Live data from Hacker News

Google Is Not God of the Web

bilge.world

201–210 of 306 posts

Re: Google Is Not God of the Web

#201

Not sure what this article is arguing. Sometimes you want to make a slow website that doesn't fit well on a phone screen? Leaving aside the fact that you can of course do that, and that if I'm using a search engine on my phone I probably (usually?) don't want to look at your slow site that I have to horizontally scroll... > Modern web design principles are very rarely directed at regular people looking to make a webs…

They’re arguing that the whole idea behind the web is that different people can try different things with different priorities and constraints.

The priorities shouldn’t be dictated by a single company.

I don’t fault Google for what they are doing (besides the AMP stuff). The problem the author is talking about is largely because google is the de facto way pretty much the entire world finds new stuff.

If website discovery was driven by blofs and the like wns people then added the stuff they liked to their RSS readers then the authors issue would be resolved.

Unfortunately that world doesn’t exist anymore and Google’s preferences end up dictating what all websites prioritize.

Re: Google Is Not God of the Web

#202
post #93

> 22.82 Mbps will reliably download very complex web pages nearly instantaneously. The author may be unaware of how ridiculously huge web pages have gotten. I just loaded Wired.com, scrolled down and let it sit for a few seconds. It downloaded 96.2 MB, requiring over 33 seconds on one of those average connections. On a pay-as-you-go data plan, it would have cost about a dollar just to load that one page. The front pa…

Even with adblocking it downloads 30 MB almost instantly, and while typing this reply it's now up to 48 MB. Oh wait, it's at 50 now. It keeps going up.

Looking at the requests it seems to be downloading a video from Cloudfront. And yes, in the middle of the page there is a video playing. I'm sure people with metered connections will love that.

That said, with adblocking at least the design looks clean enough. I'm willing to bet that this is what their designers see, and then another team adds all ad overlays on top of the existing design.

Re: Google Is Not God of the Web

#203
(Disclaimer: previously worked at Google search)

I think some commenters are attributing to Google an ulterior motive, whether ill- or good-intentioned, separate from its core business. But in this case no such motivation is necessary.

Basically, Google wants its users to be satisfied - otherwise it will lose to, say, Bing. So it measures user satisfaction - e.g., if a user clicks on a Google result, and immediately hits back button in three seconds, it's a strong signal that the user was not satisfied. And Google tries very hard to increase this "user satisfaction" (and other similar metrics), because not only does it help Google's business, but it also improves the service itself.

And, guess what? When a page takes fifteen seconds to load, lots of people hits the back (or close) button. Showing such a page is giving the user a bad experience. Unless there's literally no alternatives, it makes sense for Google to "penalize" such a page.

Of course no metric is perfect, so it will occasionally backfire and penalize a great page that takes thirty seconds to load. But that's life.

Re: Google Is Not God of the Web

#204
post #93

> 22.82 Mbps will reliably download very complex web pages nearly instantaneously. The author may be unaware of how ridiculously huge web pages have gotten. I just loaded Wired.com, scrolled down and let it sit for a few seconds. It downloaded 96.2 MB, requiring over 33 seconds on one of those average connections. On a pay-as-you-go data plan, it would have cost about a dollar just to load that one page. The front pa…

>Web developers have utterly squandered all efficiency gains of the last 30 years, and

Loading Wired.com with uBlock and no javascript the page comes in below 1.5MB for me, with most functionality seemingly intact (in that the front page looks mostly normal and I can load articles that appear to have their text completely intact). The bulk of that seems to be fonts and images, which are probably unavoidable for a media site.

Some reasonable noscript whitelisting for Wired.com and a few others (out of 12 total that noscript shows me) gives a page size that's still under 5MB.

Looking at the full page with everything loaded and unblocked, the biggest offender here seems to be not web design, but an aggressively downloading autoplay video on the front page. Without that the page itself is - while not necessarily great - at least reasonably bearable.

Truth be told, I'd started this post intending to blame advertisers, and there is still some merit to that since even before the video kicks in the various third party scripts balloon the page size several times over from the minimal functional one that loads with everything blocked. But in this case, it does simply seem to be a wasteful media stream playing without regard to whether anyone wants it to or not.

Re: Google Is Not God of the Web

#205
post #93

> 22.82 Mbps will reliably download very complex web pages nearly instantaneously. The author may be unaware of how ridiculously huge web pages have gotten. I just loaded Wired.com, scrolled down and let it sit for a few seconds. It downloaded 96.2 MB, requiring over 33 seconds on one of those average connections. On a pay-as-you-go data plan, it would have cost about a dollar just to load that one page. The front pa…

"Google should be applauded if they make these wasteful developers suffer monetarily until they shape up." As the browser author with the dominant market share, Google could start by not enabling these "wasteful developers". These ridiculously huge web pages can lock up a user's computer when the user's browser is a "modern" and "supported" one but not when the user agent is something simpler. I can read wired.com re…

> Trying to get web developers to change what they design to only use a small fraction of what Chrome and the user's computer can do is far more difficult, if not outright impossible.

