On 04/06/2018 16:14, Rick Moen wrote:
> Quoting Mark Rousell (mark.rousell@???):
>
>> To play devil's advocate, what can go wrong?
> Autonomy.

I really can't see anything different about project autonomy before or
after GitHub acquisition by Microsoft. The issue you elaborate on below
is certainly real but Microsoft's acquisition of GitHub doesn't seem
likely to make any difference to it.

> The Void Linux distribution is, according to Steve Litt, currently in
> semi-chaos because GitHub, Inc. recognises no other user other than the
> founder as having authority over the project's data. If Void Linux
> controlled its own software infrastructure with appropriate
> collaborative arrangements, this administrative hurdle would be easy to
> solve.

Well yes, but this is not GitHub's fault (nor would it be Microsoft's
fault). It seems to me that if Void Linux had ensured that multiple
users had admin access to the project's data on GitHub OR if they had
kept regular off-GitHub backups then none of this would be a problem.

> As long as they outsource software infrastructure, that infrastructure
> runs according to someone else's rules. Personally, I want my software
> infrastructure to implement _my_ policies and follow _my_ rules.

Oh I agree! This is a potential problem with *all* types of cloud
hosting, isn't it. What happens if the provider disappears or if the
person with the account username and password disappears?

All users of cloud services, be it web hosting, data storage,
processing, or code hosting should ensure that (a) they have
well-distributed admin access so that the project never relies on any
single person to access the online data and (b) that they have a
contingency plan/disaster recovery plan in case they need to move hosting.

And yes, I suspect most users of cloud services don't think through
these risks well enough. But Microsoft's acquisition of GitHub won't
make this any more of a potential risk than it is already.

Personally I dislike it when people, businesses, or projects move almost
everything to the cloud. It really is overly risky to my mind (for all
sorts of reasons in addition to the two I mentioned above). But I can't
fight how things are: Cloud hosting brings many, many benefits and so I
always caution people never to rely solely on cloud services: They
should keep their own backups on their own physical infrastructure, they
should ensure that there is no single human point of failure, and having
to move hosting elsewhere (including bringing it in house) should be
part of their general disaster recovery plans in case their chosen
service providers cease to be effective.

--
Mark Rousell