[WikiEN-l] Baidupedia (the downside to free content?)

[Tim Starling]

Anthony DiPierro wrote:
> Whether or not a grossly censored Baidupedia is better than nothing, I
> don't know.  Depends, I suppose, how censored it is, but I get the
> impression that the censorship is much tighter than say the censorship
> of Google.  From what I know of the situation I'd guess a censored
> version of Wikipedia is worse than nothing, because at least with
> nothing more people will go through the technical hoops to get around
> the firewall.
> 
> If anyone has any ideas as to what we can do to help get the real
> Wikipedia to the masses in China (no client-side setup required), I'd
> love to help.  Maybe some sort of network of distributed servers
> providing https access through dynamically rotating IP addresses.

I don't think there's any solution left which will work without client
configuration, except for a number of loopholes that the Chinese Government
hasn't gotten around to closing yet. The foremost among these is our own SSL
gateway:

https://secure.wikimedia.org/wikipedia/zh/wiki/

There are various unblocked HTTP proxies, although all unencrypted traffic
is sampled, so any popular proxy will be rapidly blocked.

Traffic within China isn't subject to the Great Firewall, which is why a
proxy like wikipedia.cnblog.org worked. More proxies like that could be set
up, but our recent experience suggests that the Government is watching for
such things, and you can fully expect a knock on your door if you set one up.

Periodically changing the IPs returned by a specific DNS entry almost
certainly won't work. They have the ability to poll DNS.

So that leaves client configuration. The Tor network is still not blocked,
but that might be only a matter of time. And the downside is that it has no
system for dealing with abuse.

Having numerous SSL tunnel servers would be useful, along the lines of
Anthony's suggestion. However, you need to have some way to distribute the
server IPs to the users without letting the Government find them out. I
can't think of any way to do this with a public protocol without leading to
a very high rate of compromise of IP addresses, assuming the authorities are
on the ball. A simple HTTPS gateway like secure.wikimedia.org could be
compromised automatically by simply connecting to it and downloading the
index page.

We can go on exploiting known holes in the firewall for the time being, but
it will certainly become increasingly difficult for people inside China to
access Wikipedia, especially for those who are non-technical or not
especially motivated.

If Baidupedia does take off, I hope they will license locally generated
content under GFDL, to allow for a continuing exchange of content between
Wikipedia and Baidupedia.

-- Tim Starling