In early January, I released my annual set of predictions for the upcoming year. Among them was the contention that “Ubuntu will emerge as the de facto alternative at the expense of SuSE.”

As might be imagined, the people at Novell had questions about this assertion. These we discussed during a briefing on January 26th. To their credit, the SUSE team was polite and respectful, even as we fundamentally disagreed.

Their view – bolstered subsequently by high profile wins such as the London Stock Exchange and Watson – was that SUSE was performing well across a number of metrics from revenue to ISV recruitment.

My own is that SUSE faces some fundamental challenges.

The disconnect is simple to explain: we advantage different metrics. Our belief at RedMonk is that more often than not, sustainable performance is most strongly correlated with developer traction, visibility and usage. Revenue, shipments and profit are excellent at measuring how you’re performing now; they are less adept at anticipating future direction.

Developer adoption, on the other hand, is, in our view, highly predictive. It is difficult to identify cases where volume adoption of a given technology has not resulted in its success. Profit potential varies, of course, according to a number of variables from licensing model to addressable market size. But historically, optimizing for developer adoption is a sound strategy; it is arguably true that this correlation is becoming stronger.

The metrics we look at at RedMonk, then, are intended to quantitatively assess developer opinion. There is no single metric for accomplishing this; instead we employ – depending on the subject – varying combinations of public and private data sources to form a larger narrative. Even trivia like Google Trends search data has significance when used in context.

Sometimes the conclusions reached merely validate the conventional wisdom; Amazon Web Services is one such example – it is just as popular as commonly believed. More often, our explorations of developer traction turn up less understood strengths, weaknesses and areas for improvement.

We perform these analyses regularly for research aimed at particular developer communities; my FOSDEM talk “The Rise and Fall and Rise of Java” [coverage] was one such. But we also perform them regularly on behalf of clients, whether they want us to measure them at a corporate level or to look at specific products or platforms.

And while Novell is not currently a RedMonk client, having prepared a few of the charts in preparation for my call with the firm, I thought the embedded slides might be useful both for those wondering how I came to the conclusions I did regarding Novell and for those curious about how RedMonk tries to quantitatively measure developer traction. This is just a sample report, and does not include the executive summary, remediation recommendations or backup data, but it communicates the essence of what RedMonk Analytics Custom offers.

As W. Edwards Deming once said, “In God we trust; all others must bring data.” The State of Novell and indeed all of our “State of $YOURCOMPANY” reports are us attempting to comply. The slides above are a measurement of how we see the firm performing according to a few of the metrics we believe are important. As such, it is a complement to, rather than replacement for, your traditional research services. We do not discount wholesale cited revenue performance metrics, marquee account wins and so on. We simply believe that relative to bottom up adoption and usage, they are less likely to be predictive.

If you’d like to see how your company or product is performing, we’ve got RedMonk Analytics Custom packages available for existing clients and a la carte options for those who aren’t working with us.

For the developers in the audience, we hope you find this research of interest, and feel free to let us know if there’s something you want us to look at.

Disclosure: Novell, as disclosed above, is not a RedMonk client. Mentioned competitors Canonical (Ubuntu), Microsoft, and Red Hat (Fedora) are RedMonk customers.

In March of this year, the hedge fund Elliott Associates offered to buy Novell for approximately $2 billion. This set into motion a number of processes, one of which led to the news that Novell had agreed to be acquired by Attachmate for a slight bump over the original valuation. Andy Updegrove has the most detailed dissection of the transaction I’ve seen, for those interested in the deal mechanics.

Assuming the sale proceeds, the implications are many because Novell’s portfolio is broad. The concerns we’re primarily asked about, however, center around two assets: SUSE and Novell’s intellectual property.

SUSE was originally obtained by Novell via acquisition in 2003. Though it has been largely overshadowed by Red Hat from a public perception standpoint, SUSE was the de facto alternative for commercial Linux implementations through the beginning of this year.

This role has substantially eroded following the news of Elliott’s offer, however. The market has already largely reacted to the pending sale, in fact. Canonical, the commercial sponsor of the rival Ubuntu distribution, is pointing to recent Netcraft surveys as validation of its claims of strong customer growth over the past three quarters. Approximately the same timeframe as the news that Novell was being shopped, in other words. Although correlation does not prove causation, it is likely that Ubtunu’s growth is coming at least in part at SUSE’s expense.

Novell’s partners, likewise, have already hedged themselves in anticipation of a new home for the SUSE asset. IBM, for example, has been conducting webinars with the aformentioned Canonical. While IBM has supported Ubuntu previously, support centered around the desktop. The recent server level engagements can be read as a direct consequence of the uncertainty around the future of SUSE.

Rumors originally had VMware as a likely destination for Novell’s Linux assets, a home that IBM and others might be less than happy with. What remains to be seen is whether Attachmate will prove more favorable, either for partners or customers. Initial reactions have been less than favorable, an attitude not improved by a general unfamiliarity with Attachmate.

The future of SUSE, then, remains uncertain. It maintains substantial commercial account penetration, and the acquisition may resolve questions about the long term viability of its corporate parent. On the other hand, it is possible that while the uncertainty stemmed SUSE adoption it also forestalled migration. Current customers may have been waiting for definitive information on SUSE’s new home before deciding whether to proceed with migration; we received inquiries on the viability of Debian and Ubuntu the day the news of the acquisition closed. Attachmate may be viewed as less threatening to potential partners than an industry player like VMware, but this neutrality is less encouraging to customers and developers alike. At a minimum, the two parties will be scrutinized by shareholders contemplating class action litigation.

The next two quarters should tell us much about SUSE’s future role in the commercial Linux distribution landscape.

