https://lists.launchpad.net/launchpad-users/msg05078.html

Now I can't fault the reasoning for the delay. They want to make use of the new bzr backend as it has significant speedups.  I certaintly can't say a speed up to bzr wouldn't be beneficial. If Canonical wants to say publicly that bzr isn't up to the task managing Launchpad releases..who am I to disagree with that assessment.

But I find it a bit ironic, that development in the underlying dvcs that Canonical is funding the development of is holding up the release of the code hosting web services codebase that Canonical is funding the development of.  Oh sweet irony.

Here's how I would have written the same message:

"I'm sorry, we'd love to release this code to you on the date we promised, but you see, we just aren't that thrilled with usability of the dvcs we've chosen to use internally and have been using internally for years and years now.  We took a show of hands and yep, we all pretty much agree that the implementation we've been using for years blows monkey chunks and we don't want to inflict that experience on any external contributors. We've only been using it because we are paid to use it..we've had no choice in the matter. To be honest with you, we aren't even sure why anyone who wasn't forced to use it  by their employer was using this implementation for any serious collaborative project. It's sooooo painful to use. 

So until the dvcs implementation gets better we're just not going to bother releasing the launchpad code. We think that this codebase is so very different compared to the thousands of other codebases already using the same dvcs that it just would not be worth trying.  We think that its size and complexity of the launchpad code far surpasses the capabilities of the current dvcs implementation.  Yes, the very same dvcs that we have repeatedly encouraged and forced thoursands of other codebases(like mysql and zope) to use to gain access the very same Launchpad infrastructure we are now withholding from the community.  Don't fret however, as we've been working on upgrading the capabilities of the dcvs since last November, and the new capabilities should be reliable enough for us to make use of 'real-soon-now'(tm)...hopefully.
Thank you,
  THE MANagement" 

So even though Android running on ARM was the darling of the show at Computex it seems the established OEMs like Asus are being leaned on by Microsoft to stop talking about those sorts of things so loudly: 

http://www.groklaw.net/article.php?story=20090619161307529

Maybe the marketplace is so distorted that you really can't be just an operating system company and rely on OEM partners in the general computer device space or be just a hardware company and choose to support multiple software vendors.

Maybe the only way to really fight the herding going on in the OEM consortium talked about in the groklaw article is to do what Apple has done..and to control the entire design experience for your devices. Maybe you have to take responsibility for both the hardware and software as parts of overall product design to prevent being strong-armed by your "partners" into making compromises that are not in the best interest of your "customers." 

Is there room on the playing field for a new hardware manufacturer to step up, and making Android or Moblin a defining part of their device design? One that isn't burdened with pre-existing contractual obligations to Microsoft for other products in their line-up?  Or is the barrier to entry in the smartbook/netbook too high even for that approach to work?  Which relatively unknown ARM OEM is going to break out of the trap first?

-jef

So I poked and pulled at the Fedora Community maintainer portal for just a few minutes and I could smell the kickass.

I actually caught myself giggling a little...and I had to chide myself sternly for having fun doing packaging chores..  Remember the scene in Mary Poppins where they needed to clean up the nursery? And somehow they turned the chore into an enjoyable game.. it sort of felt like that..more than I probably want to admit.  Fedora Community is very much a delicious spoonful of sugar.

It's already sort of making my head spin thinking about where this component portal view approach can go beyond packaging tasks.
What would a reshuffled portal interface look like that was streamlined for package submission and review?

I can't wait to see how people take this and create a testing oriented portal.. or a translations portal....or a spins portal.

Exploring the space of possible widget extensions is going to be fun.

-jef

This is a pretty good article about the current state of Xandros.
http://www.theregister.co.uk/2009/06/03/xandros_custom_linux/

If you are a netbook buff there are some gems in there about what it takes to be a good OEM services partner. It's a pretty good read. I personally like these quotes:

"Doing that general purpose operating system is a nightmare, and you lose your shirt on it," Smith explains. "At the end of the day, you have to do something that puts rice in the bowl."

"We have the largest paid installed base of consumer Linux users," boasts Smith. "We've got millions of these things out there, and while Ubuntu has a lot, too, most of theirs are free."

The implication is of course that Xandros is actually building a business out of the netbook market.. not just building a brand. But as they are a private company there's no way to verify their financials..so take that shot at Canonical for what it is.  I certaintly can't verify or refute the claim. But knowing what I know from the laypress articles on the breakdown of estimated netbook sales by OEM..I would say its probably a valid claim.

