Slashdot Mirror
Opera Introduces Native Adblocking, 45% Faster Than Chrome With Adblock Plus (thestack.com)
An anonymous reader writes: A new version of the Opera desktop web browser introduces fully-featured native adblocking which is able to load adblocked pages significantly faster than rivals running the Adblock Plus browser. The new feature includes whitelisting of domains and a benchmarker to test the difference between page load-times with and without ads. Krystian Kolondra, head of Opera desktop, indicates in his post that the company's hope is to encourage the 'simpler' and less intrusive advertising which has been promised, but does not yet seem to be evident.
How about comparing it to a good adblocker instead?
Nice they keep upgrading their product, but some of their security changes have just left a bad taste in my mouth for Opera. Loved the browser. Had to move on.
Sounds like it would be a great feature. Just convincing to use it is another thing. A lot of problems I have with Opera is it's feel. Sadly as I get older I become less resistant to change. I am used to where everything is and opera to me feels clunky because keybinds are not the same. Some of the convenience features I use although likely there now (not sure) didn't exist last time I tried. If I could get over some of those issue may give it a chance.
3...2...1....
- Don't do what I do, it's probably not healthy nor safe. -
Yeah because setting up and maintaining a host file is just as easy for the Average Joe as downloading Opera or an add-on.
SJW: Someone who has run out of real oppression, and has to fake it.
because it has many flaws, here are just a few I can think of:
-can't list blocked elements
-can't disable ad-blocking for a single web site
-can't disable ad-blocking temporarily without restarting browser (or even PC)
-need admin rights
-doesn't allow complex patterns (regex), which makes the database huge
-can't block ads located on the same server (host name) as the main site
ABP is known to be a pig. A comparison with uBlock Origin would be a lot more meaningful. A comparison with a hostsfile would, of course, not reflect well on any ad blocking extension.
Log in or piss off.
who cares, it's still opera... and owned by chinese.... Vivaldi is better
Go away!
Host files also cause issues when the domain is pointing to 127.0.0.1 A request is made, but there probably isn't an http daemon waiting on the localhost, which in turn causes wait delays and IP stack issues. For those of us that have a local http, we have to deal with all these pointless requests, which will return 404s.
Anyone pushing hostfiles to defeat adverts is a twat and doesn't know what they're talking about. Steve Gibson level twattery no less.
There is one way to get hostfiles to work, and that is to have a http server on the local box, only bound to loopback, which serves up /dev/null to any and all HTTP requests. Something tiny enough to sit in RAM and not impact anything else. This takes care of wait delays and other items, and because the HTTP daemon isn't bothering with error messages, the returns come quickly.
You can lead a hos to water but you can't make it drink.
Browser based ad blockers can block ads before the request is made while hosts files let the request pass but redirect it to a black hole.
It means that in theory, browser-based adblocking can be faster.
No it isn't and it is insufficient at blocking malicious javascript that is part of the page itself.
That javascript can manipulate the CSS and load and add, or malware. So hosts file could be just one layer of the security onion, but by far not the only one.
You still need a deep packet inspection proxy, that manipulates pages (e.g. like privoxy) or you need an inbrowser object dom blocker, like adblock or ublock origin.
And if I have more than one PC at home, it is much better that I properly bock on my border router than use a hosts file, as hosts files are useless for mobile devices, and you have to apply on each PC, whereas with the firewall you only do it once. Also I redirect all DNS requests to my local resolver on my border router and have local static entries for e number of sites, that point them out to IP addresses in the 239/24 and block those out in my firewall, but have the benefit of using counters and a centralized syslog for dropped packets so I can inspect which machine is sending request.
Hosts file just don't cut it. Not by a long shot on their own. Anybody advocating hosts file as the ultimate solution, is a ultimate fool if they don't understand that is insufficient, and an ultimate tool if they understand it, yet don't speak about the rest.
While Vivaldi browser is still in beta (on Linux anyway) I've found it amazingly fast. It can use Chrome plug in and combined with the uMatrix plugin (NoSlash on major steroids) I've found it amazing .Made by the guys who created Opera.