Of far greater concern to some, however, is Novell’s intellectual property portfolio. As has been well documented, as part of the transaction Novell sold some 882 patents to CPTN Holdings, a consortium that counts Microsoft as a member (more on them here), for $450 million dollars. Although Novell subsequently disclosed that it was retaining its Unix copyrights, this does not satisfy fears regarding the patents. Copyrights, remember, refer simply to a given codebase, while patents refer to the idea or invention behind them. It’s entirely possible, then, that Novell could retain the copyrights to code as Microsoft simultaneously acquired patents that read on same. Questions, therefore, remain: what – specifically – was the intellectual property acquired? More importantly, what’s the intent of purchase: are they being purchased for offensive or defensive purposes? It’s admittedly speculative to extract intent merely from Microsoft’s recent history with respect to intellectual property licensing and litigation, but in the absence of other information this reaction is natural.

Even the cost of the asset is curious. At $450 million, the IP is clearly viewed as material: that’s almost the cost of one Azure datacenter, in Microsoft terms. But if the IP potentially reads on open source projects like Linux or Open Office, let alone commercial products like Microsoft Office, even this valuation is far too low, assuming an open sale. The lack of evident external interest in the IP has certainly raised eyebrows. One explanation for this apparent apathy might lie in the OIN licensing. Novell, you’ll recall, was a member of the consortium whose purpose is to ensure royalty free licensing and patent non-assertion for “Linux and Linux-related systems and applications.” If said licenses terminated upon sale, the motivation for OIN members such as IBM and Red Hat or licensees such as Canonical, Google or Oracle to involve themselves in the bidding process might be substantially higher. According to the OIN’s Keith Bergelt, however, “ALL obligations remain in place. So, all OIN licensees at the time of transaction close will have access to Novell’s patents through a cross license under the terms of the OIN license.” As an aside, it’s interesting to speculate on what impact, if any, this had on the value of Novell’s assets at the time of sale, and whether any hypothetical asset devaluation may have been previously offset with prior cash incentives.

Though it’s attracted less attention than other recent merger and acquisition rumors, then, it’s entirely possible that the Novell transaction has a more profound impact on the industry than the far more controversial Oracle/Sun transaction. Such is the unfortunate ability of patents to have broadly overreaching – and unpredictable – consequences. Given the length of the sale, the idea that Novell’s partners and the wider Linux ecosystem isn’t familiar to a greater degree than the public with the nature of the assets sold isn’t plausible. Which just returns us to the original question: what did Microsoft and its fellow consortium members pay approximately half a billion dollars for, and why?

Like the rest of the public, we look forward to answers on these questions in the months ahead.

Disclosure: Canonical, IBM, Microsoft, and Red Hat are all RedMonk customers, while Google, Novell and Oracle are not.

Never in recent memory has there been less focus on the operating system, a subject we examine periodically (coverage, coverage, coverage). With the exception of mobile, where the adoption of Android versus iOS is a subject of intense interest, operating system inquiries have never been fewer.

We see this qualitiatively, as the nature of our engagement shifts to the layers above, below and around the operating system; from the virtualization substrate to the middleware equivalent abstractions for application deployment to the management tools that both provision the above and keep it running.

We also perceive the shift quantitatively. In October of 2009, “opensolaris vs linux” was the #1 incoming query to RedMonk properties according to RedMonk Analytics; “linux vs opensolaris” was #7. In October of this year, “opensolaris vs linux” is #13; “linux vs opensolaris” is not in the Top 100.

The simplest explanation for this might be the decline and acquisition of Sun, and the subsequent deemphasis by new owner Oracle to the Solaris brand generally and OpenSolaris specifically. But for this to be true, we should have expected to see commensurate gains to other operating system related terminology, be that Windows, Linux or combinations of both.

We see no evidence of this. Our data shows instead a profound shift in inquiries to layers of abstraction above the operating system; Node.js, for example, is massively trending over the past six months, occupying three of the top four spots. As are terms such as “PaaS,” “cloud fabric,” and “saas cloud.” We can even observe in real time the difficulties developers are having in digesting the implications of the new development paradigms and related terminology via queries like “cloud & saas & difference” (#23 last six months).

The obvious conclusion is that the operating system market is in transition.

The deemphasis of the operating system is even apparent from a resourcing perspective. The team originally responsible for DTrace, a still differentiated piece of observability technology built into Solaris, have all left Oracle. Rather than attempting to build another operating system or helping to build a competitors, many have joined Joyent. They may continue to do operating system work there via the illumos project, but their business will be the cloud, not distributable operating systems.

There are a variety of reasons for the decline in visbility for the OS; most are likely familiar to you. The ascendance of cloud, the delivery of applications in a network context, open source led shifts in availability and procurement, the accelerating fragmentation of application development, historical trends towards further abstraction, platform consolidation. All are contributing to a shift in the way that applications are designed and deployed, which in turn impacts the role of the operating system [coverage].

The question of whether operating systems still matter is probably a wrong one: the new platforms, cloud included, cannot function without the core functionality an operating system provides. The better question is who operating systems matter to, and what this means for the industry.

It is self-evident that operating systems matter – intensely – in the mobile space. Clever solutions like PhoneGap aside, developing for Android is not like developing for iOS. Which in turn is not like developing for the BlackBerry Tablet OS. The mobile landscape is akin to server side development circa 1995, where applications are written directly to the operating system with portability an expensive afterthought.

It’s also true that operating system adoption on the desktop remains highly stratified; with the long predicted Year of the Linux Desktop yet to arrive, consumers PCs are still primarily either Mac or Windows.

The enterprise server world is a more mature market than either mobile at least, and has therefore already weathered the rounds of consolidation yet to arrive in the world of handsets and tablets. There are fewer legitimate contending platforms in the server market, even as at least one of those (Linux) shows signs of pending fragmentation.

Even these fewer choices, however, are increasingly abstracted from application developers. What is the operating system behind Microsoft Azure, Engine Yard, Force.com, Google App Engine, or Heroku? The answer is, in varying degrees, none of your business. These and other PaaS vendors are aggressively distintermediating the operating system by removing it as a choice.

Nor is Infrastructure-as-a-Service, which generally still implies a choice of operating system, reflective of traditional enterprise operating system adoption patterns. What the available cloud metrics suggest is an environment where traditional drivers for operating system adoption – commercial support, certified application compatibility, etc – are second to convenience. According to the Cloud Market, which provides relative volume metrics on the Amazon AMI catalog, this was the approximate composition of major operating systems on November 9th.



