Page 3 of 3

PostPosted: Sun May 10, 2009 10:48 pm
by Lyx
One issue with 3rd party blocking is that it is difficult to tell if that other server belongs to the same owner, or a different one. Just a few days ago, i stumbled over a situation where scripts necessary for the site to function, where pulled from a different server - which however had the same owner.

So, whilte for example blocking of 3rd party script domains may be an efficient start, its not perfect - you will sometimes block scripts from the same party on a different domain. In that way, what this "3rd party filtering" really means, is filtering of content from other servers than the one on which the request originated - and sometimes not even that.

PostPosted: Sun May 10, 2009 10:50 pm
by Lyx
P.S.: IF you need an example for this, then go to youtube and watch how the page is thinking that you dont have flash installed or javascript disabled, even though that is not the case.

PostPosted: Mon May 11, 2009 9:08 am
by mrbene
Lyx, you've provided another excellent example of the ambiguous types of configuration that exist on the public internet, situations where additional research (in this case a quick look to the domain registrar) would result in a clear understanding.

It's been my experience that every possibly configuration is pretty much guaranteed to exist, somewhere.

That said, I'd like to turn the question around a little bit: Who is in the best position to build a trust relationship with the recipient of ads? I think that webmasters can do this to a certain degree, but that it's the ad networks that could really make the changes.

Picture this - as a recipient of ads, I tell Google/Doubleclick that I don't want Flash based ads. Then, wherever Google/Doubleclick serve ads, I don't get that type of ad. They're already moving on this front (check it out).

Overall, trying to shim the webmaster in as a manager between the ad recipient and the ad network is a limited vision approach. And asking the ad blocking tools to support something that the ad networks don't is inefficient.

At least, that's my take.

Ad blockers allow consumers to send the message "your ads suck" to the networks. The networks are the ones that are really going to generate industry-wide change, in terms of user choice, if only to win back eyeballs (and revenue).

PostPosted: Mon May 11, 2009 2:33 pm
by Ares2
Wladimir Palant wrote:As to content types, I am not sure how meaningful that still is. Aren't most ads scripts these days?

3rd-party -> yes, 1st-party I would say no, maybe some but not most of them.

Wladimir Palant wrote:The other complication is banner rotation - the same ad network might serve acceptable ads at one point but something absolutely unacceptable (annoying animation, sound etc) a few minutes later.

Well, I think that's not something we should care about. If an adservers also serves annoying ads, it will be blocked as a whole in "level 1". Those ad netowrks that are only serving acceptable ads (like Google's pageads) will only be blocked in "level 2".

Assuming that "level 1" ideally blocks all annoying ads and represents what a high percentage of Adblock users want/use, site owners could decide to switch to one of those acceptable hosters and this could lead to some adservers revising their strategy. Very optimistic, I know, but possible. :wink:

PostPosted: Mon May 11, 2009 4:24 pm
by Wladimir Palant
Here comes a proposal which might work: http://adblockplus.org/blog/an-approach ... d-blocking (doesn't rule out all the other proposals but will complement them). Feel free to comment (the forum thread for this topic is http://adblockplus.org/forum/viewtopic.php?t=3923).

PostPosted: Mon May 11, 2009 4:46 pm
by prinn
Wladimir Palant wrote:That's true for any request - including the one requesting the webpage itself. Depending on the webserver, you might be connecting to a reverse proxy that will pull the information from different sources (including a third-party it will send your data to). That's something you cannot really prevent, it can only be dealt with by non-technical means (privacy policies). First-party ad calls definitely don't make a difference here, you are still talking to only one server and it up to this server to deal with your data responsibly. Third-party ads are different, here you are explicitly talking to third-party servers, something you likely don't want if you are concerned about privacy.


I acknowledge that any request may result in inappropriate data handling including the sharing of personal, sensitive, and/or behavioral data with third parties. However, lets recognize that advertising contexts are high occurance contexts due to the intrinsic value of such data for ad targeting purposes and the ad request/response phase is almost always at the heart of it. To equate the risks of first-party NON-ad requests with the risks of first-party ad requests is to neglect the context. I do agree that third-party ad requests are of greater concern than first-party ad requests for several reasons. However, not even that contrast is perfect for the user might consider the third-party entity entirely trustworthy while considering the site they are visiting as questionable. Everyone visits sites they know nothing about and many visit sites they don't really trust.