"Imagination is more important than knowledge" - Einstein
Exactly. While this enhancement is intriguing, I uninstalled Opera when that sale was announced.
Host files also cause issues when the domain is pointing to 127.0.0.1 A request is made, but there probably isn't an http daemon waiting on the localhost, which in turn causes wait delays and IP stack issues. For those of us that have a local http, we have to deal with all these pointless requests, which will return 404s.
Anyone pushing hostfiles to defeat adverts is a twat and doesn't know what they're talking about. Steve Gibson level twattery no less.
You can change the IP to anything you want. It's text, edit it if you use a local HTTP deamon.
But if you do that, you are probably also running local BIND, use that to insert authoritative zones to catch new subdomains, and speed up the process with NXDOMAIN
HOSTS files with tens of thousands of entries (like mine) do not cause any noticeable delay. None. Something is wrong with your operating system if it chokes.
The Twats are the advertisers. The whiner is you.
My hosts file has over 2,400 entries that point to 0.0.0.0.
My http server has 0 entries in the log that complain about 0.0.0.0.
I also don't get 404 errors from my server, nor do I notice any performance issues.
Based on the above info, it sounds like you're doing it wrong. Perhaps that's why you posted as AC.
Add the ads at the server.
Then how would the server know where to fetch the ads? I can think of two ways:
The publisher's server acts as a proxy to the ad network I'm not aware of any ad networks that allow such proxying. Ad networks would need to change their terms of use. Advertisers buy ad space directly from publishers That doesn't scale well: O(n^2) vs. O(n). If you have 30 publishers and 30 advertisers, use of an ad network results in 60 contracts, each between one publisher or advertiser and the network. Direct contact between advertisers and publishers, on the other hand, would require 900 contracts. Ain't nobody got time for that. In fact, ability to reach a large audience with one click is one reason that nationwide advertisers moved away from small-town newspapers.I have an even better idea, that bypasses the entire problem: Get rid of any Microsoft 'operating system' (using the term loosely here, since it's getting to be more and more like a 'botnet system' or 'advertising delivery system' than anything else) and install some flavor of Linux (or literally anything else you can get your hands on) instead. Then there won't be ads, 'telemetry' (read as: 'spyware') or any of this other bullshit. Don't really care what your difficulties with the alternatives are, either. Deal with it.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
My solution is to use uBlock Origin to filter browser content, then limit host entries to much shorter lists of known malicious domains that I don't want any process connecting to for any reason (all discussion about malicious ads aside). Keeps from having to fill the hosts file with a zillion entries, and allows more flexibility when I want to fine-tune content blocking in specific cases.
I use a Proxy Auto Config file. It is a java script program that decides what proxy to use based on the URL. If it's a known ad site, then it uses a proxy that redirects to 1x1 transparent GIFs for all requests, otherwise it goes to the real web site. One great thing about this is that I can block based on the path name, so I can even block ads served by the same host as the real content, which you can't do with a hosts file redirect.
If I notice a page is loading slowly, as I did with a local newspaper site, I'll look at the cookies it sends, and block any off-site domains that send cookies. That greatly improved things there.
The one thing I can't block on is the size of the requested image. If the web browser knows that the image is a 1x1, I would love to block it--those are always a waste of time.
No sale has been announced. An offer to buy was announced a month ago. Opera Software is still headquartered in Norway and run by Norwegians.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
Somehow that got posted to the wrong story.. not sure how that happened!
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
It is not theory. I tested it. It is about 200% faster.
It also depends on what your blackhole IP is. I found 0.0.0.0 to be faster than 127.0.0.1. Course the last time I bothered to test this XP was a fairly new thing.
With 127.0.0.1 the name 'resolves' then it times out trying to find port 80 on your own box. You are incurring the timeout for everything you blacklisted. I added in a simple redirecting apache install with a 1x1 gif returned for everything and later a simple php script to do the same thing. Obviously that broke many things but worked surprisingly well.
