Google Chrome To Identify and Label Slow Websites (techcrunch.com) 53
Is it the web page that's slow or is it your network connection? In the future, Google's Chrome web browser may have an answer for you. From a report: Google announced today a plan to identify and label websites that typically load slowly by way of clear badging. The company says it may later choose to identify sites that are likely to be slow based on the user's device and current network conditions, as well. Google hasn't yet determined how exactly the slow websites will be labeled, but says it may experiment with different options to see which makes the most sense. For example, a slow-loading website may show a "Loading..." page that includes a warning, like a caution icon and text that reads "usually loads slow." Meanwhile, a fast website may display a green progress indicator bar at the top of the page instead of a blue one. And for links, Chrome may use the context menu to help users know if the site will be slow so you can decide whether or not you want to click.
Pot, meet kettle (Score:5, Insightful)
Is Google going to differentiate between sites hosted on slow hosts versus sites which take forever to load because of all the Google ad trackers and analytics tags included on each page? Cuz I see a lot more of the latter than I do of the former.
Re: Pot, meet kettle (Score:2)
Re:Pot, meet kettle (Score:4, Insightful)
You mean like: This website downloaded the requested page in 35 milliseconds but took 1450 milliseconds to download all the attached JavaScript. It then took a further 5640 milliseconds to execute all that JavaScript and render a page suitable for viewing. This is an inefficiency rating of 0.04%. This web site is slow.
Re: (Score:2)
If you're the site owner, please call this number to talk to a Google account manager about hosting it via AMP and removing this scary warning for all of your customers.
Dat's a nice web site youse got there. Be a real shame iff'n customers was to be scared away from it. Luckily I'm about to make you an offer you can't refuse to ensure that doesn't happen...
<script defer> (Score:4, Informative)
Is Google going to differentiate between sites hosted on slow hosts versus sites which take forever to load because of all the Google ad trackers and analytics tags included on each page?
Properly engineered analytics and adtech scripts load with <script defer> [flaviocopes.com]. This means they run after the HTML has been parsed and the DOM created, and they may delay much of their work until after the "content" above the fold (CSS, images, and fonts) has been loaded and painted. Such scripts, which don't increase time to painting and interactivity, probably won't get counted.
Even though deferred heavy scripts consume battery charge and download quota, this is still better than the previous metric of "AMP good, non-AMP bad".
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
I know for a fact my ISP isnt throttling, not even overseas torrents (get the picture? does that frame it for your tiny mind?)
Re: (Score:2)
Is Google going to differentiate between sites hosted on slow hosts versus sites which take forever to load because of all the Google ad trackers and analytics tags included on each page?
The Google trackers are loaded asynchronously and don't consume much bandwidth. If your website is slow because of Google's trackers then you should slap your webmaster. There are however plenty of shitty tracking javascripts, quite often home made that are horrendous, or javascripts that loads entire libraries from a 3rd party just to display a blink tag.
I'm thinking the point is to identify precisely these.
AMP Move (Score:2)
It's not going to count Google's ads/analytics against a page that uses their proprietary AMP format. This is just a continuation of that fight, they continue to argue HTML is slow (and open-source) and AMP is fast (and Google controlled). If you want the Web to be like Go (they ask for free help, but maintain dictatorial control over the language), Google has you covered.
Re: (Score:2)
And if anyone mentions async and defer they can go to hell. That only solves the solvable problem of javascript blocking page rendering. If your browser is getting chunky from analytics.js loading more scripts that possibly load more scripts then it doesn't matter if the page rendered fast when your scroll wheel isn't scrolling.
Get a faster computer/phone!
No! I'd rather block your crap.
And besides, I'd like to know how google plan to measure this, because I'm looking at my "page speed warnings" in their t
Re: (Score:2)
Google has proposed a fix for that and is implementing it in Chrome and on their own sites.
A new tag allows the browser to cache the tracking ping-back and send it later, up to a few days later in fact. The timing is randomized and the amount of tracking data limited so that it can identify ad impressions but is of little use for tracking users, so it's a privacy enhancement too.
