which have a trackpad.
Neither IE nor Firefox supports pointer:coarse.
If your content is so crammed together that you can't tap a link with a fat finger on a 7" screen, you need to work on your design a bit.
Phones will have smaller screens and are dead simple to detect with media queries.
APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
What if it is a touchscreen device with mouse support, what then?
It goes based on whether a mouse happens to be connected. This can cause documents using the @media (pointer:coarse) media query to change styles when a mouse is connected or disconnected.
Did you just assume the size of my screen and the size of my pointing device? Stop. All bets are off on both.
Furries make the internet go.
It's ''ad nauseam'', you fucking peasant.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Did you just assume the size of my screen and the size of my pointing device?
Until web browsers support CSS4 media queries, web browsers have to use CSS3 media queries and assume that any device with a viewport narrower than 26em or so has a touch screen, which is a coarse pointer.
So, if a wireless mouse (or maybe pen?) is connected, but not being used, it'll give the targets too small for a finger?
Comment removed based on user account deletion
Mandatory Javascript (mandatory loaded from google cdn) for a restricted subset of HTML?
W.T.F.
1) Define a Subset of HTML
2) Create a Parser
3) Validate the page. When it's AMP display it with your fast parser, if not open a browser
4) ???
5) PROFIT
Why does it need javascript?! Javascript should be FORBIDDEN on "fast" pages.
Browsers need to help there. Chrome does a very good job. firefox not that good. But the worst are websites, which disable the zoomin on purpose, because they know better what fontsize is good for me.
your browser can let you choose.
Did you know, that firefox allows you to use alternative stylesheets in the "view" menu? when a site provides a stylesheet with rel="alternative", you can choose it there. And old mozilla had button navigation for rel="next", rel="previous" links. Browsers had nice features even before html5 ...
"AMP is for ARTICLES, not SITES."
ARTICLES ARE FOR MAGAZINES, NOT WEB SITES.
But I guess you failed to understand that since you've probably never picked up a piece of dead wood in your life.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
Sorry but AMP sucks even for articles.
A good question to ask, if you're generating content, is that if the only way to pay for it is by embedding advertising, then the content is worth nothing. Thus, if you do want to broadcast it to the world for some reason, hobby or otherwise, you're going to have to find the money yourself. If the advertising industry disappeared overnight, and all the content that it funded did so too, nothing of value would be lost.
Host ads and scripts on YOUR servers (or rented space). It is the connections outside your domain that cause lag that is beyond your control.
227-3517
Or as Gruber said before... you could just "fix your fucking web site".
The only people that AMP benefits, is Google. It's a lock-in tool, plain and simple.
Google values page speed very highly and that value is only increasing. They're pushing AMP landing pages and converting ads to AMP too. As a website owner, when Google gives priority to those that have AMP pages over those that don't, it's pretty hard to choose not to offer AMP. Look at the sites that chose not to become mobile-friendly back in 2014. They saw significant drop in rank and traffic. Most won't throw away that traffic just because they don't like everything about AMP. Claim maintaining AMP pages is maintaining 2 versions if you like but for most it's done automatically. Few will have to manually maintain their responsive site and AMP site. WordPress runs a huge percentage of all websites and offers a WordPress-made plugin to automatically generate AMP pages, along with other plugins offering the same, plus native implementation is on the way in future versions. The author claims comments can't be included on his site in the AMP version. AMP most certainly supports comments. His implementation just doesn't. Additionally, the vast majority of website content doesn't contain comments, so it's not a big issue for most, even if it was true that AMP didn't support it.
if I did have that problem and thought it due to lines being too long, I could just resize the browser window and the text will wrap.
A max-width on the body text column does pretty much the same thing. Putting max-width in the CSS rather than waiting for the user to "resize the browser window" does the right thing by default for the majority of people, who do have this problem.
CSS4 media queries anticipate that multiple pointing devices may be connected at once, which is why it defines a concept of "primary pointer" used to evaluate the pointer media property. If the user wants to use the touch screen, he should tell the browser to treat touch as the primary pointer.
In theory, I'm inclined to agree that hosting the ads on the publisher's server is probably the most efficient for network data volume, CPU time, and user privacy. But in practice, a publisher selling ad space directly to advertisers faces a few additional difficulties compared to the more common route of going through a broker, such as an ad network or an ad exchange.
I'm interested to read the solutions that you would apply to these problems if you were running an ad-supported website.