It’s true that this is a survey of available images rather than deployed images. But it’s equally true that this distribution would have been unimaginable even twelve to eighteen months ago: the leading commercial distributions make up only 18.5% of the catalog. Given that we’re still seeing a distribution in the catalog, it’s clear that the operating system matters to cloud deployments. People value choice enough to have created better than 7200 different options. Precisely what matters about those operating systems, however, appears to be in transition.

Within the datacenter, on the other hand, little has changed. Which is fortunate for suppliers, because the majority of revenue is still derived from infrastructure maintained and run by businesses themselves. Enterprise purchases still overwhelmingly favor traditional operating system suppliers: Microsoft (Windows), Red Hat and to a lesser extent Novell (Linux), with the majority of the remaining share divided between Unix vendors (HP, IBM, Oracle, etc). The dramatic spike in interest in Red Hat Enterprise Linux 6 speaks to the continuing appetite for new operating system features in a variety of markets.

As long as the enterprise markets remain the dominant revenue source, then, operating system vendors have little to fear. The challenge – still theoretical – will be when the shift in developer interest we are beginning to observe begins to manifest itself within an enterprise context. At present, definitive statements that the importance of the operating system will be as diminished as the commodity hardware platforms they most frequently run on today are purely aspirational. On the other hand, who would have predicted fifteen or even ten years ago that one of the major players in the market, a $750M business, would be an entity that sells software that’s available for free?

The data we have argues that change is coming. That it is, in fact, already here; it’s just unevenly distributed. It will be interesting to see how the various operating system suppliers leverage, fight or succumb to it. Just as fascinating will be whether the incumbents are able to pivot neatly from one dominant role to another, or whether that process will advantage the contenders. History is no guide here, as we’ve seen both scenarios play out repeatedly.

Disclosure: Canonical (Ubuntu), Microsoft, PhoneGap and Red Hat are RedMonk clients. Novell and Oracle are not.

The past few years, as I’ve written about, I’ve worked primarily off Thinkpad laptops, with an aging Sun Opteron workstation available for more computation heavy tasks. Neither of those pieces of hardware, however, is up to the workloads I’ve been engaged with lately. Virtualizing multiple operating system instances, working on large datasets, or tinkering with big data software such as Hadoop. To be honest, it was all either machine could do to run Chromium with my usual 60+ tabs open. Sure, the cloud helps, but when you can barely keep a browser up and running, it’s time for new gear.

Hence my call to Dell, who for the sake of full disclosure is a customer of ours. In response to my inquiries, Dell shipped me a loaner top of the line workstation to test, the Dell Precision T7500. I’ll have more on what, specifically, the machine is for later. For now, a quick rundown on the specs, setup and software choices.

The Hardware

This beast comes equipped with two quad core Intel Xeons running at 3.2 GHz, 24 GB memory, and 3×300 GB 10K RPM hard disks. It’s easily the most powerful box I’ve had locally since my mainframe days, in other words. Plugged in to this are the 30″ and 24″ monitors I already had on hand.

The Operating System Choice

The first thing I did when I got the box was try to get my beloved Thinkpad USB keyboard working with the Windows 7 Home Premium instance preloaded on the workstation. I failed. Even after manually installing drivers from the CD that came with the keyboard, Windows insisted that my U, I, O, J, K, L and M keys were instead 4, 5, 6, 1, 2, 3, and 0, respectively. So rather than waste more time tinkering, I gave up and installed Ubuntu. Being a latest and greatest guy, I picked the still Alpha Lucid release of the distribution.

As an aside, please note that I’m sure that Windows can be made to work with that hardware just fine, and that I’m not recommending that people use Ubuntu simply because they can’t get a piece of hardware to work. If you like Windows, use Windows. I happen to prefer Ubuntu, so that’s the context for this decision. Your mileage may vary, as always.

What Works

Anyway, Ubuntu recognized the keyboard perfectly, to the point that even the volume up/down/mute buttons work properly. Everything on the machine works out of the box, actually, with but a few exceptions. A quick rundown of the hardware:

• Wifi: Atheros AR5001X+, just works
• Graphics: NVidia Quadro FX 3800 just worked with the single 30″, had to enable the non-free drivers to get compositing working and the 24″ inch online as a dual monitor
• Sound: Sound Blaster X-Fi XtremeMusic (D), just worked
• Internal USB card reader: just worked
• External Hard Drives: 2xSeagate 1.5 TB, just worked (this bug has been satisfactorily addressed for me)
• iPhone: while I don’t use this functionality, Lucid sees my iPhone perfectly, will play music off of it, and even offers primitive music support

System-wise, Ubuntu 64 bit sees all of the available memory and cores correctly as evidenced by this htop capture. As mentioned, I had to enable the proprietary NVidia driver to get the fancy graphics and second monitor working, but the driver installation is completely automated.

Simply put, Ubuntu Lucid 64 bit pretty much just works on the T7500, at least with the configuration I chose.

The Software

From there, I did my usual Linux install. First, the software to be installed:

1. Amazon MP3 Downloader
2. Chromium (my default browser now)
3. Deluge (my now default Bittorrent client)
4. Dropbox (a staple of my existence these days)
5. Emacs (my editor of choice)
6. eMusic Download Manager (for downloading eMusic tracks)
7. Flash (still necessary)
8. GNOME Do (a Quicksilver like application for Linux)
9. htop (top, but pretty and visual)
10. revolution-r (R, in other words, for statistical analysis)
11. VirtualBox (a really excellent free virtualization package, I only wish that a.) I could resize hard drives and b.) that Aero would be enabled for Vista/Windows 7 as it is both in Parallels and VMware Workstation/Fusion)
12. VLC (will play anything, as they say)

Eventually I’ll get around to installing all the infrastructure stuff I use to test like Apache, MySQL, and so on, but these are the day to day basics I need. Next, the software to be removed:

1. Evolution (don’t use a mail client, and don’t particularly care for Evolution)
2. OnBoard (don’t need it)
3. Transmission (my experiences with this Bitorrent client has been very poor)

The Configuration

After using one of Bisigi‘s Pretty Themes for a while, I’ve cut over at least for the time being to one Ubuntu’s new “Light” themes, Ambiance. I know some people are a little bent out of shape about the window controls, but I just assumed what we were subsequently told: that they looked heavily at how existing operating systems did things, Apple in particular. While I agree that not everything Apple does is perfect, the fact is that they’ve invested a ton of time and energy into user interface research over the years, and they are, at least in my view, the best in the world at UI. Meaning that if Apple believes the controls should be on the left, I don’t think it can hurt to try it.

