Hello guys,
I've noticed the main adblock lists have gotten pretty massive and I suppose there are many entries that are long out of use. I think it would be a good idea to implement some sort of statistics how often a filter is being used by the community and if an entry has been seldom used like, 1 time for two months, it can be regarded as outdated and deleted from the list. Of course, there could be an opt-out checkbox for the tinfoil-hat people, but I think this would be very useful to enhance browsing speeds. I'm already employing a similar strategy locally - regularly disabling filters that were last-executed long ago and it greatly sped up the plugin. I do a lot of browsing on my computer and on various sites, and still - more than 90% of the EasyList filters have never executed for the 2 years it's been installed on my computer.
List optimizations
-
adless-user
- Posts: 2
- Joined: Wed Sep 07, 2016 3:49 pm
Re: List optimizations
I read that already, but it appears to me as a too complicated solution to do the job, plus, works only with resource rules which are the lightest(in contrast with heavy dom rules). What I propose as a solution is:
- To have a new Introduced date column next to the Last executed, so we know the new rules. It can be generated locally on updating so not to bother the list formats
- Submit reports for rules that were never executed for more than X months, so a submitted rule is not being used, an used rule is not being submitted.
- If, say, a rule is submitted(not used for several months) by 99.99%(arbitrary), we can consider noone uses it and it can be removed by the next update
Of course, it could be made more complex, for example - to check whether an incoming update would delete a used filter for the particular user(0.01% corner case) and convert the filter to a local, temporary list.
- To have a new Introduced date column next to the Last executed, so we know the new rules. It can be generated locally on updating so not to bother the list formats
- Submit reports for rules that were never executed for more than X months, so a submitted rule is not being used, an used rule is not being submitted.
- If, say, a rule is submitted(not used for several months) by 99.99%(arbitrary), we can consider noone uses it and it can be removed by the next update
Of course, it could be made more complex, for example - to check whether an incoming update would delete a used filter for the particular user(0.01% corner case) and convert the filter to a local, temporary list.
Re: List optimizations
see also this ticket https://issues.adblockplus.org/ticket/495