EasyList

EasyList Italy update

written by MonztA, on Nov 15, 2011 2:58:51 PM.

The EasyList Italy subscription has been granted recommended status in Adblock Plus, which means that it will be the default selection for Italian users.

If you want to provide feedback about EasyList Italy, including reporting adverts or problems, please use the Adblock Plus issue reporter or open a topic in the EasyList forums.

Announcing EasyList Italy

written by MonztA, on Sep 27, 2011 8:19:00 PM.

Although EasyList author Khrin started to develop EasyList Italy just two weeks ago, it already has a considerable amount of filters and it is now ready to go live. The subscription is designed as a supplement for Italian-language websites and is now listed on the EasyList homepage, from where it may be viewed or installed.

If you want to provide feedback about EasyList Italy, including reporting adverts or problems, please use the Adblock Plus issue reporter or open a topic in the EasyList forums.

Two years, seventeen days, three hours and five minutes

written by Michael, on Sep 10, 2011 12:13:00 PM.

It has been two years, seventeen days, three hours and five minutes since I became involved with EasyList by creating an account on the EasyList forums. When I joined, I did not expect that I would ever rise to my present position of lead EasyList and Privacy author, forum administrator and all the other roles that I have accumulated in my EasyBlog profile.

It was chance that lead to my first major breakthrough: only three months after I had joined the forum, I was granted contributor status and thereby discovered, via a topic in the private forum, that Ares2 was officially struggling to maintain the subscriptions (my 500 posts over my first three months on the forum had nothing to do with this). Wladimir was, at the time, reorganising the Adblock Plus repositories anyway and offered a Mercurial repository to Ares2, which was eventually accepted. The question then arose who would assist Ares2, which I replied to with the following post:

"I would be willing to help, although I appreciate that I am somewhat of a novice in comparison to the other members of the private forum."

Michael

Despite my acknowledgment of my limited involvement with EasyList, my application was accepted, along with Erunno's, and by early December I was actively committing changes.

Becoming the lead author was also a matter of being in the right place at the right time. By the end of April I was effectively the only active EasyList author, to the extent that I complained, via private message to Erunno and Ares2, that:

"I signed up for a group project, not a one man band."

Michael

Although I received assistance from Erunno, Ares2 found himself unable to continue to run EasyList and in June he therefore offered me the position of lead author, which I readily accepted. It was also acknowledged that we need more help, and three people were offered authorship at the time: MonztA, Khrin and Famlam, of whom MonztA and Khrin immediately accepted and soon took on their new positions.

After the peculiar circumstances that lead me to my position, I decided to develop EasyList further, redesigning the website, which had not been significantly altered since Rick was the EasyList author, and adding this blog. I added groups to the forum to more clearly identify significant people, such as contributors, subscription authors and EasyList authors, and added the private "EasyList Authors' Forum" to reduce the excessive use of private messages by authors. I have also been one of two authors of SubscriptionList and helped to develop the Adblock Plus documentation and check the English localisation of Adblock Plus - I am responsible for the description of Adblock Plus on the Mozilla Add-ons domain. More recently, Famlam has taken up my offer to become an author and I have developed FOP (Filter Orderer and Preener), which automatically sorts and commits changes to Adblock Plus subscriptions.

That is not to say that I have no regrets about my time with the project - I have certainly made some significant mistakes, including committing a rule that would have broken a very large number of websites had the filter not been spotted quickly and being overly antagonistic to an important company - but that I am generally proud of my involvement with EasyList.

Perhaps the most significant change to the EasyList project while I have been the lead author is to make the subscriptions available as Tracking Protection Lists (TPLs) for Microsoft's Internet Explorer using a script I developed, but which was later significantly improved by Wladimir; EasyPrivacy TPL now accounts for just over half of all EasyPrivacy downloads.

Those are my memories of EasyList - my reminisces. But I now need to move to the present and to reality. Over my two years with the project I have been successful in real life, but have realised that the resulting changes will reduce my level of my involvement for the foreseeable future to such a low standard as to do a disservice to the community. EasyList needs stability - there has been approximately one lead author a year since Rick's unfortunate death in 2009 - and I am unable to provide it. As I hinted in a hidden message in a recent blog post, I am therefore stepping down from all of my roles associated with EasyList and Adblock Plus.

