Archive for March, 2007

Beryl 0.2.1 in Universe

Saturday, March 24th, 2007

I was thrilled to see a slew of new Beryl packages land in Ubuntu yesterday.

There’s been a furious amount of activity from the MOTU and Beryl upstreams to get these packages ready for Feisty inclusion – cleaning up copyright issues as well as getting the packages themselves into first class order. Now the rest of us can test Beryl simply by:

- enabling universe in /etc/apt/sources.list (or, in Ubuntu, just use System->Administration->Software sources)

-  installing Beryl (“sudo apt-get install beryl”)

After I’d done this, I could get Beryl up pretty much immediately by just running “beryl-manager” at the command prompt.

A huge thank you to the folks who worked so hard to get this done in time for Feisty – lupine_85, racarr, imbrandon, pricechild, and others! The MOTU team’s response to this challenge was awesome.

SUN’s free software credentials took a big step forward today with Ian Murdock’s decision to join the company. Congratulations all round. Now we have to find a new CTO at the Linux Foundation, and one with big shoes to fill. Hopefully, Ian’s experience will lead to something ISV’s very dearly want – commonality and standardisation across their non-Windows target platforms. Here’s to an interesting year!

Pre-installing Linux

Wednesday, March 14th, 2007

There’s been a tremendous level of interest in the fact that pre-installed Linux (in the form of Ubuntu | Fedora | OpenSuSE) is the #1 rated suggestion on Dell’s IdeaStorm. On the face of it, there is little question that Linux pre-installation is popular with customers. Why, then, is it so difficult to buy a PC in the US or Europe that has Linux (and ideally Linux alone) on the hard drive?

The devil, as always, lies in the details.

First, margins on PC’s are razor-thin.

This has two significant consequences. Most importantly, it means that Microsoft co-marketing funds are a substantial portion of the profit margins for many large PC retailers. Tweaking the nose of the giant might be fun but it’s risky. If Microsoft reduces the per-PC marketing contribution it makes for a particular reseller, that puts them at a huge financial disadvantage relative to their competitors. This means that one of the biggest issues a computer manufacturer or reseller faces in considering Linux pre-installations is the impact it will have on the Microsoft relationship, and hence bottom line.

Also, thin margins mean that any customer interaction or support call can blow away the profit not just on that sale, but on many others as well. The worst-case scenario is a customer who buys a computer at the lowest price off your website, assuming it’s a Windows machine, and then calls, infuriated, because it “won’t work with the game they are trying to install”. One customer who accidentally gets Linux without knowing what that means is an expensive proposition for a company that makes relatively little on the low-end product range. For this reason, I don’t think it makes any sense for Walmart to sell low-cost Linux PC’s, and we’ve never pushed US / European retailers to try pre-installing Ubuntu unless we think they can segment out the market which genuinely WANTS Linux from those that are just looking for a great deal on “a [windows] computer”.

Second, we free software fans are a fussy crowd, and very hard to please. You know what you are like – you sit and configure that Dell system down to the finest detail, you want a specific model of HP laptop, you want the one that has the Intel graphics chipset not the other chipset because you prefer the free driver approach from Intel… you are in short an expert, demanding customer. This means, that in order to reach us with Linux, a reseller has to offer Linux EVERYWHERE, not just on a few select models.

Worse, we are not “Linux” users, we are users who want version 6.06.1 of Ubuntu, or 10.2 of SuSE, or Fedora 6. We want a specific distro, and in many cases also a specific VERSION of that distro. In order to please us, the vendor has to offer an enormous matrix of possibilities – machine and distro/version.

This is an expensive proposition.

So, what can we do to help address the need?

First, we can help the vendors get more detailed insight into the real nature of demand. For example, here’s a survey being run by Dell that will I’m sure help inform their decisions about how they help you get Linux on Dell:

It would be great, of course, if those sorts of surveys were less vendor-specific, so that we could express our opinions once and have that counted across the whole industry, but there you have it. (It would also be great if Dell would consider Ubuntu to be both community- and commercially-supported, but that’s a different story ;-)).

Second, we can start looking at ways to change the model so that there’s a better fit between customer expectations and the economics of the industry. For example, if you’re one of the people who voted for Linux pre-installation on Dell IdeaStorm, would you be happy to receive a Dell box with no OS and with an Ubuntu disk in the box, which you yourself installed, with no support from Dell? What if it came with an assurance that the set of components you had configured *should* work, but no guarantee? Can we tweak the parameters to get to the point where you would be satisfied, and Dell could make a reasonable profit with only reasonable risk? Solve that, and I think we could all get one step closer to fixing Bug #1.


Of course, some resellers specialise in Linux pre-installations. My favourite of them is System76, who do a great range of laptops and desktops with, amongst others, Ubuntu preinstalled. Kudos to them for spotting the market and making the most of it.

A very good start

Monday, March 12th, 2007

I’m very impressed with the results of the early work being done at the new Linux Foundation, which is OSDL+FSG with a leaner focus on getting things done.

Working in civil society, quango’s, non-profits or consortia is extremely difficult – money is always tight, it’s not as clear what the metrics of success should be, and it’s often hard to get consensus from a critical mass of players on what needs to be done. So I think it’s a credit to the folks who setup the new entity that they have been able to narrow the focus substantially and get buy-in from all the major players on the goals for 2007.

I’ve been nominated for and elected to, and have accepted, a seat on the board of the Linux Foundation, not in my capacity as founder of Ubuntu or via Canonical, but as an independent representative of the free software and Linux community. I’ll endeavor to wear that hat as effectively as possible in the role!

I’m not a great fan of consortia – they are always at risk of being divided by their own membership, but I agreed to take on this role because I think the management have a mandate from the funders to deliver something that’s really important for free software – an open process of standardisation that delivers results in competitive time frames. I’ll do my best to help them achieve that.

Tags in Bazaar

Thursday, March 1st, 2007

Just landed in Bzr trunk for 0.15:

peregrine% bzr help tag

usage: bzr tag TAG_NAME
Create a tag naming a revision.  Tags give human-meaningful names to
revisions. Commands that take a -r  (--revision) option can be given
-rtag:X, where X is any previously created tag.

Tags are stored in the branch.  Tags are copied from one branch to
another when you branch, push, pull or merge.

It is an error to give a tag name that already exists unless you pass
--force, in which case the tag is moved to point to the new revision.

-d ARG, --directory=ARG
Branch in which to place the tag.
-r ARG, --revision=ARG
See 'help revisionspec' for details
--force               Replace existing tags
-h, --help            show help message
--delete              Delete this tag rather than placing it.

Looooovely :-). Can’t wait for 0.15!