This is not only improbable to happen that web developers will change their malicious behaviour, it's also the user agent's fault for allowing that.

Why is there no connection speed detection in the browser? Why does the browser allow media playback by default? Why is there no mechanism that reflects the expectations of the user? Is the user expecting videos on the news website or just to read text and images?

I personally think that user agents are not really user agents anymore, as there's not even the idea of offering the user a choice like this.

And personally, I do not agree with the concept of trusting random websites on the internet - by default. Any website on the web should be distrusted by default, with the user having the choice on what to load and what to expect when navigating to that specific resource.

If I visit an i.imgur.com/somename.jpg, why am I redirected to an html page with megabytes of files just because the user agent accepts html then? Should this not be outright impossible?

But please take my comment with a grain of salt, I am building a web browser concept that allows filtering and upgrading the web (which is superhard to build) [1] and it's still a ton of work to get anything working in a stable manner.

[1] https://github.com/cookiengineer/stealth

Re: Google Is Not God of the Web

#206
post #123

Earlier quoted context omitted.

They tried this, I used it, no big adoption, product got killed. To which service do you refer?

I believe it was Google Contributor: https://en.m.wikipedia.org/wiki/Google_Contributor

Yes, that is the one, thank you

Re: Google Is Not God of the Web

#207
post #133

Earlier quoted context omitted.

Exactly the way it looks today, because most people aren't willing to pay for content.

Got a source for that absolutely impossible to source claim? People pay for content all. The. Time.

NY times has 3.5M digital subscribers. Netflix has 120M. There are 4.5 billion internet users.

Re: Google Is Not God of the Web

#208
Didn't expect something as silly as this, especially the assertion that page sizes don't matter and since internet speeds are going up, web people are free to use it.

I don't see internet costs going down per MiB. Not everyone is on an unmetered connection. And users expect the large amount of bandwidth to be utilized for Netflix or something like that.

And how do you say Google is being evil for doing this? Even if there was no search monopoly and imagine there were two competing search engines. I think both of them would factor page load times in search results for better user experience.

Re: Google Is Not God of the Web

#209
post #93

> 22.82 Mbps will reliably download very complex web pages nearly instantaneously. The author may be unaware of how ridiculously huge web pages have gotten. I just loaded Wired.com, scrolled down and let it sit for a few seconds. It downloaded 96.2 MB, requiring over 33 seconds on one of those average connections. On a pay-as-you-go data plan, it would have cost about a dollar just to load that one page. The front pa…

My browsing these days looks like

- JS off by default

- Web fonts off by default

- Media elements larger than 100KiB not loaded by default

uBlock origin is the god of internet.

Edit(s): formatting

Re: Google Is Not God of the Web

#210

Earlier quoted context omitted.

"Google should be applauded if they make these wasteful developers suffer monetarily until they shape up." As the browser author with the dominant market share, Google could start by not enabling these "wasteful developers". These ridiculously huge web pages can lock up a user's computer when the user's browser is a "modern" and "supported" one but not when the user agent is something simpler. I can read wired.com re…

> Trying to get web developers to change what they design to only use a small fraction of what Chrome and the user's computer can do is far more difficult, if not outright impossible. This is not only improbable to happen that web developers will change their malicious behaviour, it's also the user agent's fault for allowing that. Why is there no connection speed detection in the browser? Why does the browser allow m…

All those things you mention are design choices.

Perhaps one of the impediments to the development of new user agents is a feeling that they must be complex and meet some "standard" of features. A standard that is nigh impossible to meet for the average programmer. On top of that, web developers demand the ability to create complex web pages full of executable code.

However we have no proof that users would reject a cornucopia of different agents that did not all have the same set of features. User agents do not need to be designed to satisy web developers. User agents can be designed to satisfy users.

They can be designed to do one thing well instead of do everything.

No user agent need be intended to "replace" any other, and certainly not to replace "modern" browsers. The intent is to create choice and user experimentation.

It is still possible to access the web with simple programs. It is not gopher or gemini but it still can work in a relatively simple way. Web developers probably do not like that but it remains true. The complexity of today's web sites is optional. It is a deisgn choice. Made possible by... "modern" browsers.

Godspeed.

Post reply on HN