Because if I decide I don’t like it, I don’t have to use it.

Finally, a bit of quick configuration to pull in my emacs settings and so on.

1. ln -s ~/Dropbox/.bash_aliases .bash_aliases: pulls in my bash aliases from my Dropbox copy
2. ln -s ~/Dropbox/.emacs .emacs: pulls in my .emacs file from Dropbox
3. ln -s ~/Dropbox/emacs emacs: pulls in my emacs directory (w/ themes, etc) from Dropbox
4. ssh-keygen -t rsa: generates an ssh certificate for the box
5. ssh-copy-id username@host: copies my certificate to our various servers so I don’t have to log in each time

And that’s about it, apart from migrating a few VirtualBox harddrives from the old workstation to the new one.

I had very few problems, and having done this multiple times most of the above takes less than five minutes of effort because it’s all handled by package management tools. Everything wasn’t perfect, however.

The Issues

1. Amazon still doesn’t provide 64 bit versions of its MP3 store downloader, and the usual fix of using getlibs didn’t work on Lucid. Nor did Pymazon, a Python based alternative, work properly. Still don’t have a fix for that, as I’d prefer not to get in the habit of copying libraries around.
2. Brasero, meanwhile, the CD/DVD burning tool broke as it always does during Alphas of new releases, so I’m temporarily using the less user friendly GNOMEBaker to burn CD’s and DVD’s.

Other Comments

Two other items of note for the Ubuntu geeks in the audience. With this migration, I’ve officially dropped Banshee in favor of Rhythmbox. I liked Banshee, and still prefer it in many ways, but after a couple of ugly crashes that corrupted the library thus losing my playlists, I needed a replacement. Rhythmbox isn’t perfect, but it works and is nicely integrated into Ubuntu.

Second, I haven’t (yet) installed Pidgin, the IM client that OS X’s Adium is based on. In part because Ubuntu has transitioned to Empathy and because some of the underlying technology is interesting, I’m giving that project a shot. But there are some serious usability issues with the interface, and how it’s woven into the Ubuntu desktop. The integration into the Me Menu is suboptimal, the user interface – specifically its usage in the Indicator applet – is terribly confusing, and the account creation process was clunky.

It’s not clear to me that Empathy is ready for prime time from a UX standpoint, so I’ll be curious to see how that aspect evolves within Ubuntu and the other distributions that leverage it.

Overall Impressions

The T7500 is just stupid fast, and Lucid’s a nice interface for the hardware. I don’t have enough up and running yet to do any legitimate comparative benchmarking versus my usual hardware, but it’s impressive even on trivial applications. The disk usage analyzer, for example, scans the entire filesystem in less than ten seconds; with either of my old machines, runtime was a minute to two, depending on what else was running. The rendering of an eight minute video in Pitivi, the video editor included with Ubuntu, took about forty seconds. Chewing through the entire works of Shakespeare to count the frequency of the word “Zounds” using Hadoop took about fifteen seconds, but that was on a virtualized instance with more limited resources. And as as you can see from the screenshot above, virtualization is not much of a challenge for this machine.

I’ll have more on how the box will be used both on virtualization and big data later, but for now the Linux compatibility report for the hardware is excellent, as is the performance.

Like Jeremy Zawodny, I’ve had persistent issues with Intel Video since the Jaunty release of Ubuntu. The short version is that the fancy effects that I eventually came to depend on stopped working after the Intrepid release of the distribution. Both Jaunty and Karmic include Intel video drivers – and I’ve tried every bleeding edge PPA version I can find – that simply don’t work with Compiz. For those of you running similar Intel video chipsets (my machine is a Thinkpad X301, which reports the hardware as: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller), then, here are the current options for composited desktops on Ubuntu.

Compiz

Normally, this can be easily activated by selecting System:Preferences:Appearance:Visual Effects:Normal/Extra. The Intel video drivers included with Jaunty and Karmic, however, include a regression which disables Compiz. What you’ll get is a dialog box that says “Searching for available drivers,” a refresh of the display, and no Compiz. As mentioned above, I’ve tried in the past to use the bleeding edge repositories for Intel drivers, but have not had success making Compiz work.

I lied: Compiz now works just fine. The solution? Reinstallation of the Compiz package and its assorted dependencies. I have no idea why this should work, but it worked for someone else and it worked for me. So right now, then, Compiz is your best option.

Metacity

The good news is that composited Metacity – the software Compiz was intended to replace – works. The bad news is that it doesn’t work well. Effects can be enabled, as you can tell from the glassy and transparent panel of GNOME Do, pictured. But enabling them introducing a noticeable and unacceptable lag; ALT-TAB behavior includes a perceptible, unfortunate delay from command to window presence.

If you want to enable this, ensure that Compiz is turned off and execute: gconftool-2 -s '/apps/metacity/general/compositing_manager' --type bool true. And before you ask, yes, false will turn it off.

GNOME Shell

The most interesting of the three options, GNOME Shell is the project upon which the next iteration of the GNOME desktop – version 3 – will be based. While the developers are likely to push the release date out from March to September, you can easily build and test it out on your desktop now.

Like KDE 4, which was a substantial revamp of the previous version of that project, GNOME Shell is a fairly radical rethinking of the desktop paradigm. I like some features more than others – the application activation paradigm suffers greatly in comparison to the non-intrusive GNOME Do, in my opinion – but it’s certainly an attempt to push things forward.

