Open Source

Willow Garage’s move to sell copies of its PR2 to the public at $400,000 (discounts of up to 30% available to researchers) is being hailed as a turning point in the industry’s development. (Image from Willow Garage.)

That’s because, unlike most Japanese robots, which are either purpose-built or one-offs, and unlike iRobot (which despite its name is a cleaner and military contracting company), the PR2 is built from an open source software base, with modular hardware.

Willow Garage is not the only open source base on which to build a robot. Urbi went open source last month, with a C++ library and an API. CARMEN is Carnegie-Mellon’s open source robotics toolkit. Orocos also offers an open source platform for robotic control.

It’s clear, then, that such PC ideas as standards and open platforms are starting to pick up in the robotics industry, which until now has been proprietary, military, industrial and (frankly) a little inaccessible to the computing mainstream.

So what does the industry need now? I would suggest it needs an organization coordinating standards, providing compatibility among the various open source offerings, a base on which everyone can grow and develop.

Something like the World Wide Web Consortium.

The W3C has come in for its share of criticism, in the 16 years since the Web was spun. Its standards process moves slowly. Google’s bouncing balls, demonstrating HTML5 and CSS, were meant to speed up an adoption cycle that threatens to drag on another decade.

But the group does serve a purpose. It provides a base. Everything built up from the base refers back to the base. There are differences in how browsers render web pages, but these are minor because the base remains intact. Everyone building web pages innovates from a base of W3C standards.

Robotics needs a base.

Robotics needs a free, open source, universal platform on which to build. We need compatibility among the various open source projects now seeking programmer loyalty. We need a clubhouse in which to meet, a center, a base.

Anyone know how to build one? (Cough, cough, JimZemlin, hack. Excuse me.)

It took a while, and that delay may be telling in terms of the underlying issue, but the Free Software Foundation has finally issued a public statement on the Oracle-Google Java suit.

They side with Google.

(FSF logo from Wikimedia.)

Not that Google is entirely on the side of the angels, writes the FSF’s Brett Smith.

He says they could have avoided the problem by forking IcedTea, which is covered under the GPLv2. They’re still not on the side of the angels when it comes to software patents. But that Larry Ellison is one bad (shut your mouth). So the FSF statement says:

An aggressive infringement suit over software patents is a clear attack against someone’s freedom to use, share, modify, and redistribute softwareâfreedoms that everyone should always have. Oracle now seeks to take these rights away, not just from Google, but from all Android users.

Smith then calls on Google to use this case as a “come to Jesus” moment, and to defend itself by attacking the very idea of software patents as impractical and a very bad thing.

Over on the other side of the pond, long-time software patent foe Florian Mueller doesn’t think joining the IcedTea party would have made any difference. Here is what he wrote to me this morning:

The idea of those GPL defenders is that the right holder (in this case Oracle in succession to Sun) of a GPL’d work can’t use his patents against GPL-based forks. So in theory if Google had used Oracle/Sun’s GPL’d Java code as the basis for its own virtual machine, the FSF argues it could now claim Oracle/Sun made an implicit patent grant. However, where the FSF is wrong is that this would relate to forks.

That legal strategy might have worked under GPLv3, he adds, but that license has not been accepted by the industry specifically because of that patent clause.

Patents have become strategic weapons. They were created by courts, not the Congress, and both the courts and Congress have refused to reject them since. Now that every large company has an arsenal of such patents, they’re assets from which they wish to gain corporate advantage.

But the whole intent of patent, as with copyright, is to encourage the production of more and better stuff. It’s not to make the heirs of inventors or writers or film actors wealthy, but to make them productive.

This should be the test. Do Oracle’s software patents encourage innovation or discourage it? Would there be more innovation with the patents or without?

Patents are not a property right, and neither are copyrights. That last sentence should not be controversial, but you’d be amazed at just how revolutionary it appears to be in 2010.

If you try to contribute to a corporate open source project, especially a “dual core” project, you will probably be given a “contributor agreement” to sign.

The agreement gives all copyrights and patents to the project’s corporate sponsor. The wording can differ, which is why a group called Project Harmony is working to harmonize them.

This is a different issue from that of the license. Many projects licensed under the GPL are still subject to contributor agreements.

These agreements have their fans, and their purpose. They let business be done centrally, without having every minor decision subject to a veto by developers.

Having a corporate center to an open source business can be a very good thing, assuring regular updates, a quality Web presence, and software worthy of use by an enterprise.

