MacPorts Statistics (was Re: usage numbers for macports vs. homebrew?)

[Clemens Lang]

Dear Marko

> first of all I apologise if my words were too harsh. It was not my intention
> to attack you or anyone else on a personal level!

Apology accepted and thank you for your kind reaction.

> I am just concerned about everyone’s privacy.

I think you were in fact asking the right questions such as who gets access to
the statistics data and how personalized the data actually is. I feel your
questions have improved the statistics setup and our policies for dealing with
the data.

> And I admit my comparison was perhaps a bit over the top, but the revelations
> of the past year have made me a bit nervous regarding the privacy topic.

I can totally relate to that and I agree this is an important topic.

> I am sorry that it came across as inappropriate, but this post regarding
> handing over the database:
> —
> > 	I'd take a copy.  Maybe you could just put it on DropBox or something
> > 	similar?
> —
> made me shudder because I thought we’d distribute the database with all
> information including UUID to anyone who can work dropbox.
> Does that not sound odd?

I agree. I was not going to distribute the database using a public download
link, especially not with any personal data in it and to non-committers. I
think as a result from the discussion we can record that no user-generated
data should leave the statistics server in non-aggregated form. This is what
I have done so far and I think it's a reasonable thing to do in the future
aswell.

I will provide (and have provided) a tarball with the database schema (that's
generated from the ruby code and open source anyway) and the data from the
 - categories,
 - ports, and
 - schema_migrations
tables, which is information generated from the rails database code and the
PortIndex. This might be useful to people working on the statistics code b/c
it isn't trivial to setup the scripts that will populate these tables from
the PortIndex (e.g. doing this on a Linux server as I do requires a source
installation of MacPorts on Linux, which is tricky).
Data from the
 - installed_ports,
 - os_statistics,
 - submissions, and
 - users
tables will stay on my server only.

> I was - from the beginning on - actually always PRO this feature, i.e. I DO
> SEE the advantage to have such a system in place. If you go back to those
> discussion on the list you will see that. So, I am not against this great
> feature at all.

Sorry, I didn't go back to re-read the archives on that -- I probably should
have instead of accusing you out of the blue.

So, let's get this matter behind us and back to improving MacPorts, which is
what we're all here for. :-)

-- 
Clemens Lang