These Things Matter To Me

I like VMware stuff a lot, and write about them regularly, usually in a positive way, so I write this list of VMware Linux anti-patterns with only good intentions.

Lately I’ve been feeling a bit bummed by VMware’s deprecated Linux experience, and Windows-centric mentality. If VMware gets too Windows-centric, it’ll only be competing with Microsoft. I want VMware to do well against Microsoft! And I believe that to do so, it needs to deliver an experience that doesn’t marginalaize Linux users.

Here’s how VMWare can improve the Linux experience:

1. Improve the vmware-tools install on Linux guests. (All VMware products, all host OS’s)
   On a Windows guest, you run a binary (by clicking it!), and you’re done. On the Linux side, you run a binary (most likely via command-line), and then visually, you‘d think you were done. but you’re not. You installed it, now you configure it. Seriously, whenever I train or help users with this part, I feel guilty/embarrassed about how idiosyncratic it feels. The output of the rpm/install script never tells you “now you configure it.” Sure the docs do. But on Windows you don’t have to do all that. VMware, I won’t hold you responsible for the command-line parts, that’s (mostly) in Linux’s hands. But you could longterm do away with the configure, and short-term, announce at the end of the install that, “Now it’s time to configure.”
2. Make a Virtual Infrastructure client for Linux.
   You made one for Windows. I think you love Windows more than you love Linux. (And yes, I am aware of the web client. I’ll pretend you didn’t suggest that.)
3. VMware Converter is way more difficult with Linux than Windows.
   • The VMware Converter program itself only runs on Windows
   • Converting Linux physical-to-virtual VM’s is a more burdensome process for Linux guests, with more rules and hoops than Windows guests have.
4. I feel like you’re slowly taking away ssh/ “service console”/ file system access to your ESX-based products. You’ve only done this to 3i so far. And there is a workaround. But the vibe I get (I hope I’m wrong) is that you’re trying to wean us off of standards communications and file-system access to this stuff.
5. Give VMWare Server 2.0 a Linux client.
   I know you took away the non-web client from both Windows and Linux, but since Windows people can use the new VI client to access VMWare Server, I’m still counting this as a ding against Linux users.

Real talk.

Gutsy’s so close I can feel it.
This weekend I installed the beta of Ubuntu Linux 7.10, Gutsy Gibbon.
A few notes to myself, and just one bug:

• Bug: no matter how many Firefox windows I have open, they never begin to “stack up,” instead spreading infinitely across the bottom GNOME panel. I never realized this before, but the act of stacking Windows up after a certain point keeps the the title of each window legible. I’m sorry I never appreciated you before, mister/miss Firefox-stacking-window situation.
  You don’t know what you got, ’till it’s gone.
• after install via upgrade, sound comes back muted. Not a big deal, but after you install Linux, you’re paranoid about hardware not working, so for a minute I thought sound support got worse with Gutsy. You really had me going there, Gutsy Gibbon! But you merely muted me. Not funny, Gutsy. This isn’t Punk’d.
• gnome-setup has crashed a couple of times. I don’t know. I can’t reproduce it at will, but it happens.
• (not a bug) but I still have to install a special driver to get graphics to work right on a Dell 700m.

(phone from keetra)

UbuntuLive is live!
My talk isn’t until this afternoon, but I want attendees to be able to have access to my slides to take notes on, without having to wait until after the talk.

• .odp slides (Open Office format, edit inside of.)
• .pdf slides (pdf format. Less editing options, but simpler for some.)

Like many slides, they don’t totally stand on their own without the verbal “filler.” I’m making a “prose” version of this presentation at some point to address that

Yesterday I began what will probably end up becoming a series of posts about how y’all need to rethink the meaning of the word “enterprise,” and related, the value of support. In specific, I called out how “Enterprise Linux,” (usually meaning Red Hat Enterprise Linux and Novell’s Suse/SLES/SLED) is overvalued because the packages you pay for often need to be replaced with packages you don’t pay for.

Another problem with Enterprise Linux is the way it’s sold. You buy the bits and support together. People might think it’s hard for an open source software vendor to just sell the bits, when technically, so much of it is “free,” and its easier to just imagine all those software dollars are actually paying for “support” (representing commercial man-hours, not free) but the simple fact is many organizations would love to pay for the bits they could technically get for free, and just do without the facade of expensive support, when the support they get from other resources is more responsive.

Photohosting site smugmug was in that boat, and blogged about their issues with Novell and Redhat.

