Wednesday Jul 02, 2008

Pedal-pup-propulsion, Novell style

Joe "Zonker" Brockmeier, the puppy-kicking Novell community manager for OpenSUSE, has clearly been reading from the same press briefing as Novell CEO Ron Hovsepian.

At the Red Hat summit in Boston, pooch-punting Zonker (not being disrespectful, I understand Zonker prefers to be called "Zonker") revealed that he is "very disappointed" that Sun hasn't "bitten the bullet and participated in Linux". By which he means, presumably, the Linux kernel, as Sun certainly does participate in many projects which are upstream of Linux distributions; indeed, it's been recorded that Sun has more code in the average Debian distro (including the software in the repositories) than anyone else.

Beagle-booting Zonker's comments also seem to echo criticisms made of various popular Linux distros by employees of companies who produce less popular, green-themed Linux distros when they complain that the work is all downstream: "The work that's done for Solaris is done only for Solaris." Well, let's get two things clear:

1. a lot of the work funded by Sun on OpenSolaris finds its way into projects upstream from Linux, including, GNOME and important applications
2. a lot of the work funded by Sun on OpenSolaris finds its way into other open and closed-source projects including, DTrace into Linux

Zonker, if you must kick that puppy, please remove your misinformation boots first.

ps. the views expressed here are not necessarily those of my employer

Tuesday May 27, 2008

Ghost Writer

Seriously, is there a worse technology journalist than Sam Varghese? His column Open Sauce - A GNU Perspective (geddit?) is a train-wreck. He's very poorly informed, self-important (he considers that if he gives a product a poor review, it's "a major PR disaster"), but what's even worse: he simply can't write.

This is the open-minded individual who insists on writing "(sic)" when quoting americanised direct speech. This is the wordsmith who crafts nonsense like Bruce Perens made what would be a move with huge repercussions, and the brainless Stormy by name, not by nature (on Stormy Peters of OpenLogic). I could go on. And indeed, I have.

Anyway, I'm revisiting Mr Varghese's column after reading his review of OpenSolaris. I will admit that I was waiting for this one, fully expecting it to be littered with factual errors and what we might kindly describe as failures of intellectual curiosity. And I wasn't disappointed. His opening salvo:

One still has to go through a requester/sponsor arrangement to submit a patch to the OpenSolaris project. (By contrast, the Ubuntu Linux distribution started by Canonical is now a little more than three-and-a-half years old - and there is no need to detail what it has achieved).

We all admire Ubuntu, and no one is going to deny it has done a fantastic job in both gaining admirers within the Linux community and, crucially, growing the popularity of Linux with people who might otherwise have not used it. But still, (and ignoring the massive differences in starting points for these projects) the truth of Mr Varghese's statement really depends on who you are. There are, after all, non-Sun committers to OpenSolaris. Sure, we want more, but I challenge Mr Varghese to put back a patch to Ubuntu without a sponsor. One will always have to go through a request/sponsor arrangement unless one has committer status.

Mr Varghese then demonstrates his ability to work the Google machine, regurgitating criticism of OpenSolaris (much legitimate - we're not claiming perfection here) from IBM employees, and then his inability to RTFM by complaining that he can't find But then he unleashes his final barb,

But the licence is what jars the most. It pops up in all its glorious detail right at the start of proceedings, the Community Development and Distribution Licence. It isn't compatible with the General Public Licence, an indicator, again, that Sun is still in two minds - should we (really) give it away or should we still continue to be control freaks?

Like many people, I have a lot of respect for the GPL, but let's be clear: only the GPL is compatible with the GPL. And the CDDL is an open source license by any - any - reasonable definition, and certainly the one to which Mr Varghese appears to subscribe. Does he even read his own column? So when he writes:

There are two Sun components that would be of interest to Linux developers if they were licensed under terms that made them portable - the ZFS filesystem and DTrace. But by the time Sun decides on whether it will open source these two, it may be time for me to bid goodbye to this world.