The most interesting aspect of the project, from my perspective, is Mutter. Mutter is essentially Metacity with the Clutter technologies baked into it. As Joe Gregorio noted back at OSCON, Clutter’s a popular project these days, and for good reason. Built by the Intel-acquired OpenedHand team, among others, Clutter offers aesthetically pleasing, thoroughly modern graphics capabilities that, even better, are fast. Indeed, of the available compositing options for my hardware, the only one that performs adequately is the next generation option in the Clutter based GNOME Shell. That is most unusual; the bleeding edge stuff is typically very slow.

I’ll have more on Shell in future, but for now, just be aware that it may be – depending on your hardware – the only compositing option that works.

Buoyed by the success I reported running Karmic on an antiquated X40 (and this positive report), I decided to take the plunge this past weekend and upgrade – against the recommendations of many, given that a beta is still two weeks away – my production laptop, a Thinkpad X301, to an alpha version of Ubuntu’s forthcoming Karmic Koala release. While I would not encourage anyone to follow my example at this juncture, because I’m experiencing the typical growing pains of ever-changing, still alpha software, there is good news.

Herewith, a quick review of Karmic in its pre-release state.

What Works

Pretty much everything, happily. While it must be acknowledged that I did an in place upgrade rather than a fresh install, and thus am not subject to some of the more aggressive fundamental changes such as the default filesystem, the upgrade experience was relatively painless. With the exception of the bluetooth modules, which I have yet to test but are detected properly, all of the on board hardware is supported seamlessly and out of the box. Wifi, audio, suspend, the video hardware: all good, for the most part. Even the famously problematic on board Ericsson F3507g WWAN card is working for me as of the 2.6.31-10 kernel. Indeed, this post is coming to you via the magic of that function.

What Doesn’t Work (and How to Fix It)

With the caveat that many, if not most of these, are because I’m running alpha software, here are a couple of gotchas.

• Banshee:
  Certain applications seem to make Karmic mad, and Banshee – version 1.4 – is one of them. My instance would start up and play music, but crash when I did nothing more significant than clicking on a playlist. The fix seems to be upgrading to 1.5. It’s not in the official repositories yet, but there’s a PPA with daily builds available for Karmic users here.
• Compiz/Intel Video:
  This will vary from hardware platform to hardware platform, but the Intel Video drivers have been iffy recently and Karmic continues this trend. One day Compiz – or now Compiz-Fusion in Karmic – will work, the next it will be broken and you’ll be left with a sad, no-effects desktop. When it breaks, it can be disruptive, because I depend on seamless and fast switching between multiple desktops. The fix for this, unfortunately, is to wait for updates; I’ve tried all manner of Xorg.conf manipulation and driver swapping, but it’s safest just to wait for a working update.
• Firefox:
  This one is my fault rather than Karmic’s. The new Ubuntu version, as is standard with each release, upgrades the Firefox instance from the 3.0 that shipped with Jaunty to the latest and greatest 3.5.3. As part of that upgrade process, I made the mistake of having it import my settings – bookmarks, etc – from my 3.0 release rather than the other 3.5 build I’d installed from the PPA here. Rather than seek to correct this problem through importing backups, I decided to solve the problem with Weave, Mozilla’s browser syncing plugin. Directing Weave to erase my local data and replace it with that it had stored in the cloud, it successfully completed the first task but failed at the second. Unable to sync the data successfully from the cloud, I’m thus left with a hollowed out Firefox 3.5 – one with its memory wiped, and nothing to replace it. To their credit, the Weave team has been very responsive on Twitter and email, but in the meantime, I’m relying on Chromium more heavily than ever.
• NetworkManager:
  NetworkManager – what most users will know as the wifi applet in the upper right hand corner – has undergone a few changes in Karmic, most notably being migrated from an /etc/init.d script to an Upstart job. NetworkManager may or may not come back for you laptop users after suspend; in the event that it fails, try sudo /etc/init.d/network-manager start instead of the old sudo /etc/init.d/NetworkManager start. This has worked for me thus far.
• Suspend/Resume:
  I claimed above that suspend/resume works and so it does. The catch is that while the machine resumes properly, not everything else comes back so gracefully. I’ve had issues with everything from Window Decorations as discussed below to NetworkManager when returning from suspend. It can be messy coming back, although it’s gotten better with the last few days updates. The fix here depends on what’s wrong, of course.
• Window Decorations:
  I am periodically losing my Window Decorations – the top borders of windows – following a resume from suspend or a straight restart. This is problematic not only aestheically but functionally, because you lose the ability to move applications around the desktop. The fix, fortunately, is simple: head to Appearance under System, click the last tab, “Visual Effects” and select Normal or Extra. This may fail, but even if so it will act to reboot, so to speak, the GUI and in so doing it’s restored my Window Decorations every time.

Look and Feel

I’ve argued for a while now that one of Canonical’s primary contributions to the Linux distribution world is in the polish, the look and feel of the desktop. Karmic continues this tradition, with an attractive, aesthetically pleasing UI. While some of the decisions – the introduction of the Growl-like notification system, for example – have been controversial in some quarters, I find them to be welcome additions to a rapidly improving user experience.

To that same end, I’ve made two changes to my desktop in recent weeks that might be of interest to others running Ubuntu. First, I changed from the now-standard Dust theme to the more refined if unfortunately named “Wild Shine.” You can pick that and other newer eye candy themes from the PPA here. The other change I’ve made has been in the font department. For the first time since I started using Ubuntu I’ve strayed from the default Sans and Sans Mono fonts to the Red Hat authored Liberation Sans and Liberation Sans Mono fonts. A small change, to be sure, but one I’m happy with.

Ubuntu still is not, in my mind, a match for the top to bottom polish of an OS X, but the days of Linux desktops being boxy and unattractive – the old Volvo of the desktop world – are over, in my opinion. Those were the days.

When it’s compared to Microsoft Windows on the subject of driver support, Linux is generally perceived as lacking, although far less so in recent years thanks to the efforts of Intel and others. But in at least one case, Linux was the solution to a bizarre and as yet unsolved driver issue.

