Web Security: "Phrase Weighting" for Content Filtering
Add phrase weighting to the content filtration much like what is used in Dansguardian, this would much improve the current content filtration. When used right it would allow pages within an otherwise approved domain or url such as YouTube to be blocked if inappropriate content is present. As is the case with Dansguardian the aggressiveness of this type of filtering can be adjusted quite a bit.
Since the inclusion of a locally stored filter database in 8.2, it seems far more practical now to run all searches against a list a weighted or banned words or phrases. Similar to Dansguardian with customizable values for each entry.
krell hunter commented
I also would pay for this feature...
this is a good start. I would rather see squid+DG fully implemented instead of the constantly problematic proprietary proxy in use now.
Isn't DG a public license?
Great idea I think this would hep alOt
good idea, definitely
I would also buy it, even if it was a literal lift of the dansguardian platform
I think it is open source as well, so cribbing off of the code would be easy
This should really not be that difficult to include; I would buy a license in a heartbeat if it was included.
Bastian Haas commented
Yes I know this feature from other filtering solutions.
You define a certain threshold and give each phrase a certain weight (positive OR negative!).
yes indeed, those phrase lets say thriple thing can be blocked and ofcource its very inapopriate.
Bob Alfson commented
Theoretically, TrustedSource can block individual pages, but the reality is that gazillions of pages will never be classified. In the USA, where a company can get sued if one person sees an unclothed person on a co-worker's computer, this capability would make a good addition to Astaro's marketability. This really shouldn't be that difficult to add.