I know, mind blown right? Apple has a similar but incompatible mechanism that it is pushing.
Slow Websites or not-snappy websites? (Score:4, Interesting)
If it is something that you have to wait 30 seconds and get a few kilobytes back then that is slow. if you wait 30 seconds and get 50 megs back then it is just not snappy. But it is actually running fairly fast.
Sometimes you need to sacrifice UI response speed, just so you can provider the bulk of the data.
Are you displaying 50 million lines? (Score:5, Insightful)
"But there's a lot of data!"
Have you ever noticed when you do a Google search at the top of the results it says something like "3.4 million results in 0.28 seconds"?
> you wait 30 seconds and get 50 megs back
If you're sending back 50 MB (50 million characters) , it sounds like you're either:
A. Displaying a million lines of results
B. Downloading a file to the user's computer
C. Downloading your database into a JavaScript object and then running the application in the browser via JavaScript instead of having a web application
A. Is stupid and slow
B. Is okay
C. Is stupid and slow
Thirty seconds isn't a long time to wait to get a million lines of results displayed - displaying a million lines of results is slow, and it's stupid.
I don't know your app, but what you said sounds like what people say when the problem isn't the technical details of HOW they did it, the problem is WHAT they decided to do is silly.
Re: (Score:3)
Re: (Score:2)
Browsers should be caching things like that.
Thus, breaking the entire point of them being centrally hosted.
Re: (Score:2)
Browsers should be caching things like that.
Only when it's requested over http:// - if it's https:// then the browser's not allowed to cache it. This is why "HTTPS Everywhere" sucks. Fast or secure, pick one.
This is completely false. HTTPS content can -- and should -- be cached by the browser exactly the same as HTTP content. In both cases, cacheability is determined by the Cache-Control header. Years ago it used to be the case that Firefox wouldn't cache unless "Cache-Control: public" was set, but they fixed that.
In fact, if the server sends "Cache-Control: immutable", most (all?) browsers will only cache if the content is delivered over an authenticated channel... which HTTPS is and HTTP is not. See htt [ietf.org]
Re: (Score:2)
FYI to anyone doing web hosting or development: With the default configuration for Apache on most OSes, Apache will not add "Expires" headers to HTTP replies, and most browsers will not cache anything without them. IIRC the module mod_headers needs to be enabled, and it needs to be set in .htaccess/httpd.conf/apache.conf.
If there is a web site that isn't getting cached by your browser, it's 90% likely due to misconfiguration on the server, and not the browser. Then again, I'm regularly amazed by the crap t
Re: (Score:2)
Downloaded once. Then they are loaded from cache on every site that uses them. Smart browsers even pre-compile them if they are not in WebAssembly already.
That's why those standard libraries are always loaded from the library developer's site rather than using a local copy. It's not to steal bandwidth, it's to allow the browser to cache it.
Re: (Score:2)
I am actually doing "B" or "A" depending on the user's request. I Prefer that they download the data as a file. When designing a tool for direct use I will get the site to run very snappily, faster than 1/4 second.
But the point is some sites run slow to meet a business objective. And I am afraid Google will be discrediting it for doing what it needed to do as quickly as possible
Re: (Score:2)
That's interesting. I'm slightly curious what the application is. But not that curious. :)
Obviously I don't know your application.
It might be interesting for you to watch the users as they do A and see if they don't immediately:
Ctrl-F the one record they actually wanted
Or
Copy-paste it into Excel
As you said, in the second case it makes more sense to just give them a CSV file and skip the copy-paste. I can't think of a use case where displaying a million lines at once is the best way to go - but maybe I'm
Re: (Score:2)
They should be only using the Web View when the Search Criteria gives them a small amount of actionable data. While Excel output for larger data sets. Its main design is for fast deployment of presentable reports of data. Where in about 30 minutes. I can have a request for a report to be built tested and deployed into production. Where other tools like SSRS or Crystal Reports are often too powerfu
Re: (Score:2)
What common web activity shovels 50 megs of data across the network?
Are you using CSI's fingerprint database that needs to display every wrong fingerprint?