Let's hope not, as that would date Mr Varghese's demise as January or December 2005. Happily, DTrace and ZFS are both now available on Solaris, Mac OS X and BSD - DTrace is even available on QNX. But should Mr Varghese read this (although the evidence is that he reads very little, at least, until he has decided what his opinion is), "open source" does not equate to GPL compatibility. If it did, why would we even have the Open Source Definition? We would just have the GPL.

ps. the views expressed here are not necessarily those of my employer

Tuesday Apr 01, 2008

Last time I checked

Novell CEO Ron Hovsepian, ("ill considered" - Redmonk), not content with the murkiness of his own company's patent arrangements with Microsoft (the two companies laughably "agreed to disagree"  on whether or not their November 2006 agreement was an ackowledgement that Novell was infringing Microsoft patents), has decided to throw some mud in the general direction of OpenSolaris.

"Rather than focusing on Microsoft, Zemlin pointed Hovsepian to Sun and Open Solaris, where his comments were unusually pointed.

"I believe OpenSolaris has had about 60,000 downloads last time I looked," he said. "When you look at Linux downloads just last year [there were] over two million of just SUSE.” Hovsepian also attacked the OpenSolaris license directly.

"I would suggest to the customers and to the community, be careful. The way they’ve written their contract as soon as you look at it, you can’t go back and look at Linux.

"It’s a very dangerous contract from my perspective for someone who wants to work on Linux."

Attacking the license?  Well, any old fule knows that only GPL'd code is compatible with GPL'd code.  But presumably, Mr Hovsepian is scaremongering that those who work with OpenSolaris code have been exposed to patented methods and may face encumbrances if they want to put back to another code base.  Is that what he's getting at?  (Incidentally, his criticism remind me more of Microsoft's covenant not to sue open source developers who use Microsoft-patented methods for "non-commercial" use than anything else.  Now that's dangerous.)

Anyway, last time I checked, Sun is pretty clear on software patent policy.  You can read about it from
Mike Dillon (General Counsel), Simon Phipps and again here (Chief Open Source Officer and, ahem, my boss), Greg Papadopoulos (CTO) and Jonathan Schwartz and again here (CEO).

Secondly, Ron Hovsepian's reported figure on OpenSolaris downloads wasn't too clear.  Did he mean source code?  Or binary distributions?  When did he last check?  And who did he check with?  Of course, since Sun opened the Solaris source code, there have been millions of downloads of Solaris 10, hundreds of thousands (if not millions) of downloads of Solaris Express, and I don't know how many downloads of source code, but I would put it in the tens or hundreds of thousands (I tried to calculate this, but gave up - you have the full tarball, and specific consolidations...hits per file, data transferred per gets messy and I couldn't generate a figure I could stand behind).  We've also had tens of thousands of downloads of the developer preview of the binary distribution coming out later in the year.  Which isn't bad for an operating system that hasn't been launched yet.

Ron Hovsepian

Ron, left, not right

But enough about us.  Last time I checked, Microsoft's intepretation of Novell's patent licensing agreement was,
"If a customer says, 'Look, do we have liability for the use of your patented work?' Essentially, If you're using non-SUSE Linux, then I'd say the answer is yes," -Steve Ballmer
More on this in the latest episode of the Register's Open Season podcast (27:20 minutes in) -this is not necessarily historical issue.

Now this isn't Mr Hovsepian's first ill-informed and misleading attack on OpenSolaris, but I don't think that's it's really our fault.  Simply put, he has a shortage of targets.  I mean, who else is he going to attack?  Given his position, he can't very well attack another Linux distro, and he's hardly going to attack Microsoft.  Apple?  For successfully porting of OpenSolaris technology into their operating system (without licensing any patents, we might point out)?  Doesn't really work, does it?

Open source needs better leadership that this.

ps. the views expressed here are not necessarily those of my employer

Monday Aug 13, 2007

Journalist, read thyself

Self-styled Cyber Cynic Steven J Vaughan-Nichols published this piece suggesting that recent developments in the IBM-SCO case may compromise the OpenSolaris project.