We don't hear a lot about Xandros.. even though they were the software company that helped create the commercial netbook market as we know it by partnering with Asus and providing the linux based operating system for the EEE PC line. They've been running under the radar for a while now getting very little press, but the article sheds some light on what has been keeping them busy.  It's nice to see them talking about their support for the ARM based Qualcomm Snapdragon....something Ubuntu's Mobile effort seems to be lacking at the moment.  How is Xandros doing in terms of mainlining their hardware support enhancements for ARM?  Hopefully Xandros is active in the upstream kernel ARM porting community.

So it seems like Computex is gearing up to be a cage match demo-off  between Moblin equiped Intel netbooks
http://www.earthtimes.org/articles/show/novell-to-demo-moblin-based-operating,845054.shtml
http://blog.internetnews.com/skerner/2009/06/moblin-v2-coming-to-ubuntu.html
http://www.linpus.com/press_1.php?gid=2&nid=14

and Android Equipped ARM netbooks.
http://chinese.engadget.com/2009/06/01/computex-2009-a-taste-of-snapdragon-netbooks-to-come/
http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9133698

Sort of makes sense why Canonical is scrambling to make an Android compatible kernel available. Seems the ARM OEMs have enmasse chosen to roll out Android as their netbook interface for their first round of product offerings into the netbooks space.  Computex will be the big unveiling event which starts the clock to market in a lot of ways.  If Canonical wants to catch that train they are going to have to run hard to get up to par on hardware compatibility.  It's so important in fact that Canonical is talking about maintaining two different kernel trees...the normal distribution one with the Canonical secret sauce.. and an Android enabled kernel with Google's secret sauce. That's a lot of secret sauce. I hope their kernel team doesn't drown in it. 

-jef

You know what's weird... May 2009 was a record setting month for unique IP traffic for fedoraproject.org

1,996,040, just edging out November 2008.... and we haven't actually done the release yet.  Did we drive that much traffic just from building pre-release interest in the F11's feature set?

I'm very curious as to what the peak will be for June will be with F11 releasing this month.

-jef

So we are coming up on a new release again and i find myself asking myself the same question I ask myself around every release...well maybe not every release, every Spring/Summer release.

Why do i spend any time whatsoever contributing (or pretending to contribute) to Fedora?  I could be out off-the-grid paddling in the river late at night right now. (Land of the midnight sun and all) instead of fielding bug reports for the few packages I maintain in Fedora.  Why do I do it?

Because it still feels like it matters. Even when the project hits a collective snag on policy issues and there's a dust up of what is the best direction to take, it still feels like it matters. I still feel like Fedora as a project is making an impact on pushing  the entire open ecosystem forward towards a better technology future for everybody. Or maybe Fedora is reaching into the future and dragging it back for everyone to use now. Anyways, it still feels to me like Fedora is playing an important  role that other projects are not.  My personal contributions to the effort maybe miniscule, but as long as they aren't counterproductive I get some personal satisification out of knowing I'm associated with the effort. More satisification that being out on the water with paddle in hand. Though I continue to find that sort of odd. I really like being out on the water. Maybe its more the fact that its marginally easier to work on Fedora while drinking, than it is to kayak while drinking. Maybe I just need to get a beer hat as standard kayak gear. Hmm. Yeah screw it, I'm getting a beer hat and going kayaking all summer.

-jef

Who's going to be the big winner in the ARM based netbook market which is going to show up later this year?  Android will be.. its pretty much a fargone conclusion.  Don't believe me? Here let me quote a Canonical employee.

https://wiki.ubuntu.com/KernelTeam/ARMBeijingSeminar2009Report

"Although there is no Google guys to talk about Android, Freescale, Mavell and Thundersoft demoed Android. I have to admit Android caught more eyes than us. People asked about our ARM netbook product/solution and business models. Because of some hardware issue of Freescale babbage rev1.0 board, the performance of Ubuntu on Babbage (compared to Android on Babbage) is not very good. Android demos runs on Qualcomm Snapdragon/8xxx, TI OMAP3, Freescale Babbage and Marvell PXA3xx. "

"Customers are very interested in Android and Android has better performance than Ubuntu on Freescale Babbage "

"Ubuntu can only support Freescale Babbage, we have to support more platform"

With the ARM chip manufacturers like Freescale and Marvell demoing Android on Google's behalf without Google sending anyone to that conference  to do the salespitch...that speaks volumes as to where the entire ARM OEM landscape is headed..

Okay "garunteed" may not be strictly true, but clearly Canonical has an uphill battle with ARM netbooks against Android even though they have a track record with intel netbooks.  More importantly, Canonical actually recognizes this as a problem.. they clearly don't smoke the stuff they've been handing out to the laypress...which is a good thing to see. It's always bad for business when you are hooked on the same drugs you're peddling.

