boballab wrote:See Amazon.com would get banned because they sell autobiographies that in some of them descripe just what Spec listed.
(snip)
The only way to get them is to use a very wide range of key words to Ban the site. Something a Computer can do after a program is written form it. It would just log on from site to site scan for banned words and move on. Once the conditions are met to Ban a site it automatically copies the URL to the Ban list.
I doubt Amazon would get banned - however it may be required to enforce Office of Film and Literature Classification* standards and not sell material that was Refused Classification* to Australians.
With regards to the keyword filter - they're not talking about pro-actively crawling the web and blacklisting sites using some spider software (although, that may be done also), they're talking about when you type in a URL into your web-browser, or click on a link and your browser requests a page - the filter would kick in, and scan the page as it was being delivered to you. There's other steps before the dynamic filter (like checking against the existing blacklists), but that's one of the killers.
Now, for non-tech savvy folk it might seem like an okay task to set up a bunch of beefy servers that filter this content. After all - if I can install filter software on my home PC and it makes little or no noticable difference to a web browsing experience, then it can't be a particularly onorous task. And that's true, to a certain extent.
What makes it such a difficult task is that you'll have a huge list of words to scan for - and you'll need to scan every single request that comes through - not just 'Give me
http://www.example.com' but also 'Give me
http://www.example.com/image.jpg' and 'Give me
http://www.example.com/script.js'. If you're using Firefox, right click on a page and go View Page Info, then click on the Media tab. On this site alone there's like 30 images - each one of those would need to be checked - sure it might be an image, but the filter will still have all those requests pass through it - even if it does absolutely nothing to it. Add to that probably half a dozen style sheets and scripts that make web pages work, and most sites will weigh in at about 50-100 individual requests, just to service one page.
Of course, people browsing the web is only part of the traffic that would be monitored by the filters - Many many applications exchange data over HTTP (the network protocol that delivers all website content to your browser). Take for example those applications that let you upload photos to Flickr - they speak HTTP. Got a widget that shows the current weather or stock prices on your computer? Speaks HTTP. How about your computer automatically checking for updates. Also HTTP. RSS Reader? HTTP. Game console with an internet connection? HTTP. Use iTunes? HTTP. What about an iPod Touch / iPhone? Lots of HTTP. Media Centre or TV that downloads an EPG? Yet again, speaks HTTP.
The vast majority of applications that you use to do things online are speaking HTTP - because it's easy and relatively efficient at what it does. All of those applications and devices that speak HTTP will generate requests that will be scanned by the filter.
I think you can see where this is heading, right? There's an aboslute mountain of requests every second, many of which are quite legitimately requesting or sending content over the internet. All of which need to be scanned.
For a large ISP to scan millions of requests per second at anything close to real-time, they would need hundreds, if not thousands of very expensive servers, consuming huge amounts of power to run. The Government sure isn't going to pay for that, so what'll happen? ISPs will put costs up. Because the filtering can't happen without you noticing a slowdown, internet access speeds will drop - most noticably during peak times.
So, we'll have more expensive and slower internet access which will semi-randomly block access to legitimate sites because it matched a few wrong keywords for what? Anyone wanting to access something illegal isn't going to be hampered, they're using either Encrypted HTTP (Slower than regular HTTP, but faster than filtered HTTP) which by it's nature can't be filtered, unless the government is willing to completely kill all online commerce**.
The sum total use of the ISP level filter is a 'feel good' blanket for concerned parents, and won't do anything to stop pornography of any kind being distributed or accessed on the web.
One of the nastier side effects of this filter that really SHOULD have parents worried is that anyone that IS doing something illegal (say, distributing child pornography) will be driven to using more advanced encryption to bypass any filtering.
Thus attempts by the Australian Federal Police to perform an "internet tap" will only get to see an encrypted stream of communications.
* = OFLC is the Australian Government department responsible for applying classification ratings to books, videos, films, magazines, tv, video games, etc in Australia. They have a bunch of ratings - however material that is Refused Classification is not permitted to be sold or distributed within Australia. The OFLC's standards also apply to Internet sites.
** If you're wondering what I mean, look up
Man in the Middle Attack - that's how they'd break SSL to filter it - by intercepting communications prior to the channel being established. All modern browsers, and anything else that's well developed and is using SSL will see that the connection isn't secured correctly, and refuse to let you continue. SSL Encryption is what makes Internet Banking and Online Shopping possible, by protecting your confidential details as they're being sent to the server. If a filter were able to perform a MITM attack, then there'd be nothing stopping Mary Malicious Attacker from doing something very similar, and capturing all your details.