[Rejected] Whitelisting broken sites

Various discussions related to Adblock Plus development
Wladimir Palant

[Rejected] Whitelisting broken sites

Post by Wladimir Palant »

  • Note to Reddit users: Please read this post before believing some crap somebody tells you on Reddit. This is not about whitelisting ads, it's about whitelisting websites that are currently broken in Adblock Plus.
As I mentioned in forum/viewtopic.php?f=4&t=7551, in the long term we want to have a viable solution to allow ads that meet our criteria and are not annoying. This will take time however, so I was thinking about running an experiment. The redesigned first-run page would allow users to subscribe to a "special" filter list, via a checkbox. This list would contain a few entries of the form @@||example.com^$document - disable Adblock Plus on these domains. We are talking about a small list of exceptions for sites that meet the following criteria:
  • The site doesn't contain any advertising or only advertising that meets the criteria.
  • Ideally, we have an agreement with the website owners on the first point.
  • The site is likely to be broken by filter subscriptions and the problem won't be fixed in subscriptions for policy reasons.
Here are two example sites:

http://www.adtechus.com/
http://www.admeld.de/

Both are related to advertising and landed on the filter lists because of that. Both EasyList and Fanboy's List do not "fix" such cases (stated policy). Yet whitelisting the sites won't do any harm, quite the opposite - the few users who get there by choice won't be met with a broken website.

This is a small experiment with very limited impact. The goals are:

a) See how users take this.
b) Get us talking with advertisers and website owners.
c) Experiment with the technical side.

This will hopefully bring us closer to the long term solution. Outstanding questions on the user interface side:
  • How should that checkbox be called? It will definitely have a link next to it leading to a webpage with detailed explanation on the purpose and policies but we still need a simple title.
  • How should this option be presented in the Filter Preferences dialog?
anonymous74100
Posts: 213
Joined: Sat Mar 19, 2011 3:45 pm
Contact:

Re: Experiment: Whitelisting broken sites

Post by anonymous74100 »

Who will be maintaining this "special" filter list?
Will the checkbox be checked by default?
Wladimir Palant wrote:See how users take this.
The users intalled Adblock Plus not MaybeBlockSomeAds Minus, so they probably won't be happy.
Wladimir Palant wrote:Get us talking with advertisers
Collaborating with the enemy is not a good idea.
Wladimir Palant

Re: Experiment: Whitelisting broken sites

Post by Wladimir Palant »

anonymous74100 wrote:Who will be maintaining this "special" filter list?
Most likely Till since he would be the one talking to advertisers. We are talking about 10-20 entries at most.
anonymous74100 wrote:Will the checkbox be checked by default?
Yes, the entire exercise would be pretty pointless otherwise. It is still user's choice however - anybody who wants all ads eradicated can still have it.
anonymous74100 wrote:Collaborating with the enemy is not a good idea.
Interesting approach. I guess you will uncheck that checkbox.
anonymous74100
Posts: 213
Joined: Sat Mar 19, 2011 3:45 pm
Contact:

Re: Experiment: Whitelisting broken sites

Post by anonymous74100 »

The checkbox should be: Uncheck this to uncripple Adblock Plus
Last edited by Anonymous on Thu May 12, 2011 8:46 am, edited 2 times in total.
Reason: Big font removed
Wladimir Palant

Re: Experiment: Whitelisting broken sites

Post by Wladimir Palant »

Thank you for this constructive contribution. Now please move along, this is clearly not a feature you will be using.
anonymous74100
Posts: 213
Joined: Sat Mar 19, 2011 3:45 pm
Contact:

Re: Experiment: Whitelisting broken sites

Post by anonymous74100 »

How would this checkbox effect people who upgrade ABP, not install it the first time?
Wladimir Palant

Re: Experiment: Whitelisting broken sites

Post by Wladimir Palant »

I don't know. We have the option to assume "not checked" or to show them the first-run page again (probably in a reduced form). The former has the disadvantage that it will make the user base for the experiment significantly smaller, we might not get useful results then. The latter also has obvious disadvantages, we don't want to annoy users.
anonymous74100
Posts: 213
Joined: Sat Mar 19, 2011 3:45 pm
Contact:

Re: Experiment: Whitelisting broken sites

Post by anonymous74100 »

Wladimir Palant wrote:We have the option to assume "not checked" or to show them the first-run page again (probably in a reduced form).
Actually you have the choice of assuming "checked" or showing the first-run page.
If you assume "not checked" then this experiment might as well not happen, cause there won't be enough participants.
Since assuming "checked" is a bad idea the only thing left is the first-run page.
Wladimir Palant wrote:we don't want to annoy users.
Users will allays be annoyed, there's no point in worrying about that.

What's the target version? 1.3.7? 1.4?


