Reply to thread

Message

<blockquote data-quote="Decopi" data-source="post: 725754" data-attributes="member: 67091">[USER=51905]@Evjl&#039;s Rain[/USER] ... I tried to emulate your test (as much as possible). I tried to use your hosts for UBlock, and your phishing/malware sources (I took from your first comment in this post, describing your test). I used 10 samples for each source. I mixed the samples, taking a few from newest on the list, other from the middle, and others from the end of the list (on same day).But again [USER=51905]@Evjl&#039;s Rain[/USER] , this was a very sloppy quick informal test from my side. I need to run something more formal, perhaps on next week.At first sight, I have some unproven conclusions:1) Test methodology need to be improved (at least from my side).2) It does not exist the perfect solution, or the perfect add-on etc. A combination of tools seems to be better.3) Tools/Add-ons must be tested considering their strengths. For example, Netcraft x Safebrowsing for phishing, Norton x Bitdefender for malwares etc. I really don&#039;t see much sense in comparing everything with everything.4) Also, tools/add-ons must be tested considering &quot;zero-day&quot; samples x 1 day old samples... because test results are dramatically different even in same source5) Browsers must be updated, and also add-ons must be with data-bases updated. This is not so simple to be done. An add-on data-base not updated, will show very different test results.6) Etc</blockquote>

[QUOTE="Decopi, post: 725754, member: 67091"] [USER=51905]@Evjl's Rain[/USER] ... I tried to emulate your test (as much as possible). I tried to use your hosts for UBlock, and your phishing/malware sources (I took from your first comment in this post, describing your test). I used 10 samples for each source. I mixed the samples, taking a few from newest on the list, other from the middle, and others from the end of the list (on same day). But again [USER=51905]@Evjl's Rain[/USER] , this was a very sloppy quick informal test from my side. I need to run something more formal, perhaps on next week. At first sight, I have some unproven conclusions: 1) Test methodology need to be improved (at least from my side). 2) It does not exist the perfect solution, or the perfect add-on etc. A combination of tools seems to be better. 3) Tools/Add-ons must be tested considering their strengths. For example, Netcraft x Safebrowsing for phishing, Norton x Bitdefender for malwares etc. I really don't see much sense in comparing everything with everything. 4) Also, tools/add-ons must be tested considering "zero-day" samples x 1 day old samples... because test results are dramatically different even in same source 5) Browsers must be updated, and also add-ons must be with data-bases updated. This is not so simple to be done. An add-on data-base not updated, will show very different test results. 6) Etc [/QUOTE]

Verification