While I’m on a Thinkpad X301 at the moment, which coincidentally also runs Linux, the machine that it replaced – a Thinkpad X40 – is still kicking around. Limited by an anemic processor and a tiny 40 GB hard drive, I thought it would potentially still be useful as a netbook-style machine: handy for internet usage and the occasional work maintenance when my primary work machine was unavailable (like this past Saturday, when we were hacked).

Because my fiancée would occasionally be using it as well, I opted first for Windows XP, which would be most familiar to her. That installed seamlessly on the machine, and all the hardware – the wireless included – worked just fine, with one exception. Oddly, the Intel wireless on board refused to connect to the brand new Cisco (Linksys) access point we’d set up at our apartment. Absolutely refused. The same hardware would connect to any other network – even the pointless WEP secured one at the office – but it would not under any circumstances associate with the home network. It could see it, but simply not connect. At first I thought it was the WPA2 security, but that didn’t seem to make any difference: even completely unsecured, the network would not cough up an IP.

This not being the first time I’d experienced some wireless wonkiness from XP, I decided to abandon it in favor of its newer cousin, Windows 7. When the RTM build I applied finished installing, however, neither the sound nor the wireless worked. After Googling for a solution, I applied the Windows Vista drivers for the wifi hardware and it worked perfectly…except on the home network.

At this point you might be thinking it might be the home network, rather than the computer. That’s what I would have thought, except for the fact that both of our iPhones and my X301 running Ubuntu could connectly easily. So the problem, whatever it was, lay within either the machine or Windows. Or more accurately, the Windows drivers. Both the XP and the Vista flavors.

With no real requirement for Windows other than my fiancée’s convenience and familiarity, then, I decided to give Ubuntu a shot at the job and see if it was the hardware that was the problem. Having come off some separate Intel driver issues with Jaunty, I skipped the current production release and went to Karmic. The verdict? Works flawlessly. I’ll be damned if I know what the problem was with the Windows drivers, but with a working alternative I doubt I’ll ever try and find out. I’ll wait to see if Windows 7 drivers are released, and if they are, I’ll see how they fare.

Until then, how is the fiancée adjusting to Linux? “I kind of like it,” she says.

Maybe I won’t swap it out.

While it’s true that you hear it here last, generally, a week is a bit much, even for me. But as we’re still fielding questions about the news that Microsoft had filed a complaint over alleged infringed patents against TomTom, Dutch manufacturer of navigation systems, it seems necessary to comment. So in service of said need, please find below a Q&A exploring my current understanding of the complaint and the issues it raises.

Q: Before we continue, is there anything to disclose?
A: Yes indeed. Microsoft, the plaintiff in the case, is a RedMonk customer, while TomTom, the defendant, is not. Of the related commentary below, the Linux Foundation is also a RedMonk client. That should about cover it.

Q: Can you summarize the news for those that haven’t seen it yet?
A: Sure. TechFlash’s Todd Bishop does an excellent job of summarizing the news for you here:

Microsoft filed suit against TomTom [on February 25th], alleging that the in-car navigation company’s devices violate eight of its patents — including three that relate to TomTom’s implementation of the Linux kernel.

For the legally inclined, he’s even got a PDF of the complaint for you, with the US District Court version available here and the International Trade Commission flavor here.

Q: So what does that actually mean?
A: That is the $64,000 question. At least where $64,000 reads as potentially serious amounts of money. The initial speculation – as hinted at in Bishop’s summary above – was that this was the opening salvo in a long awaited patent war, which would pit Microsoft and its patent portfolio against the Linux kernel and its various contributors and defenders. Upon further review, however, most have concluded that this is not the case.

Q: Before we get there, can you provide some background on this “long awaited patent war?” Why would people expect Microsoft to litigate in this area?
A: Love the patent system or hate it as it relates to software – full disclosure: I personally hate it – it has been established that software is in fact patentable. As a result, virtually all large software companies – Microsoft obviously among them – actively and aggressively pursue software patents. More specifically, in this case, Microsoft has in the past stated that its belief is that certain open source assets – including the Linux kernel employed by TomTom – read on patents that it holds.

In a Fortune article from 2007, then-”licensing chief” Horacio Gutierrez had this to say on the subject:

Gutierrez refuses to identify specific patents or explain how they’re being infringed, lest FOSS advocates start filing challenges to them.

But he does break down the total number allegedly violated – 235 – into categories. He says that the Linux kernel – the deepest layer of the free operating system, which interacts most directly with the computer hardware – violates 42 Microsoft patents. The Linux graphical user interfaces – essentially, the way design elements like menus and toolbars are set up – run afoul of another 65, he claims.

As you might imagine, these claims did not make people within the Linux community specifically and the F/OSS world particularly happy, especially given the fact that they are long on claims and short on details.

Since that public interview, many in the industry have been expecting Microsoft to strategically attack one or more of the allegedly infringing projects. It was in this context, then, that the news that Microsoft was litigating around the Linux kernel was read as a shot across the bows of Linux users everywhere.

Q: So it is not a shot across the bow?
A: Probably not, although the perception that it is could certainly be argued to serve Microsoft.

Q: Let’s come back to the latter claim, but first let’s explore why this is not believed to be an attack on Linux specifically or open source generally?
A: We can begin by setting aside Gutierrez’ claims that this is about TomTom rather than open source – “I think there shouldn’t be any ambiguity on our expectations as a company. We recognize that open-source software will continue to be a part of the industry” is one of the quotes attributed to him – because that’s precisely what you would expect a spokesperson to say in a case like this. It’s interesting to note, as an aside, that Gutierrez was recently promoted.

Anyway, my read – as I related to Steven J. Vaughan-Nichols – was that this was not the long awaited patent offensive because it was too oblique. As the inestimable Andy Updegrove observes:

For benchmarking purposes, it’s important to know that while Microsoft has spent a great deal of time urging vendors over the years to take out licenses to multiple patents, including those that it claims are infringed by Linux (and other open source software), it has almost never sued anyone, ever, for infringement of any of its patents relating to any of its products. But it does make its expectations very clearly known to those vendors that it drops in to visit. According to Microsoft, its five year old licensing program has resulted in licensing deals with over 500 vendors. Hence the question that everyone is asking today: why now, and why against this particular company?