Verbage mentioning guidelines and exceptions for "acceptable advertisement" is disconcerting. For there is not now, and there never will be, a suitably objective much less practical way to rate the consequences of ad request/delivery. Privacy policies are largely worthless in many if not all countries and no one can audit what goes on behind the scenes. Furthermore, the user is the only one who can or should make decisions regarding which requests/actions/ads are acceptable and which aren't.

However, utilizing "ad descriptors" which communicate the known characteristics of ads (which is subtly but critically different from rating ads and categorizing any as acceptable) and providing users with the ability to block/fetch based on said descriptors empowers and encourages users to make the calls for themselves.

In a ratings based system, a judgement passer examines the ad characteristics and assigns the rating and that feature alone only allows the user to control things at the granularity of the rating scale (usually a few coarse levels). In a descriptor based system, the judgement passer is eliminated (which undercuts some opposition arguments) and the user can control things based on whatever characteristics can be described by the descriptors (some of which could be detectable by Adblock Plus such as first-party vs third-party).

Off to read read/contemplate the blog post...

PostPosted: Mon May 11, 2009 6:07 pm
by pirlouy
Like some people, I'm not sure it's a good solution to ask webmaster. They don't know what is displayed to visitors. For example, googlesyndication can display text ads, but also annoying gif, or even Flash... :/

I second those who think it should be only based on history. But whatever you decide, it's a good idea and a step in the right direction !

PostPosted: Tue May 12, 2009 9:18 am
by Phil
Luthair wrote:Should Television stations also be able to require viewers watch and listen to commercials?


NO. And I have a television show, not a station.

PostPosted: Tue May 12, 2009 3:21 pm
by IceDogg
After thinking about this and getting over my own anti-ad bias, it's not such a bad idea. Basically this is just a more user friendly (and discoverable) whitelist. It's still up to the user in the end and that's the important part. Only thing I still disagree on is being able to disable it easily, but if it bugs people enough they can read a little (which wouldn't hurt some people anyway) and find out how to disable it, if they want. As long as this can't be used in any way to detect ABP which Wladimir already answered for me on the blog postings
No, this wouldn’t be a practicable solution to detect ABP. Even if the website could detect a “reaction” to this tag somehow (which it cannot), the reaction would be rare enough to make this way of detecting ABP users useless
...so I'm satisfied. Thanks Wladimir.

PostPosted: Tue May 12, 2009 3:43 pm
by Picard
How interesting.

I expect it will take a good 8 days before we get as many "May I show you this?"'s as we get ads today.... :roll:

I find all ads annoying

PostPosted: Tue May 12, 2009 4:27 pm
by Joanne Mullen
I find all adverts annoying. I want ad block plus to block them all by default without any fuss and without any exceptions unless I specifically choose to whitelist a site myself without being nagged about it. If you're not interested in doing that any more just say it so your users can go somewhere else.

PostPosted: Tue May 12, 2009 6:25 pm
by greenhatch
Nearly 200 replies in Wladimir's blog the majority ckearly against the proposed new feature. However I'm doubting this will turn the tide. So long as I can disable it totally then I'll live with it.

PostPosted: Tue May 12, 2009 7:00 pm
by MonztA
greenhatch wrote:Nearly 200 replies in Wladimir's blog the majority ckearly against the proposed new feature.

I guess most of it is just venting without even reading Wladimir's proposal.

PostPosted: Tue May 12, 2009 7:10 pm
by Wladimir Palant
greenhatch wrote:Nearly 200 replies in Wladimir's blog the majority ckearly against the proposed new feature.

Given that so many commenters come from Slashdot that's hardly surprising. I am more surprised about so many useful (and positive) comments.

PostPosted: Wed May 13, 2009 9:48 pm
by eBusiness
For this to fly I don't think you can ask the user to choose for individual sites. I think I speak for 99% of users when I say that we want an "install and don't think about"-tool.

If ads are to be discriminated I say the filters will have to do it.

To pick the low fruit first start categorising the ad networks as to what standards they live up to (something like the criterias Lyx listed) and give the user a choice as to what categories to filter.

The more time consuming step is to start doing the same for web pages, a system where all users are given a simple tool to report a web site and a boatload of volunteer moderators review the reported websites to ensure that the listings are fair.

In any case, I don't see much use for the "this site is good" tag, I doubt that the users will be any better than the non-users, eventually it is just more overhead in the HTML headers.