As promised, I've had a trawl through old threads about opt-in porn filtering. I hope it's OK to re-post this here, I'm not sure Kaloki's around any more but the post below clearly explains the problems with this proposal. If you're still reading, Kaloki, thanks once again for your excellent post ...
------------
KalokiMallow Thu 23-Dec-10 18:59:06
I apologise in advance, this is probably going to be a long post, but if you are interested in the technical reasons why filtering porn cannot work, then you really should read it.
Anatomy of a website
Domain
The domain is the address you type in.
Eg. www.mumsnet.com
This is actually separate to the website itself, which is why you can have multiple domain names pointed to the same site.
Say you wanted to block www.mumsnet.com you wouldn't necessarily be blocking access to the website itself. As they could then just set up www.mumsnetisback.com without having to change where it is hosted or reload content.
You also could block domain names with keywords in the title, say you blocked "mumsnet", however this wouldn't block www.mummsnet.com - which could easily be pointed at the same site.
So that wouldn't work for filtering websites, too easy to get around.
Hosting
A website is hosted on rented or bought server space, you could block the IP address for a server, which is what the domain name points to. This is about the only way to block websites, but requires you individually blocking each website.
An ISP provides access to the internet, whether for a user or server, some have their own servers which they host sites on. But not all do. Hosting and providing net access are two different things.
URL
Essentially the same as the domain name, but with directions to specific pages or files.
Eg. www.mumsnet.com/Talk
You could block specific pages within a website, either by keyword or knowing the address. However you'd have to somehow take into account embedded information;
The web standard style of coding websites nowadays usually runs along the lines of;
Main page
¬ Header
¬ Content
¬ Footer
So while you may have blocked "main page", you haven't blocked "header", "content" and "footer", and if someone was to direct link to one of them..
Meta Data
Hidden information coded into webpages, usually keywords and a description. Not all sites bother with this though.
So although you could search the meta data for keywords and block pages where the meta keywords are to be filtered, if someone hasn't entered meta data then the computer will have no way of knowing.
HTML/Coding
Keywords could also be in the coding. This will include the text you see on the page. You could block pages with blocked keywords, however, read on to see why that is flawed.
Images
There are only two way to block images.
- Block any images that are inserted using the code. Which will block 99% of images. 99% of all images that is. Including the MN logo at the top of your page.
- Block images with filtered keywords, but this has the same problem as meta data, it doesn't have to be filled in. And the image file could be 111111.jpg. With no keyword data, you have no clue of that is a pornographic image or a pretty little flower.
There is no technology that exists that can identify what an unlabelled, generically titled image is either.
Embedded Media
Exactly the same as Images. This covers embedded video, interactive flash and audio.
Different filtering methods
Keywords
You could block keywords. But what keywords would you block for porn?
Penis? Vagina? Breast?
- there goes any website that mentions anatomy in any way, say medical websites...
Pussy?
- so no personal pages about Ginger the cat.
Porn? Erotica?
- There goes this page, and a large part of the feminism forum.
Do any of you have spam filters on your email? Does that work all of the time?
Do you ever get emails asking if you'd like to purchase v1agra?
That's the other way to get round keyword filters, just type things wrong or leave sp aces in them. You could even add in sym|3ols.
Also, if you wanted text but didn't want it searchable then you'd just use an image file with the text on and not label the image file. Easy.
So to summarise. You can block individual IP addresses, one by one - but hang on, how would you implement this? Who'd decide? Do you create a central agency to decide? Or do you leave it up to the public to report?
If you leave it up to the public do you immediately filter any reported sites to look at when someone gets a chance? Or do you wait? Could get a few complaints when sites aren't removed promptly enough? Say you decide that you will only filter a site when it has a set number of complaints? What happens when a large group of internet users decide it would be amusing to all report one site at once?
And would you block the page with the content, or the whole site? What if someone posted a pornographic picture on flickr, do you filter all of flickr?
If you use keywords to block sites then you will block innocent websites.
There's also a flaw with opting in. Say it's a family network, (because remember, the ISP's can't distinguish between separate computers) and one person wants to opt in (possibly because the filter system has blocked a website which isn't actually pornographic) then how do you protect the children?? Oh yeah, a computer based filter. One which can be set up not only for individual computers, but also for individual user accounts on one computer.
And you can set it up to block or not block exactly what you want! How exciting!
---------------
More very long threads on this topic
here and
here