On the latter question, I think it’s worth noting what products are involved. None of Microsoft’s flagship products are involved. Rather, it’s GPS mapping software (TomTom competitor Garmin Ltd. has reportedly signed a licensing agreement with Microsoft already for MSN Direct, which provides traffic, flight and weather information to GPS devices.

So while neither I nor Andy, apparently, expects this to be a major shift in Microsoft’s approach vis a vis patent litigation and open source, I would still argue that the choice was anything but accidental.

Q: How do you mean?
A: Much like the Fortune story from 2007, Microsoft can potentially benefit here by sowing Fear, Uncertainty and Doubt even if – as is presumed here and elsewhere – Linux is not the actual target. The initial media coverage, after all, was heavily focused on the Linux/patent angle, at the expense of other angles. That could benefit Microsoft even while costing them.

Q: What does this cost Microsoft?
A: Each time Microsoft raises the spectre of patent litigations around open source – or, as in this case, others do it for them – they take a step or two backwards in their ability to work effectively with a variety of open source communties. While litigation is often how deals are done with commercial entities, communities react very differently.

Microsoft, for the record, does not agree with the initial contention, however.

Q: In what way?
A: I put the following question to Microsoft last week:

Beyond the explanation provided – that this is a dispute with TomTom rather than the community behind the Linux kernel – is it Microsoft’s contention that the involvement of Linux in the dispute is purely coincidental and not at all relevant to the complaint itself?

Their answer read as follows:

Open source software is not the focal point of this action. This case against TomTom, a global commercial manufacturer and seller of proprietary embedded devices, involves infringement of Microsoft patents by TomTom devices that employ both proprietary and open source software code. It is not unusual for companies to develop products based on a mix of proprietary and open source code; like every other company, they must take responsibility for ensuring that their systems do not infringe others’ patents. Licensing agreements are a useful means for ensuring mutual respect for IP and in fact, Microsoft has licensing agreements in place with many companies that run mixed source software environments, such as Samsung, LG Electronics, Fuji-Xerox, Brother, and Kyocera Mita.

Q: What about the specific patents at issue?
A: Many, predictably, seem foolishly vague: you can patent a “Vehicle computer system with wireless internet connectivity?” Really?

But I’ll leave those to others: the three that are most germane to Linux are the implementations relating to the FAT filesystem. USPTO #’s 5,579,517 (26 November, 1996) Common name space for long and short filenames, 5,758,352 (26 May, 1998) Common name space for long and short filenames, and 6,256,642 (3 July, 2001) Method and system for file system management using a flash-erasable, programmable, read-only memory, specifically (thx, Stephe). As Larry Augustin says, while the industry has long treated FAT as a standard, it is not, and as ridiculous as those patents sound – and they sound ridiculous here – they are in fact patented.

Which might be why Microsoft considers the patents sound in the first place, though, as is Andy, I’m very curious to see what happens now that they are going to be the subject for a potentially huge prior art search.

Q: How are Linux advocates reacting to the news?
A: Urging caution, but also calm. Jim Zemlin, for example, the CEO of the Linux Foundation, wrote this post whose title included “Calm Down, Hope for the Best, Plan for the Worst.” At the very least, Microsoft can expect a vigorous retaliation from a variety of corners should it attempt claims on the Linux kernel directly.

Q: What about arguements such as Groklaw’s contention that the Bilski en banc decision render some or all of the patents questionable?
A: For those that missed the original Groklaw reaction that referenced Bilski, take a look here. Curious about that myself, I put the following question to Microsoft:

Setting aside the issues many have with the current patent system, what is Microsoft’s level of concern regarding the validity of some of the patents in question following the Bilski decision?

Their reply was a terse:

We are confident of the validity of our patents in this case.

So I guess we’ll see.

Q: Any other reactions worth noting?
A: Well, I would check out old friend Stephe Walli’s piece, in which he argues that this about the mobile internet rather than Linux. Interesting reading.

But overall, I don’t see this as a landmark piece absent further related litigation. It’s a case of interest, certainly, one to watch, but not the other shoe that it was initially believed to be.

As many of you are no doubt aware from the coverage, we at RedMonk recently participated in a survey with the good folks over at Canonical to get a better sense of where the Ubuntu community is at at present. Not shockingly, the returns are interesting and clearly – in my view – worth exploring. The following Q&A will explore some of the more interesting questions, though I encourage you all to take a look at the survey data yourselves, as I’d love to hear what you extract from it. You can get that here; you do have to register, but it’s all opt-in vis a vis the contact aspects, which is good. Anyway, the Q&A.

Q: Before we begin, anything to disclose?
A: Indeed. Obviously we directly participated in the survey in question with Canonical – a RedMonk customer. No project funds were applied to our participation in this project. Separately, a number of Canonical competitors, including Microsoft, Red Hat and Sun are RedMonk customers. Last, RedMonk is something of an Ubuntu shop, with my personal desktops running the distribution as well the production server that hosts redmonk.com.

Q: Can you provide some context for the survey? How many results were collected?
A: According to the folks from Canonical, 6819 returns for the survey were collected.

Q: And what kind of audience was it, both in terms of its composition and their exposure to Ubuntu?
A: The survey itself was generally self-selecting, being most actively promoted to an Ubuntu audience (through the website, etc). As far as the composition of the audience, it’s skewed towards smaller deployments: while there were respondents running a few hundred machines, the majority were in the 1-50 category. And while ~15% (14.85%) of the replies originated in organizations with north of a thousand employees, the largest single demographic remained businesses with under 50 people.

Q: Any specific vertical focus?
A: Not really. Technology (Services), Technology (Software), and Education were 1/2/3 in the verticals reported, followed by Other, Manufacturing, Government, Non-Profit, Finance and Retail. Pretty decent spread.