…we loved Red Hat Linux, we loved how good they were at building & testing their software, we loved their mechanism for delivering software updates. We just didn’t need support.

We got on our knees, begging and pleading with Red Hat to let us pay for a “software updates only” license. They wouldn’t have it. “Support comes bundled with updates”, I was told, “no ifs, ands, or buts”. I *want* to pay Red Hat for the valuable service they do for us and the community. I just don’t want to pay for the part we don’t need – human support.

I would really like to pay Red Hat for all their hard work building and testing the software. .. It’d be the right thing to do. But Red Hat won’t let me.

The company ended up going with CentOS, a clone of Red Hat Enterprise Linux I’ll write about another time. Again, you’ll notice the author isn’t trying to avoid paying for the software, he just doesn’t want to pay for a service (”support”) he doesn’t need.

WHAT WE CAN DO

Many of us who use Linux in commercial situations are more than happy to pay for it. Let’s feel comfortable paying for it in different ways. We need to get over the traditional model of a single, central body of developers and supporters being embodied by a single company. There are different currencies and parties involved. Give back to your providers creatively with money, bug fixes, documentation, and sharing your best practicies. Take the time to identify the upstream developers and projects and consider funding them directly. Publicly share your challenges and success stories on the internet, the attention will help future users and the developers by making their project less of an unknown quantity for future users.

But most importantly, be willing to break free of this totally broken tradition of thinking paying a bunch of money to a central body in some way solves your technical problems and protects you. It may make certain people in your organization feel safe, but take the time to run some numbers. What value have you really gotten out of support in the past? Put the burden of determining value on those who sell it.

• even old-school business analyst Forrester says organizations are using open source for “mission critical applications.” A less breathless, more helpful informative summary here, courtesy the very articulate Matt Asay.
• ya habla Español Dell. puro en directo. Dell’s new Spanish blog, Dell en Directo. You already know about the English juan.
• Arizona State uses Drupal, and documents their set-up
• Dell’s improved Linux forums (I haven’t figured out how they’re improved yet. Taking their word for it. It’s possible they’re manipulating me in a state of post-Ubuntu-announcement bliss.
• some belated notes on Red Hat Enterprise Linux 5

Some interesting thoughts over on Slashdot’s coverage of the Dell/ Ubuntu announcement.
For one user, the announcement is actually a deciding factor in a purchasing decision:

Personally, I have resisted the siren call of Dell for a long time. This changes my mind. I need a new machine and this could be just the ticket — it was either that or refurb an old HP with a new HD and a copy of Feisty Fawn. I like the idea of it pre-loaded.

And yeah, that’s the kind of response Dell and Ubuntu want to hear. But I think it’s time we really question the value and definition of support in the way it’s been thought of in the past. Support is extremely overvalued, and lack of support is too often used as a reason to squash a great tool or piece of software. Let’s examine our commercial support relationships and think of what we really get out of them. Are our bugs fixed faster? Features added more quickly? Do we find out about upcoming products from our vendros before the blogosphere does? When we have a configuration question, whose documentation is more helpful? Community sites and mailing lists, or the official documentation?

I’m very excited that Dell and Ubuntu have a relationship with each other now, and there’s no way it can hurt the quality of Linux on Dell hardware. But let’s not wait for announcements like this before we feel comfortable pursuing technologies that otherwise trump their commercially supported peers.

related:

Dell interview with Mark Shuttleworth about the announcement, how Linux gets adopted differently in different parts of the world.

(Ubuntu Linux CD originally uploaded by excellent Flickr citizen Pen9)

I’m as giddy as the rest of y’all. Yes, Dell will pre-install Ubuntu on Dell hardware.
Some of us have been waiting for this news for a while, either in the affirmative or the negative. After all, it was over a month ago when Dell closed its Linux survey. A bunch of different commercial and community-based distros rallied to get their Linux represented. And then silence. Dell didn’t let us know how, when, or what they’d do with the data. They thanked us, but dat about it.
So it was very exciting to hear this evening first a rumor via Dell, and a confirmation via Canonical that yes, Dell will support Ubuntu.

But let’s be gracious here . DesktopLinux.com wrote “Dell to choose Ubuntu,” in a tone suggesting only one Linux could be chosen. Mos def some of this language is left over from the Dell’s own Linux survey, where one could vote for only one flavor of Linux for Dell to install. But remember bug #1 people. It isn’t that Fedora has majority market share . Let’s hope Dell’s experience of offering Ubuntu is so delightful, they get involved with other flavors of Linux, and get on the right side of bug #1 ;D