suggestions/request for fsfe

[Paul Boddie]

On Tuesday 7. January 2020 15.17.12 Carmen Bianca Bakker wrote:
> 
> On a grander scale, I think a better solution would be that relying on
> a server is not necessary. Imagine instead that a distribution might
> include a `spyware-dns-hosts` package that modifies `/etc/hosts` with
> the same kind of blacklist that pi-hole uses.

I didn't address the alternatives to just delegating responsibility to 
services in my response, but it is certainly the case that localised solutions 
should also be in a position to be considered viable and usable. If they 
aren't then effort should be directed towards making them so.

I may have mentioned that I use a rather clumsy approach on a single-board 
computer than I also use as a "workstation", extracting host details from 
"excessive" Web sites and adding them to /etc/hosts. This could be improved 
substantially, and maybe I could learn from what the pi-hole project does, but 
it probably makes various Web sites usable for me on this fairly constained 
hardware already.

Having a curated blacklist does involve a certain amount of effort and 
collaboration, and it also brings certain responsibilities. Here, the FSFE 
would probably act as some kind of guarantor that various providers do things 
in acceptable ways, having clear policies and not just blocking stuff 
arbitrarily, for instance.

> Or maybe browsers could ship with much stronger privacy protection. I
> believe that Firefox is flirting with the idea of blocking more ads by
> default, but I'm not extremely well-read on that topic.

The problem with the Web and its de-facto custodians is that various business 
models rely on pervasive advertising, with some of that advert money being 
ploughed back into Web platform development. Consequently, the attitude that 
"amazing" things can now be done using Web technologies facilitates people 
using those technologies to provide overcomplicated and surveillance-heavy Web 
"experiences".

Of course, the people developing Web technology are always able to persuade 
themselves that they are merely "enriching" the platform or "empowering" the 
users. Never mind that what should be relatively simple online transactions 
turn out to involve tens or hundreds of megabytes of traffic, multiple data 
centres, and maximum CPU, for which the power has to come from somewhere.

Returning to the specific topic, there have been initiatives like FreedomBox 
that seek to make systems - Debian, in that particular case - more appliance-
like and requiring less end-user maintenance. However, FreedomBox seems to 
have taken a path where people do need to care about what it is doing, 
arguably following the same path as some other router-like distributions. 
Maybe pi-hole is more appliance-like, less prone to failure, and less 
susceptible to burdening the user with maintenance by dressing it up as 
customisation.

But as I noted in my previous message, having a broad and coherent strategy 
means providing "actionable" solutions, and if a solution such as pi-hole 
isn't practical then organisations like the FSFE should be looking to 
facilitate improvements in order to strengthen and deliver that strategy. 
Saying "Free Software is great" and then "you're on your own" is not a 
credible strategy.

Paul