MonztA has kindly agreed to take over the role of lead EasyList and EasyPrivacy author (he is already the lead EasyList Germany author), effective immediately. He has been writing subscriptions for at least twice as long as I've even been involved with EasyList and I know that he already has plans to continue to develop the project. I wish him and the community the best of luck and thank the numerous people (too many to list here) who have supported me as an author. But now I must leave. After two years, seventeen days, three hours and five minutes.

New EasyList author: Famlam

written by Michael, on Sep 9, 2011 3:02:24 PM.

I am pleased to announce that we have a new author for EasyList, EasyPrivacy and EasyList Germany: Famlam. He has been a member of the EasyList forum for nearly a year and a half and is already a contributor to AdBlock for Chrome, one of several advert blocking add-ons for Chrome that use the EasyList subscriptions. He was officially granted the status of EasyList author yesterday evening and I hope that you will welcome him into his new role.

I would also encourage anyone else interested in becoming an author for EasyList, EasyPrivacy or EasyList Germany to contact the project. As I noted in a recent article on EasyList usage and downloads, we have very few authors relative to the millions of users of the subscriptions and would be glad of additional assistance.

EasyList Statistics: August 2011

written by Michael, on Sep 1, 2011 12:00:00 PM.

Looking back, it has been a long time since anyone has evaluated or released the statistics for the EasyList subscriptions, so I thought that, as August has just eneded, it might be time to estimate exactly how many users there are of the filter lists.

Even though I am aware that most people are interested simply in the user estimates, I should start by explaining how the statistics are gathered for the users of EasyPrivacy and any other privacy conscious individuals. When a browser connects to a web server, it automatically sends information to request a copy of, in this instance, a text file. This request includes the user agent, which for Firefox states the name of the operating system (although this information is not present in my statistics) and the version of the browser. The Adblock Plus website, like so many others, logs these requests with the time and date the request was made. These logs are then analysed by AWStats. No active tracking scripts have been installed.

Although there is a lot of information, calculating the number of users of the EasyList subscriptions cannot provide a complete picture for a variety of reasons. Adblock Plus updates subscriptions between four and six days after they were last updated. However, if a user does not use a browser for a day, then an update will be delayed for a day, up to ten days after a subscription was last updated. This being August, many people are likely to be on holiday, which means that the subscription updates are likely to be less frequent than my assumed value of five days. It is also possible, but extremely unlikely, that results may be overestimated as a result of many people forcing a download from the Adblock Plus preferences window.

Viewing the statistics at the time this algorithm was implemented, Wladimir hoped that the changes would prevent a “Monday peak” from overloading the server, the result of office workers all updating on Monday morning if the five day interval passed over the weekend. However, it is clear that there are a number of users of the subscriptions who do not have a recent version of Adblock Plus installed. Using the four full weeks of August 2011 as an example, Mondays show on average 12% more downloads than would be expected (week total / 7), ranging from 110.8% to 118.5% of the expected traffic. In contrast, downloads tended to tail off as the week progressed, the most marked drop being on Saturdays and Sundays, with an average of 92% of the expected traffic each day. This ranged from a low of 91.6% on Saturday 13 August to a weekend high of 95.8% on Sunday 28 August.

In spite of the variation based on week day the total number of downloads per week was fairly stable: in order, each week achieved 98.1%, 99.0%, 100.7% and 102.2% of expected traffic, the initial low results likely the results of holidays without computers and therefore without updates to EasyList.

Now, however, it is time to answer the big question: how many people actually use the subscriptions? These values were calculated by taking the total number of downloads in a month, dividing the resulting total by thirty one to find the average number per day and then multiplying by five, as this would be, if everyone used their browser daily (which they don't), the average update period.

Going through the subscriptions, it is no surprise that EasyList has the largest user base. Taking into account all combined subscriptions, versions and variations, such as the one with element hiding rules, the one without rules for adult websites and relevant tracking protection lists, we find that there were seventy three million downloads of the file in August, suggesting that there are twelve million users, of whom 99% use Adblock Plus filter format. EasyList Germany has also been popular, with two and a half million users (99.9% Adblock Plus format), as has EasyPrivacy, despite the comparatively paltry total of "just" half a million users. However, unlike EasyList and EasyList Germany, which have seen little usage as Tracking Protection Lists, just over half of the users of EasyPrivacy use EasyPrivacy TPL, its popularity likely due to Microsoft's promotion of the list, which we are very grateful for.

