Re: In support of Jonas Bernoulli's Magit

[phillip.lord@russet.org.uk (Phillip Lord)]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > which might warrant
>   > improving the copyright assignment mechanisms.
>
> I don't think the mechanism, as such, is much of an inconvenience any
> more.  Isn't all the commnication is done digitally now?

Only if you live in Italy, Germany, India or the US.

At least according to

https://www.fsf.org/blogs/licensing/fsf-to-begin-accepting-scanned-signatures-for-copyright-assignments-from-india

This post also says the full list is in information for maintainers

https://www.gnu.org/prep/maintain/maintain.html#Copyright-Papers

Although, this only mentions Italy and the US.

> It could be improved, but that's like speeding up the fast parts
> of a program -- not going to make a big difference.
>
> The substantial part of the job is finding those people and asking
> them to please sign.  But it's not terribly hard.

It is terribly hard. The process consists of this:

1) Get list of all contributors
2) Email them all, recording the date
3) Email those who respond instructions
4) Email assign@fsf to find out current status
5) Email those who havent again
6) Reply to people who say "haven't heard anything"
7) Repeat some variation of 4,5, and 6

8) Work out how many are left
9) Once n < 3 or 4 check how big their contributions are
10) Write them out

Now, for dash, this wasn't too bad (of the 40 contributors I had to get
about 10 through the process, and do only one write out). A quick check
of my mailbox suggests it took around 100 emails over the course of
three months. Not too bad, but still painful. Magit has 4-5x as many
contributors, plus dependencies. Assuming linear scaling, we are looking
at ~1000 emails.

So, we can all bitch about the process; here is a suggestion to improve
things using current software that we already use -- debbugs, installed
with non-public archives which I presume it can do.

1) Get list of all contributors/contributor emails
2) Email assign with full list to find out who needs doing
3) Email each person with new debugs ticket
4) One of:
     assign@fsf cc's electronic communication to debbugs
     assign@fsf acknowledges paper communication *in either direction*
     to debbugs. assign would probably need to attach an ID number to
     outgoing assignment forms.

This would allow tracking the status of the current situation with
respect to each author in need of assignment. It would be more work for
assign@fsf, but would reduce their need to respond to "do we have
assignment for x yet" emails. It would work imperfectly: debbugs is
painful; not everyone would keep debbugs in the CC; and debbugs has a
X-Debbugs-CC field, which means the initial request would be CC'd to the
person when really you want X-Debbugs-TO. But it might work okay. And it
would mean that anyone could take on the task for a given package.

If you will get someone to set up debbugs to support this, help with
using it, and get assign@gnu to buy into the process, I will test it out
with the new assignments need to update dash. If you and GNU are bought
into improving the process, we can refine the details overtime.

If you think "it's not terribly hard", that's probably because you're a
very driven individual who has managed to achieve much that I would have
thought is hard to the point of impossibility. I'm not.

Phil