But they also have detractors. Count former Sun open source executive Simon Phipps (above) among their number. The agreements are coercive, and make some pigs more equal than others in what should be a shared development experience, he writes.

I’m of two minds on this.

In theory  Phipps is right. Projects like the Linux kernel and Mozilla run quite well without contributor agreements.

On the other hand, some pigs are more equal than others, in that they provide the bulk of the work and expense a project may need to survive. In most projects run by Google, the contributions of Google employees go far beyond the combined efforts of their communities.

This is true on smaller projects as well, even those under the GPL. Projects like Appcelerator began as a single company’s dream, and much of the work continues to be done by that firm. That’s one reason CEO Jeff Haynie felt he could unilaterally switch to the Apache license in 2008.

License and contributor agreements don’t have to be done by fiat, of course. Red Hat amended their agreements with some transparency earlier this year. The process can be reassuring and increase the amount of community development.

My own view is that this is much like the open source incline itself, or the open source development incline. That is, there are many places you can fall on the incline, but the further down you go — the less dependent you are on contributor agreements for instance — the greater your community contribution is likely to be.

The answer for your project depends on how much of the development burden your company honestly expects to take on, and how dependent you are on your community for coding. It’s a question, in other words, you need to go into with your eyes wide open.

Whichever side of the table you happen to be on. [poll id="117"]

Firefox 4 beta 5 is now available and offers new hardware acceleration, multimedia and security features.

The latest beta, for example, offers hardware acceleration through Direct 2D by default. Direct 2D was introduced with Windows 7, was made available for Vista, and exploits built-in graphics hardware in Windows PCs with DirectX 10 to enhance performance on graphics heavy websites, said Firefox lead developer Mike Beltzner in his blog today.

“You should notice that some pages are a lot faster and more responsive, in  particular, pages that use advanced animated graphical effects,” added other Firefox developer Bas Schouten in his blog about the feature.

Beta 5 also includes a new Audio API for visualizing audio that exposes raw audio data within the video and audio elements of HTML5. Mozilla claims that developers can use this API to change how users experience the web.

“Until now, people havenât had the ability to interact with sound on the Web in all the creative ways that video and images allow,” Beltzner wrote. “With this new API, developers can read and write raw audio data within the browser, presenting audio information in completely new ways that could allow, for example, for people to visually experience a speech or a song through Firefox.”

Beta 5 also supports the HTTP Strict Transport Security (HSTS) protocol for addiitonal security. “This allows web sites to “insist that they only be loaded over SSL,” Mozilla notes.

Beta 6 is very tentatively set for release on September 10 but that may slip by as much as a week, project leaders indicate. Firefox 4 entered beta testing in June.

I’ve learned one important thing from my 19 year old.

Gaming is important to his generation.

It wasn’t to me. Yes, I knew Steve Jackson (right) at Rice, but he seemed an unusual person. To my son Jackson is a God — it’s like I went to school with Randolph Scott.

Many in my son’s generation feel the same way. Gaming is a big deal to them. It’s like TV was to my generation, like the Internet is to both of us.

But gaming is not like the Internet in one important respect. It is highly proprietary. It is far more proprietary than the worlds of PC and enterprise software, where I make my living.

Most people my son’s age have more respect for game companies’ “stuff” than their parents’ “stuff.” (He’s not constantly bumming $20 off Electronic Arts, I can tell you that.)

Which brings me to projects like PSGroove, a “jailbreak of the PS/3 game machine that lets you make copies of your games and gain ownership of them. It means you don’t lose your games when you lose or break the disk (as people my son’s age are wont to do).

How important are open source projects like this, and to what extent do kids follow them?

The answer, in my house, is disappointing. My son is terribly inconvenienced by Digital Rights Management, and sometimes even complains about it at dinner. But he accepts the concept and has never tried to get around it. He says he’s not a programmer. (Most people my age don’t produce TV shows, either.)

Is he typical?

A poll here at ZDNet might get misleading results, because the fact you’re reading this site means you’re interested in technology, and the fact you’re reading this blog means you’re interested in open source. The sample is skewed.

But some of you doubtless know some people under 25. My son won’t believe this, but I suspect some of you may even be under 25.

So how important is open source to your gaming experience? Have you tried to jailbreak a machine, perhaps because you lost one too many disks or wanted to share something the DRM said not to share? Do you follow the efforts of open source gaming, even a little bit?

