gnunet-developers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] FPS paper & browser study


From: Christian Grothoff
Subject: Re: [GNUnet-developers] FPS paper & browser study
Date: Mon, 09 Sep 2013 19:32:16 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130827 Icedove/17.0.8

Hi Ludo,

As you can verify by running our script (before & after doing a search) from

https://gnunet.org/gnsdatacollect

it is indeed the case that following a search result like this would be
counted as following a link (and thus not be in the 8%).  We cannot tell
from the data what percentage of followed links were from search engines.

However, I do not see this as a problem.  GNS will work fine with search
engines: once you've gotten "search.gnu" in your zone, you will resolve
links from the search engine for say a search for "ludu" using 
"ludo.search.gnu",
which would go to Ludo's zone (where "Ludo's zone" is defined by the
search engine, which is fine --- you will get the Ludo corresponding to
the search result you clicked on).

Now, we might not like people exposing their browsing habits like this from
a privacy perspective, but that's another story.  And obviously the 8% is
given on a limited sample for today's behavior; how people may evolve to
behave tomorrow is another story.  I mean, sample the percentage of encrypted
e-mails you got two months ago vs. today... Sometimes these values change,
and if you need to import a public key into a zone, I can imagine that users
may change their surfing behavior in a way that reduces the 8% further.


Happy hacking!

Christian


On 09/09/2013 07:16 PM, Ludovic Courtès wrote:
> Hello,
> 
> Congratulations on the nice FPS paper!
> 
> Section 4.5 states:
> 
>   Based on a limited and most likely biased survey where users volun-
>   teered the output of a simple shell script that inspected their browsers
>   history database, we determined that given current Internet behavior,
>   approximately 8% of domain names would require introduction via some
>   out-of-band exchange.
> 
> I think many users would just type (say) “gnunet” in the address/search
> bar of their browser (rather than “http://gnunet.org”), which leads them
> to a Google search result page.
> 
> In that case, ‘gnunet.org’ is counted as not needing an out-of-band
> introduction, I suppose.  However, that does not capture the fact that
> Google is an undesirable way to get introduced to a web site from a
> censorship-resistance viewpoint.
> 
> Does your study account for such uses somehow?  What impact would it
> have on the 8% figures?
> 
> Thanks,
> Ludo’.
> 
> _______________________________________________
> GNUnet-developers mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/gnunet-developers
> 




reply via email to

[Prev in Thread] Current Thread [Next in Thread]