P.S. I still don't like this.
Wladimir Palant

Re: Experiment: Whitelisting broken sites

Post by Wladimir Palant »

Adblock Plus 1.3.7 is already released. The target version is 1.4 because that's where we implement the big user interface changes - that makes sure that the user can easily get an overview over his settings.
Ares2
Posts: 1275
Joined: Fri Feb 15, 2008 12:47 pm

Re: Experiment: Whitelisting broken sites

Post by Ares2 »

The only thing I find strange is that the reason for and the main purpose of the list (whitelist sites that explicitly agree on using acceptable ads or no ads at all), is mixed with a completely unrelated and hardly as important issue that could theoretically be misused to whitelist ads:
Wladimir Palant wrote:[*] The site is likely to be broken by filter subscriptions and the problem won't be fixed in subscriptions for policy reasons.
Can't those things be separated from each other? I think that would also help the reputation of the list as it would only contain "good" sites and not a collection of all the adserver homepages out there which perfectly promotes the "ABP sold out" conspiracy.
Last edited by Ares2 on Thu May 12, 2011 9:55 am, edited 2 times in total.
User avatar
Hubird
Posts: 2850
Joined: Thu Oct 26, 2006 2:59 pm
Location: Australia
Contact:

Re: Experiment: Whitelisting broken sites

Post by Hubird »

Assuming "checked for upgrades" is not a very good option IMO, this is a big change to the way ABP works (going from zero tolerance to allowing ads is not something that should be forced upon users). A reduced first run asking users to opt in seems like a fair approach to me (you don't want to force people to play guinea pig).

And what of those users who don't use any subscriptions, will they have entire sites whitelisted ?
Till
ABP CEO
Posts: 242
Joined: Tue Oct 12, 2010 2:16 pm

Re: Experiment: Whitelisting broken sites

Post by Till »

Can't those things be separated from each other? I think that would also help the reputation of the list as it would only contain "good" sites and not a collection of all the adserver homepages out there which perfectly promotes the "ABP sold out" conspiracy.
We want to make the internet better, don't break it. The reason for testing this functionality is to make sure that ABP does not worsen the surfing experience for the user. There are mainly two reasons why people currently need to switch ABP on and off all the time: some sites are broken, some don't have intrusive ads.

If the test works out well we can hopefully come to a solution that fixes the internet automatically without any action required from the average user who doesn't want to be bothered with whitelisting etc.
Ares2
Posts: 1275
Joined: Fri Feb 15, 2008 12:47 pm

Re: Experiment: Whitelisting broken sites

Post by Ares2 »

It all depends on the sites that will end up on that list. The current definition gives me personally all reasons to be worried about it and the given examples don't convince me either. The average user doesn't visit homepages of adservers and I can't say that I have seen many of those issues coming up in the past year. Also it wouldn't be THAT hard to fix them in a subscription anyway if you think those are that important, I always had the impression that we (EasyList) simply don't do it because they aren't even worth those medium-effort modifications (although we do have added the $third-party to all the adserver rules back then exactly for that purpose).

Afaik all other sites get fixed in the subscriptions if it possible in any way (even by allowing certain ads) - so what other sites except homepages of adservers are we actually talking about here? The definition allows the worst of the worst (those that break intentionally beyond repair and those that have no other content but ads) to be completely whitelisted - now if that is the purpose of this list, count me in on the people that hate and oppose it with a passion.
Michael
Posts: 1361
Joined: Sat Dec 19, 2009 12:29 pm

Re: Experiment: Whitelisting broken sites

Post by Michael »

Generally, we only fail to resolve issues on domains if it isn't compatible with the more general and, I believe, primary policy of blocking adverts. This is an exceptionally rare occurrence, and the suggestion that there are multiple "broken" domains is unfair on subscription authors. If we are notified of issues, they are almost always fixed within 24 hours. Pre-emptive whitelisting goes as far to break the internet, in the eyes of what I would suggest are the majority of users, by letting adverts through.

The above statement does not include domains for adservers, which are not included, at least in the context of EasyList, because we do not envisage that Adblock Plus users are likely to either want or need to visit the websites. If people hate adverts so much that they install an extension to remove them, why are they visiting domains exclusively about such content?

I also agree with the other subscription authors who have posted. On the topic of whether or not to provide a minimal first run page, I believe users will be more annoyed to discover that adverts are being whitelisted without consent than by having to visit a page to confirm that they want Adblock Plus to provide an advert free environment. On the topic of whether or not the checkbox should be automatically selected, I would suggest not. It is my belief that defaults are meant to represent the normal value, not the program author's intention, and the norm for Adblock Plus users is undoubtedly, in my mind, that they do not want adverts under any circumstances.
Till
ABP CEO
Posts: 242
Joined: Tue Oct 12, 2010 2:16 pm

Re: Experiment: Whitelisting broken sites

Post by Till »

Sure, not many ABP users visit these sites, they just serve as examples of currently broken websites. And we most certainly are not talking about unblocking their ads here. What we should keep in mind is that ABP is growing to a userbase that is less technical savvy, most of them don't care at all about filter rules (they are probably not even aware of the fact that ABP itself doesn't block anything). I agree that you are doing a great job in fixing false positives very fast in EasyList, other filter list maintainers may not. If there is an issue, ABP gets the blame. So some kind of a quality assurance backup is important imho.

Regarding the unblocking of certain ads we need to think ahead: At some point ABP will have so many users that some websites can't afford producing high-quality content for free anymore. I don't expect the majority of the users wants paywalls on every website (which btw would only favor the big players and eliminate all the small blogs etc), I do believe most users will favor to accept unobtrusive ads. The ones that don't can continue to use ABP as it is. Of course, nobody wants the wrong sites and ad-formats ending up on a whitelist, that's why we need clearly defined criteria which we can only get by starting this experiment. And as far as the integration of the checkbox is concerned, I think we should all trust Wladimir to do what makes most sense for the experiment as well as the users.
Locked