Or is the proprietary nature of gaming just something you accept?

Could Google’s PR machine be hiding a weak legal case?

Google made its case recently to The New York Times, arguing that Oracle is trying to “take back” Java, and the Times got a law professor to call the case part of an “open source proxy war” in which open source is used as a weapon.

Bad Oracle. Nasty, nasty Oracle. (Then they play right into it by hiring Mark Hurd.)

Product is also a weapon in this war, I would argue. Google was dumping its Wave project. So how big a deal is it that it puts Wave in a box? Isn’t that more of a paper-or-plastic question? It’s like the rich kid who gives poor kids his old toys for Christmas.

For Google, I suspect, image is everything. This chart makes clear that Android phones are catching up with Apple’s, in terms of mobile data use. But are these really Android phones? Or are they carrier phones built on top of Android?

It is telling that, on the legal case, we still haven’t heard from groups like the Free Software Foundation. Some response is said to be forthcoming. But if this is as easy a call as Google makes out, why hasn’t it been made yet?

I’m as big a fan of Google as anyone. I like their stuff and I like their style. But there’s an old legal saying that when you don’t have the law on your side you argue the facts, and when you don’t have the facts you argue the law, but when you don’t have either you pound the table.

Google is doing a lot of table pounding.

The continuing shortage of Internet bandwidth which drives the network neutrality debate has always puzzled me. (An OpenBTS development kit, from the project’s Sourceforge site.)

Reason being there is no real shortage. The bottleneck has always been in the “last mile,” the on-ramp of your cell phone or your PC, or the router connection your home network uses to reach the outside world.

This is an artificial shortage, the product of a proprietary mindset.

Phone and cable companies own these on-ramps, and the right to create new ones. They use this control to create the idea of a shortage everywhere, to keep prices high, and to threaten content owners with new charges for “premium access” to “their” customers.

In theory they are easy to bypass through the air. But because frequencies are “sold,” meaning rights to use them are offered at auction by the government, the same phone and cable companies wind up controlling the air as well.

We can, if we want, have a virtually unlimited number of on-ramps, wherever we need them, at minimal cost. Proof, again, is being delivered to the Burning Man festival in Nevada this week.

OpenBTS provides the answer. It’s a simple, open source framework that can create a GSM cellular network at one-tenth current costs. It’s licensed under the AGPL.

This year’s set-up uses a third less equipment and half the power of last year’s, but with twice the capacity. With a single LMR-900 tower and a weatherproof travel rack, Range Networks will be able to give all 50,000 participants free cellular calls during the festival, then take the whole thing down when the show is over.

OpenBTS is not the only solution to this problem. OpenBSC also offers a “GSM network in a box,” which can also deliver service on-demand.

With licensed frequencies, an entire urban network must be built-out at once and constantly maintained by one company, which is why cellular bandwidth costs so much. With open source, anyone can add capacity as needed.

If systems like OpenBTS didn’t have to say “mother may I” with licensed carriers in order to serve demand, then demand could be served, defined by hardware instead of property, and the bandwidth shortage would quickly disappear.

We know that’s true because WiFi, whose frequency allocation hasn’t increased in over a decade, can now deliver efficient 100 Mbps networks to hospitals and corporate campuses, which move critical imaging files without interference.

Carriers like AT&T encourage customers to use WiFi whenever possible, claiming they just don’t have the capacity to deliver, even though they own more frequency than WiFi occupies in most areas.

The problem is that we have a regulatory regime which assumes scarcity, which creates bottlenecks, and which rewards monopolists with money coerced through a political process rather than earned through the market.

What’s hilarious is how defenders of this system call it “free enterprise,” and call open source “socialism.”  Open source creates vast markets with lots of players. The current system is government-enforced monopoly.

An open source, and open frequency, mindset in Washington can change that. Something to think about this Labor Day weekend.

Asia Dent (right) must be the most famous face in Poughkeepsie today.

She was photographed by IBM PR recently putting a probe to a new 5.2 GHz chip that is at the heart of the company’s new zEnterprise mainframe, shipping next week.

Florian Mueller calls this the most dangerous product announcement of the century. That’s because zEnterprise could let IBM create a cloud monopoly among large enterprises, assimilating Linux under its mainframe patents.

All this goes back to the Turbo Hercules case, he writes.

