multiuser knoppix configuration (other than root & knopp

[Albretch]

Hi *,

lately I have been toying with the idea of somehow having a multiuser knoppix configuration (especially to protect kids access to the net). Say, instead of basically having two users, 'root' and 'knoppix', I would like for kids' access to the net to be filtered/monitored and or restrict the websites they can access, etc.

I need this kind of functionality because at my school we are working on providing our students with (old, basically discarded ) computers and customized knoppix CDs, to be used by families

I have spent some time thinking about it and checking for knoppix customizations and don't know of anything like that.

I think two things I should look into would be:
1._ an editable Internet Access protocol for a functioning piece of code (kind of an application firewall + content filter (from what I see, Firefox(the browser) has something like that built-in (notice their search feature)))
2._ a way to set some 1.-based context for each “user type”
2.1_ user types should be preset and adjustable
2.2_ the Internet access watchdog feature should work regardless of their being 'root' or 'knoppix'
3._ this type of functionality must be somehow authenticated through
3.1_ a unique, burned ID in the CD (probably the 'Volume' so that the content is not altered)
3.2_ set and editable by the 'admin' person

How do you think this should/could be done?

Thanks
Albretch

[Albretch]

Here is what I have in mind (I haven't had time to dive into it yet, at least on my first attempt)

I work for a 'non-for-profit' school, where I am trying to push the use of open source. My boss raised the issue to me of children using knoppix to go online, bypassing any parental constrains ... (family literacy programs are very important)

So, I am thinking about:

1._ a knoppix based CD, with actually 'roles' instead of 'users'

2._ hacking the KDE libraries to parse and 'monitor' URLs and content going to or coming from the Internet (firefox does this itself)

2.1_ URL and content which are not ?good for children? should be logged and parse off

2.2_ a hash table keeps URLs ...

2.3_ an internal program using 2.2 retrieves 'URLs sh!t lists' and sends logs to the central server ...

3._ Optionally users may save their favorites (for konqueror, firefox and IE in case they use WINE in the back end server) I think I read once about an XML based standard Bookmarks format

4._ 3. should start working by itself once the user connects to the Internet. If not connected local files should be used

Thanks
Albretch