missed ad

Guest · Post by **Guest** » Sat Feb 11, 2006 1:30 am

Filter: "http://www.firstload.de/Affiliate/"

Go to http://www.fixx.de

Expected result: banner on top of the page should have been blocked, because it points to http://www.fixx.de/out.php?url=[b]http: ... Affiliate/[/b]log.php?log=3987

What really happens? Ad is not blocked.
When you delete the beginning ( http:// ) of the filter, the ad will be blocked correctly.

pirlouy · Post by **pirlouy** » Sat Feb 11, 2006 3:05 am

Like you don't use a regxp filter, the search begin at the first letter.

So, http://www.firstload.de/Affiliate/ is not found in http://www.fixx.de/out.php?url=http://www.firstload.de/Affiliate/log.php?log=3987

If you add an * before: *http://www.firstload.de/Affiliate/ , then it will work.

However, I don't understand very well why www.firstload.de/Affiliate/ manages to block...

Paulfox · Post by **Paulfox** » Sat Feb 11, 2006 9:58 am

Something as simple as firstload.de will work. Just for information purposes, the biggest ad companies with German .de endings (now including this one you stated, which I didn't have up until now!) are:

adtech.de
affiliatecash.de
deluxelink.de
firstload.de
gonamic.de
ivwbox.de
jambal.de
partnersuche.welt.de
pay4klick.de
pearl.de
popdowncash.de
popexchange.de
ptadsrv.de
red.ads.t-online.de
superclix.de
tfag.de
view4cash.de
webmaster24.de
zanox-affiliate.de

and the regex I was using was:
/(adtech|affiliatecash|allsponsor|deluxelink|firstload|gonamic|ivwbox|jamba|partnersuche\.welt|pay4klick|
pearl|pop(downcash|exchange)|ptadsrv|red\.ads\.t\-online|superclix|tfag|view4cash|webmaster24|zanox-affiliate)\.de/

that is broken into 2 lines for forum width, but entered as one line filter.

Running them through the Optimizer, we get a more "efficient" but more complicated version than my filter:

/((popexchang|red\.ads\.t-onlin|zanox-affiliat)e|((affiliate|popdown|view4)cas|adtec)h|(ivwbo|supercli)x|p(a(rtnersuche\.welt|y4klick)|
earl|tadsrv)|deluxelink|firstload|gonamic|jambal|tfag|webmaster24)\.de/

but mine is a little easier to read and alphabetized, which makes it easier to add "firstload" even if it does "waste" a few characters. Also, mine's TESTED; I ran those companies through the optimizer to produce the 2nd one above but have not tested it (I'm sure it works, although looks "messy.") If you visit many German sites, either one of the above filters will "nail most of 'em" in terms of ads. I stay away from "http" on filters as it's redundant and "wastes characters" which makes the list a lot longer, especially over time. If that site you mention is the only .de site you visit, then simply firstload.de or /firstload\.de/ would work fine and obviously be a lot shorter & simpler.

NJH · Post by **NJH** » Sat Feb 11, 2006 12:07 pm

pirlouy wrote:Like you don't use a regxp filter, the search begin at the first letter.

So, http://www.firstload.de/Affiliate/ is not found in http://www.fixx.de/out.php?url=http://www.firstload.de/Affiliate/log.php?log=3987

If you add an * before: *http://www.firstload.de/Affiliate/ , then it will work.

However, I don't understand very well why www.firstload.de/Affiliate/ manages to block...

I would be interested in WP's comment here. In the old Adblock versions a leading wildcard was never needed in filters. The search looked for an occurrence anywhere in the string. This would appear to be a major change in how filters work. I do, however, remember a thread where rue said filters beginning http:// were quicker at matching than other filters, which I could not understand at the time. Perhaps this has something to do with it and it is a carry over which I never understood and needs some explaining.

Paulfox · Post by **Paulfox** » Sat Feb 11, 2006 12:16 pm

NJH wrote:The search looked for an occurrence anywhere in the string. This would appear to be a major change in how filters work. I do, however, remember a thread where rue said filters beginning http:// were quicker at matching than other filters, which I could not understand at the time. Perhaps this has something to do with it and it is a carry over which I never understood and needs some explaining.

The actual properties of that ad mentioned in OP are:
http://pics.firstload.de/123.gif

NJH: I hope I understand you correctly. I think YOU are right that the search looks for anything in the string, which is why just "firstload.de" would work. I can only offer that starting all filters with "http://" would NOT speed anything up because 112 filters all starting with http:// seems ridiculous. So I'm with you - I can't understand that either. Several ".de" sites (or any that share similarities) can be combined and preferably in alphabetical order(ing), and that has to be more efficient AND more editable later on when additional sites fitting that "category" can be found. There is something to be said for the "aesthetics of the list," i.e. how quickly it can be amended "on the fly" without spending hours trying to find the right filter to append or having a list of single blocks 3 miles long (I'm talking about "most" users," who might just right-click block anything they see). I cannot see using http:// for single sites or incorporating it in a multi=site, multi application filter.