Re: (Score:2)
Accountants hate summarized data. They want to see it all, so they can well account for every penny in the report.
Re: (Score:2)
Financial data is also pretty compressible.
Believe someone who does just that for a living.
Comment removed (Score:4, Funny)
Re: (Score:2)
How should the website learn that it needs to improve if you coddle it with soft language?
Re: (Score:1)
Darn it, meant to post it on YouTube thread.
Trivial (Score:5, Informative)
"Google hasn't yet determined how exactly the slow websites will be labeled"
Simple, does it use QUIC or other Google technologies? Yes? Then it's fast. No? Must be a competitor...I mean slow.
Re: (Score:3)
Well, to be fair to Google - they haven’t been able to drive AMP adoption to anywhere near the level they’d like... so it’s understandable that they’d need to come up with other ideas on how they can subvert those well-established open web standards which have Managed to survive for almost 30 years.
Re: (Score:3)
You are exactly right, but that hasn't stopped them from continuing to push on AMP in standards bodies despite how bad it is and how often is has been rejected. For a full report on their AMP "work", see this writeup [shkspr.mobi].
Re: (Score:2)
>"so itâ(TM)s understandable that theyâ(TM)d need to come up with other ideas on how they can subvert those well-established open web standards which have Managed to survive for almost 30 years."
Oh, just give it a bit more time. Since the only modern, multiplatform browser that is NOT controlled by Google is Firefox, I am sure they will come up with something that will ram their "standards" down everyone's throats.
Re: (Score:2)
Agreed. This sounds to me like just another way to try to force AMP on sites. "Nice site ya got here. Shame if someone was to label it... 'slow'." They tried the carrot, by giving AMP sites boosted ranking and a nice little AMP symbol. Now comes the stick, where everything not AMP gets a pejorative label.
And I'm guessing that maybe, just maybe, rendering the loading screen that says "Usually loads slow" is going to add just a little bit to the loading time and making the site feel slow whether or not it
Cease and Desist (Score:4, Interesting)
Re: (Score:2)
What fold? My monitor does not have a fold. How can something be loaded "below the fold" when there is no fold?
Re: (Score:1)
Re: (Score:2)
Guess you're using neither their cloud service nor their trackers.
So your site is slow. Can't you see?
uBlock plus uMatrix (Score:3, Informative)
uBlock + uMatrix, suddenly the whole internet is fast. All these useless scripts, autoplay videos, and incessant ads are ruining things for us.
bloat bloat bloat (Score:3)
Use AMP (Score:1)
Well, obviously this is intended to harm self hosted websites. But - anyone who turns over his infrastructure to AMP will never worry about it!
This is such an elegant way to be evil; I really have to take my hat off to it. Well played Google.
Re: (Score:1)
Well, obviously this is intended to harm self hosted websites. But - anyone who turns over his infrastructure to AMP will never worry about it!
This is such an elegant way to be evil; I really have to take my hat off to it. Well played Google.
Why is your "Score:0" right now? No visible negative mods, though your comment is the kind of target the "friends of google" like to mod down. My primary response to that form of censorship is to quote the target (to increase its visibility).
However I think the main censorship here is the AMP promotion, which is getting some play in the discussion. This is why I advocate a progressive profit tax based on market share. If the google (and all other abusers of their dominating positions) could only increase th
Will they use their webcrawler response time? (Score:3)
Re: (Score:2)
Don't talk to Lenny (Score:2)
He's a bit slow if you know what I mean.
Sounds like discrimination to me.
Well, that explains their labeling slow sites idea (Score:2)
How can you create a nearly instantaneous web experience in Chrome? Not by some new fancy technology. You just use your monopoly power to get the masses in line by labeling less than blazing sites as "slow" [slashdot.org]. That will help drive people to faster sites (thus creating a "faster Chrome experience"), all while forcing the labeled sites to optimize and remain relevant online.
It's a dick move, but it's probably going to be pretty effective to make Chrome appear faster.
Re: (Score:1)
This will practically be every site then (Score:2)