[fab] Re: splash page + tracking image example

[Greg DeKoenigsberg]

Apologies in advance for a very long-winded email.

First, let me first offer a personal apology about this whole tracker
business.  I know that a lot of people here disagree with both the idea of
a tracker, and the method that's being used to push it through.  The
hardworking volunteers on this team are being asked to do something
unpalatable, at the last minute, outside of the team's process, without
the benefit of real debate.  Generally speaking, it's pretty much the
opposite of how decisions should be taken in the community.

That sucks.  As the person who is strongly advocating this solution, it
makes me feel like a jackass.  Generally speaking, I don't like to feel
like a jackass.  In almost all cases, "Greg feels like jackass" is usually
a strong signal that "Greg is wrong".  In this case, though, "Greg feels
like a jackass" makes me feel more like "Greg probably went about things
the wrong way and needs to explain himself, but Greg is still fighting
hard to do the right thing."

So let me explain.  We've still got some time to discuss "the right
thing".  It's not much time, but at least I'll feel like I made an honest
attempt.

I also want to make this very clear: I admire and respect this community
and its leadership.  I consider many of you to be longstanding friends.  
I am still at Red Hat because I believe in the importance of the work that 
we all do together.  I hope you all understand that, even if I advocate a 
position that you strongly disagree with, I do it because I believe that 
it will benefit Fedora's users and contributors in the long run.


1. WHY ARE METRICS SO IMPORTANT?

Really, this question should be asked this way: "are metrics so important
that you're ready to risk alienating some users and contributors to get
them?"  And the answer to that question, from my perspective, is "yes".

Why?  Because, like it or not, every funding conversation inside of Red
Hat's walls begins and ends with metrics.  If it isn't measurable, it 
doesn't exist.  Fact.

This is especially important in the case of Fedora, because Fedora doesn't
make any money directly for Red Hat.  We continue to develop Fedora
because it serves other purposes.  Research and development.  Quality
Assurance for RHEL.  The ethics of continuing to provide free software,
which is important to all of us.  And, most importantly from my own
perspective, *community mindshare*.

If we can't quantify Fedora's mindshare in some way, we lose one of the
*major* rationales for making the Fedora Project stronger and more
independent.  Every time a Red Hat executive asks "how many Fedora users 
are out there?" and we answer "oh, somewhere between 100k and a few 
million," we make it *that* much more difficult to defend Fedora from bad 
Red Hat decisions.  If a Red Hat executive has to choose between 
giving resources to RHEL and giving resources to Fedora, and if he's got 
dollar figures on one side of the ledger and hand-wavy "mindshare" guesses 
on the other side of the ledger, he's going to choose RHEL.  Every single 
time.  I've seen it happen, again and again and again and again.  And 
again.

Part of my job as "the community guy" is to fight my ass off to make sure
we don't continue to make those mistakes.  To win that fight, I need
metrics.  And the better metrics I have, the better a fight I can put up
on Fedora's behalf.

So will better metrics suddenly fill the Fedora coffers with gold, and
allow us to hire engineers to solve all of Fedora's problems?  No.  Not
hardly.  But these metrics *are* a necessary precondition to further
funding: of developer conferences, of infrastructure projects, of more
hardware, of more engineers.

In short: Red Hat asks Fedora to be accountable for delivering value to
the company -- value that must be quantified beyond dollars and cents.  
If we cannot be accountable in this way, we make it difficult to justify
further investment.  Investment that is, in my opinion, *badly needed* to
ensure ever more meaningful community involvement.

These are, in the end, political disputes.  Karsten has argued to me
privately that "the Fedora community shouldn't have to care about Red
Hat's political issues."  In an ideal world, that would be true.  But in
the real world, Fedora is *incredibly* dependent upon Red Hat's largesse
to be successful.


2. OKAY, FINE, WE TAKE THE POINT ABOUT METRICS -- BUT DON'T THE PROPOSED
FIREFOX METRICS SUCK?  ARE UNIQUE IP ADDRESSES REALLY THAT USEFUL?

Maybe not.  There may be better ways, and it's clear to me now that we
should have been discussing all of these better ways a long time ago.

In my judgment, this solution had three primary benefits that recommended 
it above all others:

  a. Dead simple to implement;

  b. Will gather the most data, since more computer users run web
     browsers than just about any other application (which is
     arguably anecdotal);

  c. Will gather data that may potentially be sliced across
     several meaningful axes, including language and region.

Seth Vidal also points out that we can gather much of the same data from 
yum, since yum talks to a central server to obtain mirror data.  This is 
another useful approach, and if we'd been having this conversation in 
public over a long period of time, it would have become evident much 
earlier.  I would argue that this data should be collected *in addition* 
to the Firefox data, because it will tell us a different story: not 
necessarily desktop users, users who understand that updates are 
important, users who have bandwidth to retrieve updates online.

Every metric is imperfect.  The Firefox metric is the best one I've been 
able to come up with in this short timeframe.


3. CAN THIS NOT WAIT UNTIL FC7, WHEN WE CAN SOLVE THIS PROBLEM "THE RIGHT 
WAY"?

I don't believe we can.  In my opinion, we need *urgently* to start
somewhere.  It's the difference between building a funding plan now and
building it 6-9 months from now.

Now, that said, I think that solving this problem *correctly* is vital.  
Ultimately, I think a new page in firstboot in FC7 is the right solution.  
But that will require a robust backend, and those are seldom as simple as
they seem when you're drawing them up on the whiteboard.


4. WHY IN GOD'S NAME DIDN'T YOU COME TO US WITH ALL OF THIS EARLIER?

The honest truth: I never expected this firestorm of controversy.  
Setting the start page to go to a vendor site is one of the oldest tricks
in the browser world; Netscape was using this technique as one of their
primary business strategies over ten years ago.  We collect exactly *zero*
personally identifiable information.  It just seems so clearly defensible
to me that I never even considered the possibility that reasonable people
would object to it.

But you are all reasonable people, and many of you object to it.  And it's 
possible that some of our users will object to it as well.  I understand 
that.  But I believe it's a risk we need to take.

It's a tricky thing, you know?  It's clear to me now that we should have
been having this conversation weeks ago -- and I believe that many people
feel more strongly about the way this solution has been shoved down
people's throats than they feel about the proposed solution itself.

For that, I deeply and humbly apologize.  I screwed up.  I'm sorry.

Screwups aside, though, we're at a decision point.  I've done my best to 
convince you all the best way I know how.  The decision is ultimately in 
Max's hands; he gets paid to make exactly these kinds of decisions, and he 
ultimately is held accountable by Red Hat for Fedora's successes and 
failures.  

Anyway.  If you all feel so strongly that my analysis is *so flawed*, and
that the perceptions of this "tracker" have the potential to be *so
damaging* to the goals of the project, that we should wait another several
months and bet on another solution in FC7... well, then, I'll take your
side in the matter, and trust you to help us solve this problem in FC7.  
When we've really needed you, you've never let us down before.

--g

-------------------------------------------------------------
Greg DeKoenigsberg || Fedora Project || fedoraproject.org
Be an Ambassador || http://fedoraproject.org/wiki/Ambassadors
-------------------------------------------------------------