Next, it is possible to put these values in perspective: how dominant is EasyList as an Adblock Plus subscription? Looking at the public statistics page for the add-on and ignoring the recent erroneous readings, we find that there are twelve and a half million users of Adblock Plus. Taking into account the usage of AdBlock for Chrome and Adblock Plus for Chrome, which according to user agents accounts for, at most, 15% of installations, we find that 80% of people using Adblock Plus use EasyList as a primary subscription. This is despite the random selection of EasyList or Fanboy's List for a new English language user. I would attribute this popularity to historical reasons (people installed Adblock Plus when EasyList was the only recommended subscription for English language users); the large number of recommended supplementary subscriptions, including EasyList Germany, RU AdList, Liste FR and ChinaList [*], to name but a few, which are selected automatically when the subscription and browser languages match; and its position as a fallback subscription: if the browser language does not match any recommended subscription, EasyList is automatically installed.

Of course, the minor subscriptions also deserve a mention. EasyList without element hiding has accumulated around seven thousand users, EasyList without rules for adult sites has over two thousand users and EasyPrivacy without international filters has just over one thousand users. Finally, EasyTest, an EasyList supplement used for examining filters known to cause problems on websites, has accumulated just 33 users.

While there is great variation in the usage of the EasyList subscriptions, the statistics emphasise the great responsibility that we authors have - a minor mistake can quickly affect thousands of people. I would therefore appeal to anyone who has made it to the final paragraph to consider getting involved with the project to assist the very few authors we have - even suggesting rules for problematic sites helps us to reduce the time it takes to commit a fix. Regardless of this challenge, I am optimistic: provided the requisite bandwidth is made available (4.2 TB was required this month for the Adblock Plus version of EasyList alone), I see no reason why the project should not continue to be successful.

[*]

I am not going to discuss the usage of the affiliated subscriptions as I do not have the statistics for the files hosted third-party, but the subscription authors are welcome to write their own analysis with the assistance of statistics for the EasyList mirrors - you should know how to access the logs, if not, please private message me.

What is acceptable first-party tracking?

written by Michael, on Aug 31, 2011 6:17:49 PM.

Although all third-party tracking is blocked outright by EasyPrivacy, the subscription's policy is slightly more lenient with first-party tracking, specifying that that items should be blocked only if they "collect a significant amount of personal data". As the wording is ambiguous, I have decided to convey my interpretation of the policy with regard various pieces of information available from the user and whether the subscription should specifically prevent the collection of such data.

IP address, user agent and referrer

Although this information is recorded in system logs, the details make users identifiable, making this information personal. I would therefore suggest that EasyPrivacy should continue to regard the first-party collection of such data as unacceptable and prevent it being sent to any additional systems on the domain wherever possible.

Screen resolution, colour depth and browser plugins

This information is not located in the system logs and pertains only to the user's system. I would therefore suggest that tracking this information is unacceptable.

Time zone

Time zone provides a very approximate location of the user, and I would therefore suggest that this information should be classified as personal and prevented from being collected where possible.

Language

If the language is that of the website the I do not have an issue with the information being collected, as it is about the site rather than user. However, to record the value set in the browser oversteps the mark in my opinion.

JavaScript status

I would regard the JavaScript status as generic information provided that trackers do not attempt to determine the reason why scripts are unavailable, and I would therefore suggest collecting this type of information is acceptable.

Advert blocking status

There is no intrinsic issue with detection of advert blocking software (we do not object to websites adding unobtrusive messages to their pages specifically for users of such programs). Therefore, provided that no detailed information such as the name of program responsible is collected, I would regard this type of tracking as acceptable.

Page status

Information about the page itself, such as the address or whether the content has loaded, is not personal information, and I would therefore suggest that collecting information about the site status is acceptable.

"Random" numbers

I have no issue with genuinely random numbers being generated on each page or the use of a timestamp. However, I would not consider it acceptable to use a number to track the user's visit trail around a website without their consent, as this would make the value little better than an IP address.

Object tracking (e.g. using Flash or Java to track users)

In the vast majority of instances objects are unnecessary alternatives to other methods of tracking, such as JavaScript. Furthermore, these objects can set cookies that cannot be easily controlled by the browser. I would therefore regard the collection of information using plug-ins as unacceptable in practically all circumstances.

The definition of "advert" and link exchange policy

written by Michael, on Jul 11, 2011 3:14:00 PM.