My biggest gripe was maintaining the list. As some people have pointed out you need basically admin/root privs to do it. Also hosts files do a linear search thru the list. For every single lookup. Not very efficient. So as your list grows you pay a penalty on every other site out there that is supposed to work.
I then moved on to this http://www.schooner.com/~loverso/no-ads/ this was before adblock was a thing so I sort of rolled my own with a combination of this guys code and a massive hosts file. I modified the code to be a binary search instead of a linear search. His pre vetted list is small enough that a linear search is not too bad. But when you ad in 5000 sites it is considerably slower. It was much faster after I added in a binary search. But maintenance was even worse as well. It did have one advantage over hosts files in that you could block just particular parts of a website instead of whole domains as it used regex matches.
I then moved onto adblock then adblock plus and now ublock with noscript. There are plenty of lists to chose from that are fairly well vetted out and they automatically update. Much less work. You can do whole domain, sub domain, sub website, or individual blocking on particular elements in the html, or combine them all together. It is very flexible and little to no work and easily disabled if need be. The downside is malware that does get in is not as easily blocked. But then you could use a hosts file for just that.
For malware I run as a reduced privileged user. Also with a decent virus checker. Have not got anything in 10 years. Mostly because these days malware comes in thru ad networks and needs javascript to run. Someone determined probably could get in. But at that point I am probably screwed anyway. So I just get the easy drive by stuff.
Browser based ad blockers can block ads before the request is made while hosts files let the request pass but redirect it to a black hole. It means that in theory, browser-based adblocking can be faster.
The most common type of hosts file blocking is to point the domain to the machines loop-back interface 127.0.0.1
When trying to fetch "http://ad.doubleclick.net/annoying_ad.gif" the machine will try to fetch 127.0.0.1/annoying_ad.gif
Most likely you don't even have a web-server running on port 80 and the connection attempt will be refused. Since there is no extension taking CPU cycles pattern matching ads this way is actually faster, of course since content instead of HTTP elements are being blocked the results are often not as pretty. It is also a very blunt tool, if ads are being served from the same domain as the rest of the content then everything will be blocked
Add on which auto disable all video, overlays and popups on page with a simple set of buttons to white list where needed. Then whitelist is kept and reused on future visits?
Maybe option to then share list with others.
I should have read more carefully because I thought the deal was done and dusted. No matter. I've been using Vivaldi since and quite like it, and I wouldn't have tried it without some prod to try something else.
it's a nice browser, but it's not really worth using vs Chrome unless you do so for the sake of privacy, at which point you'd be drawn to FireFox. The problem with web browsers is primarily bad webpage coding and we can all fix that easily by just not going to website that load slow or use too much ads. That's really the better solution. We the people train the marketand coders to work how we want it instead of just being sheep spending their lives running from once fence to another because they heard it was better over there. Sites likes slashdot and reddit aren't slow. If you can't make a site without too much trash I just don't go there. We shouldn't have to engineer around the problem. The other big problem is just bad handling of multiple open tabs. They should be able to drop resources on tabs that aren't being used with minimal overhead. You may have 20 tabs open, but only one needs to be fully rendered and using any real resource, yet that is not how any browser works. Finally, we should all disable flash and leave it off until they get the fucking message. It's complete and total crap and flash has caused BILLIONS of dollars in damages due to being massively insecure and massively bloated and inefficient. Flash is basically a virus that can run some cool games.
Most likely you don't even have a web-server running on port 80 and the connection attempt will be refused. Since there is no extension taking CPU cycles pattern matching ads this way is actually faster...
How do you think that the connection attempt gets made and refused, magic non-CPU woo-woo?
It may be true that if the extension is much less efficiently written than the browser and underlying OS services to look up the domain and do the connection attempt, that it will be slower, but all things being equal -- which is the point of this feature -- it should be faster, since you pretty much get to do a single if() rather than going through the entire rigmarole of a connection attempt.