To be clear, I think that people like Steven J Vaughan-Nichols do an important job. There are times when self-conscious cynicism should be one's guiding light (as anyone who has recently wasted three hours of their weekend trying to get rid of a door-to-door vacuum cleaner salesperson can tell you). Moreover, I can't comment upon the legal merits of his argument, being under-qualified (i.e. not qualified at all), not authorised and generally lacking the requisite ambition. But I can comment on the article itself.

To wit:

"Sun's Jonathan Schwartz -- then Sun VP of software and today Sun's president and CEO -- said in 2003 that Sun had bought "rights equivalent to ownership" to Unix.

SCO agreed. In 2005, SCO CEO Darl McBride said that SCO had no problem with Sun open-sourcing Unix code in what would become OpenSolaris."

As I read this, Mr J Vaughan-Nichols seems to be implying that Sun believes it bought the rights necessary to open source Solaris from SCO.

However, clicking on the link he provides that quotes Jonathan Schwartz, one quickly realises that Mr Schwartz appears to assert rights equivalent to ownership to Unix based upon Sun's agreement with AT&T in 1992. Which is not the impression I got from Mr Vaughan-Nichols' piece. With that in mind, the article may, perhaps, seem rather less Slashdottable.

ps. the views expressed here are not necessarily those of my employer

Wednesday Aug 08, 2007

Something must be done

Disparaging? Possibly. Disingenuous? Never in a million years.

Linux kernel maintainer Andrew Morton, reports ZDNet, was rather dismissive of OpenSolaris at LinuxWorld yesterday, inviting all to rally behind the cause of, er, monocultures.

Well, we don't expect Mr Morton to trumpet the great advances that OpenSolaris has made (which, not unreasonably, many in the GNU/Linux community are unaware of), and indeed, leading figures in the Linux community have in the past made even gloomier prognostications for Solaris.

But Mr Morton is right on the money (in the parlance of our times) with this comment:

It’s a great shame that OpenSolaris still exists...They should have killed it...They've fragmented the non-Windows operating system world and they continue to do so

Quite right. And let it not end there. Surely Sun might be further castigated for the UltraSPARC T2, which is continuing to fragment the non-Intel processor world. And it's a great shame that the Opera browser still exists, for they are continuing to fragment the non-Internet Explorer world. And let us not forget Gmail, which has fragmented the non-Hotmail world. And I really think it's time that O'Reilly pulled the plug on OSCON, fracturing, as it is, the non-LinuxWorld world.

Mr Morton also goes on to say that SystemTap will eventually have all the capabilities of DTrace. For reasons why that may not prove to be the case, you can have a look at Adam Leventhal's recent postings here and here.

ps. the views expressed here are not necessarily those of my employer

Thursday Jun 14, 2007

Three Anniversaries

There are three anniversaries I'll be celebrating this week. First, it's 10 years since I joined Sun. My first job? Cold-calling customers who had stopped calling Sun. Yes, I was a tele-marketeer, and I can tell you, it is tough and thankless work. And so, whenever I get a tele-marketeer on the line at home, I try to be as sympathetic as posible while I tell them to get stuffed. Anyway, that was 1997 and I had just arrived in the Netherlands. I was looking for work and grateful for the opportunity, just as I'm grateful now that I could transfer with my family over to Sweden.

The second anniversary this week? OpenSolaris. It's 2 years today since it was launched on a suspecting public. As a user, I think the experience of an OpenSolaris distribution on the desktop is coming on in leaps and bounds. And although I don't want to dwell on the negative, I was pleased to hear Linus Torvalds talk about OpenSolaris this week, even if I don't especially like, or agree with everything he says. I think that contrasting his comments this week with those from February 2005 shows the relevance and success of OpenSolaris.

Let's hope he and Jonathan can have dinner soon.

Which brings me to my third anniversary: on the 21st, our daughter Emma turns one. There is a concept in philosophy called "qualia", such as yellowness, or the taste of coffee, the properties of the senses which seem irreducible and can only be described by themselves. I'd say that being a parent is similar: the feeling of parenthood can only be described, I am sure, by being a parent. I don't think that you know love, or fear, until you have a child.

ps. the views expressed here are not necessarily those of my employer




« July 2016