The basic principle of EasyList is that its "filters must remove adverts". However, several discussions, usually with website owners, have highlighted an obvious issue: "advert" has not been precisely defined. A topic was therefore opened in the new EasyList policy forum and culminated in the definition of an advert as "the promotion of third party content in return for goods or services". This has several implications, as explained below.

"Promotion"

Specifying that an advert must promote means that a factual link (e.g. "Adblock Plus is a Firefox add-on.") cannot be an advert as it lacks a marketing aspect. Furthermore, an independent review is also unable to be an advert because it critically examines all facets of the content, not just the positive aspects.

"Third-party"

The aim of a website is to entice people to read more of its content, which visitors presumably desire to do, as indicated by their presence on the domain; they should therefore not be hindered in doing so by a filter subscription. Whether websites are third-party or first-party (what happens if one person owns two domains?) is a question for the EasyList authors to consider.

"... in return for goods or services"

An important question that has been raised is whether an advert requires the exchange of money, to which I would respond that it is not an essential element. Money may be one of the goods supplied, but there may also be services provided in return. Therefore a link to, say, a web host would be removed if this is a sponsored link required as a provision of service, but not if the description is merely factual, as this would not be promotion.

This definition includes link exchange schemes, whereby websites reciprocally exchange logos and links to the other domains for no other benefit, which Tarun, a website owner himself, claimed was not worthwhile removing given the large number of domains which use the link exchanges and their comparatively unobtrusive nature. After further discussion it was agreed that, provided links exchanges are unobtrusive and free of any tracking parameters, they will not be specifically removed by EasyList.

Advertisers' and tracking companies' websites

written by Michael, on Jul 1, 2011 6:44:00 PM.

Problem:

Some websites relating to advertisers and tracking companies (predominantly the former) do not function correctly when EasyList or EasyPrivacy are enabled respectively.

Affects:
  • Google AdSense

  • Adcash

  • Other domains are also likely to be affected and will be listed as they are reported.

Cause:

EasyList is intended to remove adverts, just as EasyPrivacy is intended to remove tracking. The subscriptions often include general rules to remove the unwanted content and also selectively block domains that exclusively serve such items. However, as these filters are often and unsurprisingly activated by websites belonging to advertisers and tracking companies, some parts of the domain may fail to function correctly.

Solution:

Subscription policy is that it is not worthwhile to resolve issues on domains belonging to advertisers and tracking companies, beyond requiring that filters blocking an entire domain only apply on third-party websites, as we believe that very few users regularly visit such websites. However, it is possible to disable Adblock Plus on the affected website using the option in the menu activated by clicking on the icon in the browser.

Cashback schemes and EasyPrivacy

written by Michael, on Jun 9, 2011 6:35:46 PM.

Problem:

Some cashback systems do not function correctly when EasyPrivacy is enabled.

Affects:
Cause:

Cashback websites work by either setting a third party cookie or directing people to a special address to inform the shopping website that the user is eligible for a discount. Shopping websites then need to be able to view the information to identify the user and reduce the item cost if applicable. However, EasyPrivacy is designed to prevent such details being collected; the shopping website therefore lacks the information to suggest that a special offer should be applied and charges to full price instead of the discounted rate.

Solution:

I have previously recommended, and would still advocate, that Adblock Plus should be disabled on important websites and when shopping. This will ensure that subscriptions do not interfere with the purchasing of goods and will therefore prevent issues from arising with cashback schemes. However, regardless of whether or not advert blocking software is installed, I would always recommend that you confirm the price of an item before paying.

DutchAdblockList and ABPindo updates

written by Michael, on Jun 6, 2011 6:06:00 PM.

DutchAdblockList, authored by Famlam, a contributor to Adblock for Chrome, has been a recognised subscription for a little over a year now. However, we have been impressed by the filter list, which caters for Dutch-language domains, and are therefore please to announce our affiliation with the subscription, which may be either viewed or installed from the homepage. Furthermore, after a review of the combination by myself and Fanboy, the subscription has been granted recommended status in Adblock Plus, which means that it will be the default selection for Dutch users.

If you want to provide feedback about DutchAdblockList, including by reporting adverts or problems, please use the Adblock Plus issue reporter or open an issue in the Google project page.

In other news, the affiliated subscription ABPindo, designed for Indonesian-language websites, has also been granted recommended status recently and now has its own section in the EasyList forums for feedback.