Hacker News Reader

Comment by Sanddancer | Hacker News Reader

I don't have this problem, blank pages don't take "seconds" to load, even on my little atom-based tablet. The big question I have for people noticing this behavior is how many add-ons do they have. How many times is your "blank page" parsed by an enormous amount of code written in a language not known for its performance?

View on HN

The only add-ons I use are ones to remove ads. If I didn't have them those ads would easily make the my overall performance much, much worse.

I have a custom blank page loaded from SSD that has only a small amount of html5/pure js and it takes maybe 1000ms on an intel i5/3xxx. If I open a new tab without any other windows/tabs open it's very fast, maybe 100ms. A new private tab (doesn't load add-ons or my custom new tab) opens quickly but not as quickly as a fresh restart w/ no windows/tabs.

Actually some adblockers were found to increase CPU and memory usage significantly.

In particular, ad blocking rules that match CSS selectors instead of URLs are a challenge, especially when the the CSS rules apply to any domain and thus have to be loaded for every page. IIRC uBlock's primary advantage over AdBlock Plus was in ensuring that the CSS rules were being shared between tabs instead of requiring memory proportional to the number of tabs. Popular third-party filter sets like EasyList have grown in recent years to use a really incredible number of broad CSS rules.

And some scripts eat cpu with adblock, I never looked into it but it's like they are trying over and over again to reach a host. Try opening 50 tumblr tabs with ad block on.

> Actually some adblockers were found to increase CPU and memory usage significantly.

And then it was fixed. No reason to assume it still applies, unless you have new data.

On one of my machines, Chrome has at least a second delay before actually starting to load a page if and only if hw-acceleration is kept enabled. I have also seen odd delays on other machines on other browsers that were related to hw-acceleration.

Yes, having a great number of intrusive extensions can cause issues, but there is no need to jump to conclusions as these issues can be present without the user doing anything "wrong".

If your browser is designed from the ground up to support add-ons, but can't handle a couple of them without slowing to a crawl, your browser sucks. Don't blame the user for a problem with the software.

> your browser sucks

Why not all three? The browser implementation, user's choices, and the implementations of each individual add-on suck.

The browser didn't limit the resource use of greedy add-ons (or inform the user of conflicts between them), the add-ons were coded either to do too much, or to perform their function with suboptimal code, and the user is the one that chose that set of software to use together.

Reply on news.ycombinator.com