Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Verge's web sucks

blog.lmorchard.com

Post reply on HN
Print view View original

121–130 of 137 posts

Re: The Verge's web sucks

#121

Post by solidpy »

eevilspock wrote: post #93
gorhill wrote: post #58

Earlier quoted context omitted.

Default-deny is the best answer against bloat -- though not for everybody unfortunately, but I did try to make it as straightforward as possible with uBlock Origin[1]. Using default-deny in uBlock Origin/Chromium[2] and with the page displaying properly, I get from Network pane in dev console: - 62 requests - 508 KB transferred Once you start using default-deny mode, it's difficult to go back to anything less restric…

Raymond, how did Chris get to take over the original branch, and yours appears as a derivative work, when you are the originator? What the hell happened? I want to switch to uBlock Origin, but I currently use Safari. It uses the least amount of resources on my Mac compared to Chrome and Firefox.

He transfered the repo then made a fork.

Re: The Verge's web sucks

#122

Post by manigandham »

Terr_ wrote: post #119
manigandham wrote: post #76

Earlier quoted context omitted.

This is complicated because of how ad networks work and all the layers involved. We have somewhat of a standard called RTB (real-time bidding) where an exchange will see each impression and run an auction in 100ms for each ad slot and then take the winning bid and render it on the page. That was easy and efficient when it was just an image. Now ad formats have become more "engaging" and require javascript tags to run…

You might attack it from a non-technical perspective, where the "Decent Ads" certification says something about how their business practices are set up with SLAs or policies for who is at fault if an advertisement contains malware. Alternately, a technical benchmarking-only system, where you try to identify how much their contribution is to page-load without respect to security/etc.

Yes the IAB (Internet Advertising Bureau) and MRC (media ratings council) and 4A's (American Association of Advertising Agencies) all have various "standards" and certifications. There are similar groups in Europe and smaller orgs that focus on niche areas.

Unfortunately for the most part these groups all charge lots of money for a nice seal you can put on your site, without any real checking of a business. Some of the most offending and insecure networks are members of the IAB. [1] Not to mention this has no bearing on actual technical implementations. The MRC does audits of companies to certify them for measuring viewability but there can be 50% or more discrepancy[2] between vendors which ultimately makes the entire thing pointless. And buyers also don't care, they have their own "favorite" vendors they want to work with so they're not even interchangeable.

The adblock plus company in Germany (while I dont agree with their entire stance) has an "acceptable ads" standard that at least lays out something in terms of user experience. However they accept money for whitelisting which is why all those crummy clickbait widgets like outbrain/taboola still show up through the default ABP filters.

What I want to see is actual legal regulation that covers how a business deals with data/privacy and then some kind of technical certification led by a real tech-savvy company to vet ad networks and their implementations. However this industry is full of political motives and other shady dealings so it takes a herculean effort to do anything... Maybe with some consumer backing there might be changes given the wide usage of adblock today but (not) surprisingly most digital ad execs see it as some kind of plague rather than the direct consequences of industry actions. That's not a very cooperative attitude for enacting change.

[1] http://www.iab.net/member_center/1521/1534 [2] http://adexchanger.com/data-driven-thinking/viewability-lets...

Re: The Verge's web sucks

#123

Post by manigandham »

lmorchard wrote: post #106
manigandham wrote: post #76

Earlier quoted context omitted.

This is complicated because of how ad networks work and all the layers involved. We have somewhat of a standard called RTB (real-time bidding) where an exchange will see each impression and run an auction in 100ms for each ad slot and then take the winning bid and render it on the page. That was easy and efficient when it was just an image. Now ad formats have become more "engaging" and require javascript tags to run…

Thanks for all these details! My rant was largely based on guessing what's going on with all the stuff.

Of course. The more consumers know the better we can solve all this.

Honestly, I don't agree with the blocking all 3rd party cookies and adblock. It's a blunt-force approach that is harming more than helping right now, but I do agree with the reasons why it's happening and I'm hoping that it will finally change the situation.

Re: The Verge's web sucks

#124

Post by hajile »

JustSomeNobody wrote: post #37
mr_sturd wrote: post #28

Nilay Patel has the audacity to say that the mobile web sucks [0]. No. It's just your website that sucks, mate. [0] - https://www.theverge.com/2015/7/20/9002721/the-mobile-web-su...

He blames mobile browsers for at least half the article (I had to stop reading). Sure, mobile browsers may not be great, but web sites these days seem to be written by people who just don't give a shit. Their hipster attitude is ruining people's experience.

Mobile browsers are fine. The problem is that the A9 processors from a generation ago (and current low-end phones) have similar processing power to an old Pentium 4 processor (with less bandwidth and RAM). Put the same website behind something running an A57 and watch those intolerable numbers become reasonably acceptable for a >3W processor. Don't blame the browser, blame ARM.

Mobile processors are slow because Moore's Law is grinding to a halt. I think we should blame the process engineers, but the process engineers are being halted by the size of atoms and weird quantum mechanics issues.

Nilay, don't blame browsers, blame the universe.

(or alternately, blame your inefficient website).

Re: The Verge's web sucks

#125

Post by yellowapple »

krstck wrote: post #95
gorhill wrote: post #58

Earlier quoted context omitted.

