GitHub considered harmful

[Michael Gratton]

On Fri, 11 Oct, 2019 at 10:57, Alexander Larsson <alexl at redhat.com> 
wrote:
> 
> Its already costing them money. They did a $500k donation to "offset"
> the $200k causing it to be a net loss deal, and thats not even
> considering all the bad PR they are already getting over this from
> various sources.

Yes, but they still haven't cancelled the contract, so the response so 
far clearly hasn't been enough.


>>  You see how effective this is when (say) bigoted news readers say
>>  something particularly heinous, such that people start contacting 
>> their
>>  advertisers and getting them to cancel ad their ad spend the show -
>>  there's always a quick reversal.
> 
> Well, cancelling ads is very easy to do, and easy to reverse. A large
> infrastructure reboot is not that easy, so it has to be considered
> much more carefully.

Sure, and that's why I'm bringing it up, so it can be considered.


> And there is also the question of what guarantees that the new host is
> any better (see the link to gitllab policy in my other mail). In fact,
> am I any better? I'm currently employed by the company that literally
> tabulated the holocaust...

Right, which is why I suggested hosting on FDO's GitLab. It is much 
less hassle for people using the service to move a self-hosted service 
to a different host than it is to shift platforms entirely.


> I find what ICE is doing reprehensible, and I would prefer if github
> dropped ICE as a customer. But I realize the complexities of business,
> and this contract is imho not something fundamental or defining of the
> github organization either in terms of money or in moral values, nor
> is it (I assume) a major thing for the ICE operational capabilities. I
> also think that flatpak moving elsewhere will not really add a
> significant effect to this particular case as this is already being
> fought in a lot of ways.

As before, I disagree with all of the above. If significant projects 
like Flat* migrate away from companies that are supporting ICE 
operationally (in whatever capacity), and do so in a loud way such that 
it's clear why they are moving, it puts an enormous amount of pressure 
on these companies to drop that support. Further, it discourages other 
companies from engaging with organisations like ICE in the future, 
having a positive flow-on effect.

This sort of pressure works exceptionally well in the real world when 
enough people care and buy in. For example, the 2014 Biennale of 
Sydney, a major arts and cultural event of national significance in 
Australia, was sponsored by Transfield, a company that was running the 
Australian Federal government's offshore refugee concentration camps. 
The Biennale's board was asked to drop the sponsorship, but they 
refused. So participating artists and many others organised and began 
to boycott the event. As the momentum of the boycott picked up, the 
board relented and dropped the sponsorship, and later as a direct 
result many other organisations did the same. It worked *very* well, 
because enough people bought into it. In the end, Transfield stopped 
running Australia's concentration camps, changed its name, and ended up 
being bought out.

So I'm asking the project here to buy in, so that the existing pressure 
and momentum can be kept up. It's inconvenient, but it's the right 
thing to do: *Actual human beings are being harmed.*

So, for a concrete plan, how about this:

1. Move development of Flatpak itself, the portals and related projects 
(but not FlatHub app reops) to FDO's GitLab, provide some documentation 
for people developing for Flatpak about how to migrate, update existing 
documentation, announce the move and the reasons for it widely so that 
people hear about it, and set the projects on Github to be read-only.

The cost of this is low, since as Mathieu pointed out, the 
repo/ticket/PR migrations can be easily done using existing tools, 
people can log in to GitLab using their GitHub credentials, and 
developers simply need to update their origin, and the number of people 
actively developing for it are low.

2. Start implementing support for GitLab (and other services?) in the 
Flathub toolchain, so that Flathub-hosted apps can use FDO's GitLab, 
the commercial GitLab service, or their others (GNOME's? KDE's?) as 
they see fit.

3. Once that's in place, move any Flathub apps that don't opt out to 
FDO GitLab, or elsewhere as appropriate.

This would be low cost, high impact, and minimally disruptive.

//Mike

-- 
⊨ Michael Gratton, Percept Wrangler.
⚙ <http://mjog.vee.net/>