You may recall that Turbo Hercules is an IBM mainframe emulator that works on PC-type hardware. When it was a labor of love IBM had no problem with it. When its maker tried to productize it, in the way other open source projects are productized, IBM’s lawyers were on him in a flash.

Hercules founder Roger Bowler then filed a complaint with the EC’s antitrust authorities, saying IBM was illegally tying its mainframe software to hardware.

The new chip makes those chains less burdensome to customers. They’ve got the fastest chip in the world on their side. And if you’re processing bank or credit card transactions (or health claims) that’s a very big deal.

This kind of transaction processing continues to grow rapidly. IBM has driven everyone out of the old mainframe business. Nearly everyone in the space — all the biggest global trade enterprises — have their key functions riding on IBM mainframes.

They’re thrilled with the new IBM mainframes.

Which means that if any scaled enterprise is going into the cloud, it’s taking its mainframe with it. IBM has kindly allowed this new mainframe to assimilate what Linux and Unix can do, without offering any way back.

It’s precisely what critics were accusing Microsoft Sharepoint of doing, but under complete patent protection and control.

In this way, IBM hopes to embrace and extend the cloud into its mainframe monopoly, and keep filing patents on the technology so as to make it an eternal lock on the top end of the business, Mueller writes.

Who is going to rewrite their core processing systems in order to gain the price benefits of true cloud technology?

Which may be why IBM doesn’t want to step up to the plate and be an open source hero.

It is common currency in open source to say that patents are an American problem.

That’s not true. Software patents, or patents on what is expressed in software, are a global problem.

(Picture from our Apple Core blog, co-starring Jason O’Grady and David Morganstern. Always filled with Apple-flavored bloggy goodness.)

This is especially true in the case of Apple, which has sued HTC (and by extension Google) for violating its claimed rights to multitouch technology.

As Florian Mueller explained recently, Apple filed international patent applications for how it operates its touchscreen display in early 2007, and how you unlock the device with gestures on the locked image, in late 2006. It applied for patents on its touch screen interface late last year.

From this it’s clear Apple thinks it has a worldwide monopoly on how the iPhone works, one that could last until late in the next decade. The questions courts must ask are:

1. Does this cover any portable touch screen system, as Apple contends, or just this particular system?
2. Should the patents be considered valid, since Google asserts it was working on its own Android system before the iPhone patents were filed.

There is another important question. Does it respect and reward innovation to give Apple control of all portable touch screen devices, for as long as touch screens may be an interface of choice? Would society have benefited if Microsoft had to wait until the 21st century to deliver Windows, or something like it?

Patent suits are most commonly filed in the U.S., Mueller writes, because this is still the largest technology market, because lawyers are comfortable with the legal system here and because victory usually leads to quick negotiations on global rights.

This leads me to two further questions:

1. If China creates a reasonable patent law framework, will its market eventually draw patent litigation there?
2. If U.S. legislators do return to patent reform, how will that impact technology markets worldwide?

Discuss.

Before Israel was founded in 1948 it made sense to conclude a Passover seder with the words “Next year in Jerusalem.” With Israel a reality the arguments over the phrase have changed. Yet they endure.

Desktop Linux is the same sort of deal. Linux believers always assume that next year will be the year of desktop Linux. Windows followers often chide those who seek Linux with that belief, both here and elsewhere.

Before anyone starts thinking this Catholic boy has changed his stripes, my point is simply that, in the case of desktop Linux, Jerusalem is here.

This is the year.

This is also the year where the definition of a desktop has changed. Apple changed it with the iPhone and, now, the iPad. Microsoft has failed to deliver in both these key areas. Linux has not.

Google gets the credit for that. As I noted yesterday Google Android has soaked up the excess demand for Internet hand-held devices that the iPhone left on the floor. My guess is that, once Chromium comes out, you’ll have the same experience there.

Linux has broken through because Google has the size to go toe-to-toe with either Microsoft or Apple, and push product through distribution. (Remember, there is a price lower than free.)

It’s the compatibility between Chromium and Android, based on Linux, which I think gives the old mouse-and-keyboard upright posture desktop Linux yet-another chance.

Linux Mint and Ubuntu are building the kind of simple-then-power relationship that will exist between Android and Chromium, and which existed in the past between Windows and Windows NT.

Mint offers simplicity and a full application suite. It abstracts all the complexity of the command line, much as Android and Chromium do. Even our own Jason Perlow likes it (and he is hard to please).