Q: So what statistical significance does the survey have?
A: As a self-acknowledged survey of a self-selected audience, I don’t know that anyone would argue it’s useful for drawing wider market conclusions. What it is useful for, rather, is taking the pulse of the Ubuntu community. What do their workloads look like? What are they deploying Ubuntu for? And where?

And so on.

Q: What are other folks saying about the survey?
A: The most common comment I hear – see Dave’s post as an example – is the assertion that Ubuntu is now ready for mission-critical workloads. And based on the results, it seems clear that the Ubuntu audience believes this to be true: the worst showing statistically had better than eight in ten believing that Ubuntu is ready for mission-critical tasks.

But while that’s certainly positive news for the Ubuntu community and their commercial backers Canonical, because of the nature of the audience, I’m actually more interested in other aspects of the survey.

Q: Such as?
A: The diversity of use cases, for one. As I’ve told the media in commenting on the survey, one of the criticisms generally leveled at the Ubuntu distribution is that it is relevant only in the desktop context. And while I’ve never subscribed to that idea – not least because we employ it at RedMonk in server capacities – it’s nonetheless been a persistent myth.

The data obtained via this survey, however, indicates that Ubuntu is being leveraged against workloads that have little to do with desktop use cases. Among the workloads running on Ubuntu servers, respondents mentioned clustering, CRM, data mining, database, directory server, media streaming, security (firewall, IDS, etc), and web server. Specific applications cited include DB2, Domino, Oracle, SAP, WebLogic, and WebSphere. Granted, the sample size is small, and the overall volume of customers leveraging Ubuntu in this capacity is limited.

But it’s still indicative, to me, of a distribution that is increasingly deployed in a server context.

Q: What else did you find interesting?
A: There are lots of interesting tidbits. Take virtualization: the top four replies to the question “If you use your Ubuntu servers as a host for virtualization, which product(s)/technology(ies) do you use?” were, in order, VMware (by a sizable margin), Xen, none and KVM. From which you might conclude that a.) VMware remains well positioned in that market, b.) Xen is surviving in the face of strong competition from KVM, and c.) that in spite of (b), KVM is coming on strong.

I also found it interesting that package/update management was the second most important feature – behind proven security – for mission critical deployment. While I hate to keep banging this drum, I still find it bizarre that Linux distributions – all of them – undermarket their abilities in this regard.

Q: What about the questions regarding cloud usage and, more specifically, Ubuntu’s readiness in that capacity?
A: This was one of the areas of the survey that I was most curious about, actually. Of the seven thousand-ish respondents, better than seven in ten answered “No” to the question “Do you plan to deploy workloads of any type to cloud environments?” Which was interesting. Equally interesting, however, was the fact that better than six in ten considered cloud environments ready for mission-critical workloads (eight and ten said Ubuntu was a viable platform for cloud deployments).

Q: Your general takeaway, then?
A: The survey is precisely what I expected: an interesting cross-section of the Ubuntu community. I’m still crawling through the statistical data that Canonical kindly shared with us, so look for further insights at a later date. But for now, I recommend giving it a read. Self-selecting audience or not, the report is an interesting datapoint.

With the twin caveats that I’m an Ubuntu user and the admission that I have yet to actually hold a device of any type actually running Google’s Android software, I have to be honest: all this talk of Android netbooks baffles me.

Not the part of getting it to run on the platforms, of course: the appeal of the challenge is obvious, even to me. No, what I can’t quite grasp is this: what need would Android fulfill in the netbook market that couldn’t be better and more efficiently served by alternate platforms. Why anyone would want one, in other words, and favor it over either Linux or Windows? And I suppose I should throw OS X in there as well, as I’m told that it will run on the MSI hardware.

No question, Android is an interesting project from virtually any angle. The Linux base layer gives it an appeal to certain audiences, the cleanroom JVM reimplementation is fascinating for any number of reasons, and its market significance cannot be overstated…as long as we’re talking about handsets. But I must be missing something that makes it suitable for the Netbook market, because from here it just looks like a poor idea.

Why?

It’s all about the applications. As usual.

Look, I’m no big fan of rich clients, but to paraphrase Jason Priestley’s character from Tombstone: “I’m sorry, we got to have some apps.” Android’s got a browser, true. But it’s not Firefox. And while handset users are used to not having any choice in the matter (if I could use something besides mobile Safari, which crashes like that’s its job, I would), desktop users are. If I’m going to be using a desktop – and especially if I’m going to rely heavily on the browser as an application platform – I’d like it to be one that I choose.

Why doesn’t Android – and why won’t it – have Firefox? Because Android applications have to run on top of Dalvik, Android’s JVM implementation. So until you reimplement Firefox on top of a JVM or find a way to bypass it to access the underlying operating system, no Firefox. Which means you can count out OpenOffice, too. And Pidgin. Probably Skype as well. And so on. Virtually every staple application Linux or Windows users are used to using on their desktops will be unavailable.

To be sure, there will be Android alternatives available. Just as many of the Linux applications are alternatives to their Windows cousins (think OpenOffice). But who’s going to be building for Android on netbooks, at least initially? My guess is that the bulk of Android development will be focused on the handset market, not netbooks. And while I admire some of the iPhone application implementations, I certainly wouldn’t use them over a version designed for my Mac Mini.

Application volume would seem to be a massive issue for Android based netbooks. There are tens of thousands of applications available for anyone running, say, Ubuntu Netbook Remix. How many are there for Android? Dozens? Hundreds? And that’s just looking at the Linux apps: imagine what the Windows crowd might think. Who’ve become accustomed to using one of the myriad of applications available for that platform.

Ultimately, the question I can’t answer is this: if you had to choose between:

a.) Linux and the C, C++, Java, Python, etc applications of your choice and
b.) Linux and the Dalvik applications of your choice

Why would you pick b? Or put more simply, what precisely does the reimplemented JVM buy you on the client side? What about Android will make Java desktops succeed where they’ve failed so often before? And no, user experience doesn’t count: you don’t need a JVM to have a good UX.

Anyway, if you’ve got an answer to that question, I’d love to hear it, because as I said, I’m kind of at a loss.