It's putting in "/affiliate" that skips the filter - that isn't in the properties to be blocked. The ad company is firstload. So block 'em! It has nothing to do with "affiliates."

pirlouy · Post by **pirlouy** » Sat Feb 11, 2006 12:54 pm

NJH wrote:remember a thread where rue said filters beginning http:// were quicker at matching than other filters, which I could not understand at the time. Perhaps this has something to do with it and it is a carry over which I never understood and needs some explaining.

Yes, waiting an answer for Wladimir but I think you're right.

http:// is different apparently: it is searched only at the beginning.

@Paulfox: you can also block this ad with a regxp for "affiliat".

Paulfox · Post by **Paulfox** » Sat Feb 11, 2006 8:30 pm

Sure! or "de" for that matter, but then you'd break the browser cause no site in Germany would come up. Blocking "affiliate" is quite intriguing actually - I wonder how many false positives you might get vs GREAT results - I'll check that out. I'm about to add a stylish script for "iframes." without breaking them down to specific "ad types," on the same principle - an "affiliate" probably means ad-related. What happens by the way if you block ALL iframes? Is anything useful ever in an iframe?

Back on topic: Im saying pick the basic culprit, organize "like culprits" in order, and move on to finding more ads to block.

For example: You might have a situation where there are several subsidiaries of ad "affiliates" under one roof; such as
firstload.pics pics.firstload, ad.firstload, and firstload.aggravating.readme.

Yahoo's infamous for this. Each, let's say, has 40 ads running if we could magically go to every web page in the world and find them all RIGHT NOW. That's 4 x 40 or 160 ads. Now imagine right clicking all of those - you'd have 160 http's for ONE ad company. NOW multiply that by the 19 culprits I mention in my list above. Roughly 3,200. Talk about a confusing, inefficient list. But firstload.de would get all for that culprit, AND - MY REGEX gets 'em all! Done. AND - be incorprated either in an "F" list of like "F" culprits, or ".de" for a general "German site" List, etc.

There's textbook pooh-poohing of regex style and theoretical argument, and then there's ONE damn regex nailing thousands of ads. Which one do you want? Theory or results?

Wladimir Palant · Post by **Wladimir Palant** » Sun Feb 12, 2006 12:55 pm

Ok, so: the behavior here changed a little with Adblock Plus 0.6. Any filters starting with http:// or https:// will only match at the beginning of the address. If you wish them to match anywhere, you should write *http://

Guest · Post by **Guest** » Sun Feb 12, 2006 5:31 pm

perhaps this would be worth for the faq; there you could also describe regex and show how adblock is working and which filters are better (regex <-> simple) etc.

pirlouy · Post by **pirlouy** » Mon Feb 13, 2006 12:10 am

Paulfox wrote:Sure! or "de" for that matter, but then you'd break the ...
.................................. Theory or results?

Sorry, I'm not english-native, and you use terms I didn't understand...

I've understood than you consider "affiliat" as a bad filter because of a lot of "FALSE-positive". I don't know... Using this filter since at least 1 year, I've never had problems. But I don't go on deutch site like I do not understand ONE word.

Anyway, I didn't say "firstload.de" was a bad filter... I just propose an alternative...

Guest · Post by **Guest** » Mon Feb 13, 2006 2:43 am

@pirlouy: No no - your filter may be GREAT - my argument is that something like yours (affiliat) or a general ".de" category, are both better than singly naming all terms separately. I may actually try "affiliate" out!

No disagreement with you at all Brother - I was simply saying that filterlists canget very long very fast unless you use terms such as "general regex for that category," or indeed YOUR "affiliate," or something of that nature.

Cheers/

Guest · Post by **Guest** » Mon Feb 13, 2006 2:55 am

Guest · Post by **Guest** » Mon Feb 13, 2006 3:32 am

Wladimir Palant · Post by **Wladimir Palant** » Mon Feb 13, 2006 5:37 pm

Starting with ABP 0.6.2 the first version (every domain listed for itself) will be much more efficient than the second (everything put into the same regexp). Just a note...

pirlouy · Post by **pirlouy** » Tue Feb 14, 2006 1:45 am

What about this filter ?

/http://([^/]+\.)?(a(1\.yimg|ddfreestats|llosponsor)|b(aventures|luestreak)|c(asalemedia|omclick)|daooda|e(stat|uroclick)|fa(lkag|stclick)|gestionpub...)[./:]/

I use special beginning (http://(...)?) for better efficiency. Can you confirm that with your changes, if I divide all domains in several filters, it will be more efficient ?

Adblock Plus

missed ad

missed ad

Re: missed ad

Re: missed ad

Re: missed ad

Re: missed ad