How can we continue to believe in a just universe and freedom to eat crackers if we have no ale?
Since there is no extension taking CPU cycles pattern matching ads this way is actually faster
The browser isn't using CPU cycles for that, but the OS still is. It still has to scan through the hosts file any time a request is made, so increasing the size of the hosts file will not only introduce a slight delay for anything that is blocked, but it will introduce a longer delay for anything that is not blocked where it still has to scan through the entire file. Even with "frequently-used" sites at the beginning of the hosts file pointing to the correct IP addresses, for any lesser-used site it incurs the full delay of scanning the entire file.
At this point though we're splitting hairs, and the delays we're talking about with either an extension or hosts file will probably only be able to be measured with tools to specifically measure those (assuming the hosts file doesn't have millions of records or something, in which case it probably is noticeable).
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
I'm waiting on the switch to Vivaldi for the sale to go through, hopefully they can hit a stable release by then. I'm sure it's coming soon.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
Not as full featured as say Ublock Origin but you can disable by domain or site and uses the standard adblock lists.
While host file are not ideal for blocking ads most modern Linux distros support updating ïthe host file without restarting ïthe browser.
Adblock + NoScript is my solution and so far it seems to work fine.
Adblock kills the ads and NoScript kills off all the naughty little javascript bits that bog everything down.
Just cruising through this digital world at 33 1/3 rpm...
You do realize that querys to DNS are cached by the OS and the cache file will grow much larger than even the largest hosts based ad-blocker if you been browsing the internet for any amount of time. Most sites these days fetch content from a large number of hosts
Most likely you don't even have a web-server running on port 80 and the connection attempt will be refused. Since there is no extension taking CPU cycles pattern matching ads this way is actually faster...
How do you think that the connection attempt gets made and refused, magic non-CPU woo-woo?
It may be true that if the extension is much less efficiently written than the browser and underlying OS services to look up the domain and do the connection attempt, that it will be slower, but all things being equal -- which is the point of this feature -- it should be faster, since you pretty much get to do a single if() rather than going through the entire rigmarole of a connection attempt.
A TCP SYN packet is sent and a TCP RST packet is received. If you want to be splitting hairs yes this does use a few Hertz of cpu time
How does hosts alleviate that problem? For all URLs that are not blocked (and, for the sake of argument, not in the DNS cache) the OS still has to scan the entire hosts file, then the entire cache. If you're not using hosts then it only scans the cache. Using a hosts file only takes the cache out of play if the domain you're looking up is in the hosts file, for every other domain it only adds additional processing time.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
without you-know-who pedding like there's no tomorrow
As for "can't disable ad-blocking for a single web site", true but irrelevant - why would you do that?
I have disabled ad-blocks on a few sites that a.) I wish to remain alive and b.) I've used regularly for years and never had one complaint about their advertising. If you've never felt altruistic in this regard then I can also mention that sometimes ad-blockers get false positives, preventing me from using some features of sites.
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
There is no DNS caching in browsers or OS that could interfere?
Also ad-blocking sometimes break content. A quick workaround is to white list a specific web site.
I use dnsmasq and rather than 127.0.0.1 it points to a local web server where every request returns a zero byte file.
The hostfile is auto updated from multiple sources, some white listings, and the domains with many subdomains are parsed out and added to dnsmasq.conf to block the whole domain.
All very fast running on a raspberry pi 2.
The spirit of resistance to government is so valuable on certain occasions that I wish it to be always kept alive
And to make things worse, request to add MRU is "won't fix" for Chrome, since "addons can do it", except that, they can't.
Weird... your fervent and zealous supporter, whose style is extremely similar to yours, also posted to the wrong story.
I have noticed more and more instances where ABP is inactive. Right click, and there is no "Block image" choice. In yahoo mail for example, unblockable animated ads now appear in the right margin. What's up with that??
Either way it's still a good idea...
playmoney.me - The free alternative to paper board game play money
You can lead a horticulture, but you can't make her think.