Default-deny is the best answer against bloat -- though not for everybody unfortunately, but I did try to make it as straightforward as possible with uBlock Origin[1]. Using default-deny in uBlock Origin/Chromium[2] and with the page displaying properly, I get from Network pane in dev console: - 62 requests - 508 KB transferred Once you start using default-deny mode, it's difficult to go back to anything less restric…

The difference in page load times before and after implementing uBlock is staggering . Having to use a computer you don't own (or mobile) feels like traveling back to 1999 using 56k dial-up.

What's worse is that there are still plenty of places that haven't actually progressed beyond 1999 using 56k dial-up. A lot of these multi-megabyte JavaScript-laden sites are basically a big "fuck you" to those in developing countries and even rural/underdeveloped portions of developed countries that don't even have so much as DSL, let alone cable broadband or fiber. For them, it's either 56k dialup or paying absurd amounts of money (in a multi-year contract, no less) for a satellite connection (which tends to have significant performance problems due to the higher latency).

Thankfully, this is starting to become less of a problem as cellular networks continue to expand, but there are still quite a few places where cell coverage is spotty at best (including here, in the Sierra Nevada mountains, where the mountains themselves cast all sorts of coverage shadows no matter how many towers get put up).

Re: The Verge's web sucks

#126

Post by yellowapple »

giancarlostoro wrote: post #47
JustSomeNobody wrote: post #37

Earlier quoted context omitted.

He blames mobile browsers for at least half the article (I had to stop reading). Sure, mobile browsers may not be great, but web sites these days seem to be written by people who just don't give a shit. Their hipster attitude is ruining people's experience.

I agree. I've wanted to write my own blog post about the things that suck about Mobile web, it's not the device or the software (typically), it's the way websites are designed (or not) for phones that sucks. You get advertisements that hijack all your screen real estate, open up the Android Market out of nowhere (why do they have this much priviledge? That shouldn't happen on the desktop without asking for permission…

> open up the Android Market out of nowhere (why do they have this much priviledge? That shouldn't happen on the desktop without asking for permission).

This is based on how Android browsers handle URLs to the Play Store. Some browsers (like Firefox) don't do this automatically (at least for me); instead, they'll just present you with the web page version of the Play Store (as if you were on a desktop), but with an icon in the address bar indicating that you can open the URL in the Play Store instead.

But yeah. The mobile web is a hellhole. My "favorite" is when a site will throw a browser dialog asking to install the app version of their site.

I feel like web developers nowadays are treating XKCDs #1174 [0] and #869 [1] as sincere, genuine advice instead of the scathing observations and condemnations they are of mobile web behavior.

[0]: https://xkcd.com/1174/

[1]: https://xkcd.com/869/

Re: The Verge's web sucks

#127

Post by josteink »

krstck wrote: post #95
gorhill wrote: post #58

Earlier quoted context omitted.

Default-deny is the best answer against bloat -- though not for everybody unfortunately, but I did try to make it as straightforward as possible with uBlock Origin[1]. Using default-deny in uBlock Origin/Chromium[2] and with the page displaying properly, I get from Network pane in dev console: - 62 requests - 508 KB transferred Once you start using default-deny mode, it's difficult to go back to anything less restric…

The difference in page load times before and after implementing uBlock is staggering . Having to use a computer you don't own (or mobile) feels like traveling back to 1999 using 56k dial-up.

Firefox mobile supports extensions like ublock just fine. Use a browser which doesn't suck :)

And if you're using an Iphone: you made the choice that apple should make all the choices for you. You've made your own bed. Sorry about that.

Re: The Verge's web sucks

#128

Post by mdevere »

vvanders wrote: post #83
mdevere wrote: post #56

Earlier quoted context omitted.

Temporarily disabled by default but they turn them on for certain articles. They claimed this was because comments had become too negative. It is possible they are genuinely just trying to protect the mental health of their staff who are usually the targets of more aggressive comments but somehow I doubt it.

It seemed to happen after they posted the SpaceX article(www.theverge.com/2015/6/29/8863121/spacex-falcon-9-rocket-explosioauses) that was poorly thought through. They got some criticism(but not abusive) feedback in the comments. They started deleting comments in that thread and it all went downhill from there once people realized posts were being removed. Kinda a shame, I liked Topolsky & Co but I don't visit it nea…

Deleting comments, locking threads created by reasonable people - in general, over-zealous mods suppressing the voices of users who don't toe the party line - is an unforgivable sin if you're trying to manage an online community, and The Verge is really terrible at this.

Re: The Verge's web sucks

#129

Post by ised »

lmorchard wrote: post #101
ised wrote: post #69

"263 HTTP requests" Seems reasonable. bwahahaha! Shall we discuss the number of DNS requests? And how many of those offsite servers are using something convoluted like Amazon for DNS? (which I find is more and more prevalent thanks to AWS) The blog author, e.g., is using Amazon for DNS. Alas, for each and every name, this adds more than a few lookups to what could be a 1-2 request process. Amazon uses multiple levels…

Actually, FWIW, I'm using pairNIC.com for DNS. I do host on Amazon S3, though.

That means the full lookup for your personal domainname is for an Amazon S3 domainname. PairNIC just handles the first lookup to give the CNAME for blog.lmorchard.com, pointing to Amazon.

Then there are multiple other CNAME's that Amazon uses. It is about the same as looking up a Akamai customer domain. Lots of extra DNS queries.

Post reply on HN