Max (spevack) wrote,

i promise i'm not trying to start a flame war

Normally I just ignore this kind of stuff, so as not to start a flame war. But I think the Distrowatch guys were polite in the way they wrote the article, so I shall write back and attempt to be equally polite, and just leave it at a one post each thing, and not start a flame war.

The Distrowatch piece centers on a fact: there have been surveys conducted asking people what Linux distribution they would like to see shipped on various laptops (Dell and Lenovo). Ubuntu has handily won these surveys.

Additionally, in its reporting of this latest survey, Distrowatch brings up an interview that I gave to some folks at the Red Hat Summit, which was published back in July.

In that interview, I state that various popularity polls and rankings do not really concern me. I state that Fedora keeps its own statistics which it publishes weekly. In short, the list of things that I have the opportunity to worry about as Fedora Project Leader is quite long, and so I simply choose not to worry about popularity polls taken on Distrowatch, or elsewhere. That's no disrespect to Distrowatch, it's simply a matter of my own bandwidth.

However, the Distrowatch article takes what I said and extrapolates from it two things:

(1) I am unable to admit that Ubuntu has done something right.

(2) I am not concerned about hardware vendors choosing Ubuntu over Fedora.

I take issue with these two conclusions, and I would add to them three questions of my own:

(1) What does it *mean* to win one of those surveys?

(2) What does it *require* to win one of those surveys?

(3) How are the results of those surveys acted upon? What happens if you follow the money?


First off, I have been very public about admitting that Ubuntu has done many things right. I'm not going to dig it all up via Google, but trust me when I say that I have always spoken and written highly of what Ubuntu has achieved.

The simple fact is that there was a time when Red Hat had lost its way with regard to Fedora and the community. This is undisputed. Over the last few years, I would say that the *primary* thing that the folks who are in leadership positions within Fedora have done is work to overcome and repair the mistakes of the past.

And I believe that we have done that job well. That is not to say that Fedora is the Paragon of Linux Distributions, but we've done well.

But recent history doesn't erase older mistakes. It simply creates some new goodwill, and it will be up to the future leaders of Fedora to ensure that goodwill is not squandered. Recent history also does not mean that Red Hat didn't open the door more widely for competition to come in and try to Do Things Better. Ubuntu seized on that opportunity and has had lots of success.

My entire point with the statistics angle is one of transparency, not one of result. I would love to see how much money Canonical has spent on ShipIt. I would love to see their claims of how many million users they have backed up by some real data, with a disclaimer (like ours) of where the potential fudge factors are. I would find it incredibly interesting to see how much money Canonical spends on Ubuntu every year, and compare that to how much money Red Hat spends on Fedora. It would be interesting to be able to calculate some sort of "cost per user" metric, if you will.

I recognize that Canonical has no requirement to provide any of that data. I suppose Red Hat doesn't either, but we go out of our way to do so as much as we can.


What about hardware vendors choosing Ubuntu over Fedora?

To win a survey like the Dell or Lenovo one requires you to have lots of users, who care enough about the distribution to go and vote for it. But what does it actually MEAN to win a survey like that, from a corporate and financial point of view? Once you are talking about selling machines with a distro pre-installed on them, then someone, somewhere along the chain is getting paid something. The question is who makes the money, how much are they making, and what is the margin? By margin I mean "how much money do you have to spend in order to make 1 dollar?" Are you spending 50 cents? 80 cents? 95 cents? And how do you make the margins tilt as far in your favor as possible?

Canonical has relationships with hardware vendors selling Ubuntu. Red Hat also has relationships with hardware vendors, but instead focuses on Red Hat Enterprise Linux and not Fedora as the revenue-generating product. Two different models. In my opinion, Canonical invests more heavily in a desktop-oriented market than Red Hat has in the past. I would argue that Ubuntu's primary focus is on the desktop, and I don't think you can say the same thing about Fedora or RHEL. That is a strategic decision on the part of Canonical.

Red Hat is a public company, so the successes or failures of its business strategy are transparent. Canonical is not a public company (and that's fine, good for them), but as such it is basically impossible to make comparisons between the two. Maybe they will experience some of the same growing pains that Red Hat has already gone through, as the pressure to produce significant revenue and profit increases. Will that happen? If it does, will it have any effect on Ubuntu? Maybe instead they will become so big and successful that they one day purchase Red Hat and set me to work doing QA on xbiff. We'll just have to wait and see.

Anyway, the point is that Red Hat allows Fedora to sink or swim on its own merits. It provides the financial support necessary to engineer and develop the distribution, but so far Red Hat has limited itself to Red Hat Enterprise Linux when it comes to OEMs and hardware vendors. Fedora serves as the upstream for Red Hat Enterprise Linux, and that is the financial side of Fedora's value proposition to Red Hat.

So the fundamental strategy behind Canonical and Red Hat is a little bit different. Any discussion of the two companies (and the distributions they help to produce) should be framed in that context. Unfortunately, it's kind of hard to do all of that in only a few words.

Fedora is about building community, about continually working to make a Linux distribution in as open and transparent a way as possible, and about pushing the envelope of innovation in free and open source software. Red Hat's contributions to free and open source software include NetworkManager, libvirt, SELinux, Liberation fonts, One Laptop Per Child, more contributions to the kernel than any other company, a fully open build system and compose tool (Koji and Pungi), and significant contributions to GNOME, RPM, and yum. And that's just scraping the surface of the list.
Tags: redhat
  • Post a new comment


    Comments allowed for friends only

    Anonymous comments are disabled in this journal

    default userpic