The burning question is, will there be enough of a market leftover to pay Canonical to do the necessary  work to bring the ARM port of Ubuntu up to par with Android as an competitive option?  Or is Android going to win the ARMs race simply by out performing everyone else at the starting gun? 

-jef"I fully expect that wiki page to be removed from public view as soon as Canonical realizes that I'm using it to question the sustainability of their business model"spaleta

If you are not familiar with the Fedora Client statistics effort take a moment and read:
http://fedoraproject.org/wiki/Statistics

I'd like to take a moment and talk specifically about how to do a better job at interpreting the total unique IP connections listed here:
http://fedoraproject.org/wiki/Statistics#Total_repository_connections

There are two competing factors which influence how unique IP counts can be interpreted as client counts.  On the one hand there is the effect of private subnets which map multiple clients to a single IP address. This would lead to the unique IP address count to be an undercount of the actual number of clients.  On the other hand we know we have clients which roam across networks and those clients could easily be counted multiple times in the unique IP logs, leading to the unique IP counts being an over estimate of the actual number of clients.

So which is it in reality? Is the 14 million+ unique IP counts sitting in the Fedora MirrorManager logs an over or under count of reality?

I'm here to tell you friends, that its an undercount..by about 15%.  There are probably about 16 million Fedora clients in the wild in reality. How do I get that?

Easy, I had my buddy Mike "Chops" McGrath do a little data mining of the Smolt logs and come up with an aggregate ratio of Smolt UUIDs to unique IPs.  That ratio can be taken as a scaling factor to convert unique IP counts to unique client counts given the following assumptions.

1) The smolt userbase represents a sampling of the overall client base which is no more likely to be on a private network than the average Fedora client.
2) The smolt userbase represents a sampling of the overall client base which is no more likely to have a dynamic IP address than the average Fedora client.
3) The ratio is reasonably stable over a release cycle timescale, but may be subject to a slowly varying drift.

If those three assumptions hold the ratio of UUIDs to IPs is an adequate scaling factor.  We looked over the last 16 months of aggregate Smolt logging data here is what we found:
Mean Ratio: 1.16
Ratio Stdev: 0.0263

Here's is a graph of the Smolt ratios calculated monthly.



I'm pretty confident in the validity of scaling factor. I'm also very pleased to see that the number is greater than 1.  This means that the currently unique IP address statistics we are showing are a conservative estimate of the actual client numbers.  No caveats, no soft-selling.
There are 14 million+ Fedora clients out in the wild and its time we start making that point loudly and confidently.

-jef"Measurement methodology matters"spaleta

 

 

Okay, why exactly was steroid use by major league baseball players such a scandal? Objectively speaking, steroid use actually helped the for-profit businesses which keep major league baseball going make more money. Better hitting, better pitching...better ticket sales...better free agent salaries...better business. Steroids was good for the business which makes major league baseball possible... why wasn't it ultimately good for major league baseball?

I'll tell you why.. because baseball is more than just the business of selling sports entertainment, baseball like all professional sports competitions is grounded deeply on the sense of fair play.  The ground rules of fair play are easy to summarize.  Give competitors equitable and fair access to similar equipment and resources and let them compete against each other based primarily on personal dedication and skill under an established set of rules.  Steroids upset the ground rules of fair play as it was perceived to be a quick fix to shore up shortcomings in either dedication and/or skill.  Though I could probably argue it took a certain sort of personal dedication to the concept of achievement to inject yourself in the arse with a needle every day.  I hate needles, I could never be a major league baseball player.

Steroids were great for the business of baseball..but it was bad for baseball itself.   That's an important point...a point I think the open software ecosystem and the businesses which back it need to understand.  I think proprietary web services maybe the equivalent of baseball's steroid scandal. Proprietary web services may in the short run be good for the business interests which support the open software ecosystem..but it might ultimately be very bad for the open software ecoysystem itself.  Maybe not for exactly the same reasons, analogies are only good up to a point, but I think its worth reflecting on as an example of how business interests can work against the interests of the larger community. 

I think the Franklin Street Statement on the freedoms of network services has come a little late as a a fully preventative measure, but it might not be too late to have a positive impact on aligning long term business interests with community interests to prevent a deep debilitating rift.  Going back to the baseball analogy, its sort of like its 1988 when Tom Boswell broke the silence concerning Canseco's use of steroids in an interview with Charlie Rose on CBS.  The Franklin Street Statement maybe that sort of moment.  I'd like to hope the open development ecosystem will take less than the decade it took baseball to deal with steroids, to confront the business interests which are leveraging the concepts of proprietary web services which undermine the sense of fairness in the open development model.

-jef"play ball"spaleta