What’s still missing is the financial wherewithal to push this through the distribution channel. But with the success of Google as a patron for hand-held Linux, are Microsoft followers certain one can’t be found for the old-fashioned desktop?

My larger point is it doesn’t matter. Either Mint and Ubuntu will gain desktop traction or Google will simply bypass them.

A friend had trouble with their iPhone yesterday and enlisted me in a trip to the Apple Store.

(The Apple store in Lenox Square Mall, Atlanta, from Apple.com.)

Three hours later I realized that Apple is back in the same box Steve Jobs put it in over 25 years ago.

To continue the morning’s baseball theme, It was deja vu all over again.

My friend’s WiFi was on the fritz. The battery was losing power faster than a politician under indictment. No problem, he said. I have an appointment.

The store was tightly packed with people, even though it was Monday afternoon. We were called at 3:18 for an appointment scheduled for 3. After examining the unit our hyper-friendly Apple geek suggested a reboot. No good. Sadly he suggested reloading the operating system. Some 15 minutes later, still no good.

OK, he said, we can fix it, but it will take time because it’s a hardware problem. Wait, my friend said, that’s my home phone. Can’t I just buy another?

Sure, the geek replied. Just get in this line here. How long is this line here, my friend asked. About an hour-and-a-half to two hours, came the reply from the line monitor.

Some 45 minutes later, while my friend frantically used his AT&T data minutes to try and order a new phone online while standing in the Apple phone ordering line, his girlfriend arrived like cavalry to the rescue. She wasn’t under Apple’s spell. She pulled us out and said my friend could buy something later.

Suddenly, in the mall parking lot, a miracle occurred. There, right across the street, was an AT&T store. A company-owned store, its happy little death star sparkling in the sunlight.

Eureka, my friend said. They sell iPhones. So we went over.

It was night-and-day. By which I mean the AT&T store was nearly empty. The help was not overwhelmed. They were waiting for us. We were taken to a man named Scott, who engaged my friend in earnest conversation while I perused the inventory.

Look, I said, this Samsung CaptivaCaptivate costs just what the iPhone would. It’s an Android phone designed to look just like the iPhone, and it seems to have all the same features as the iPhone. Hint, hint. (Thanks to ITGuy08 for catching the misspelling.)

Well, Scott replied, we don’t have any iPhones in stock, but I can get you into a Captiva right now. A half-hour or so later my friend was a happy Android user, asking me if I wanted an iBrick.

There are some important lessons here:

1. Apple claims to be unworried because it is selling iPhones as fast as it can make them. Even faster.
2. Apple is not scaled to meet demand for its product, and certainly not for its retail services.
3. Alternatives with the same look-and-feel are available now.

Back in the 1980s, PC users had to live through 6 years of FUD, waiting for Microsoft or IBM to get their act together and deliver a graphical user interface similar to the Apple Mac, introduced in 1984. Apple had 5 years to own the market, yet its insistence on complete control meant it couldn’t meet demand. Microsoft won.

It’s happening again, Steve. Only it didn’t take Microsoft 6 years to match you. Open source did it in two. And that’s why Android phones now out-sell the iPhone. They’re not better, they’re just available, and you don’t have to go into the 7th circle of Apple Hell to get one.

Over at his other job, our David Gewirtz suggests that, with the absorption of Sun into Oracle, open source badly needs an open source patron and that IBM should apply.

I previously suggested Dell for this role, saying it would be in their business interest to commit to this course. The problem with IBM is somewhat different. (Picture from Wikipedia.)

IBM has learned over the last two decades that it can succeed while avoiding the trips and dramas, the strum and drang, which pass the news cycles in the computer press. Sometimes no news is indeed good news, especially in computing, because it’s not about you but the customer.

But David has a point. All of open source benefited from having important projects in safe hands. With those projects no longer in safe hands a pall has settled, threatening to become a malaise.

IBM is in a unique position to fight that. It has invested heavily in Java and Linux. It passed its Symphony suite over to OpenOffice.org years ago, and now sells support while offering it for download there.

IBM has also benefited from open source through Eclipse and other projects. No other company has earned as much money from open source as IBM. No one else does a better job of giving the lie to the idea that open source is a money loser than IBM.

IBM has become the Stan Musial of open source. (That’s The Man himself, on Wikipedia, during 2008’s Stan Musial Day in St. Louis.)

It’s an open source Hall of Famer, with an excellent reputation, but few people outside its home base know the story, just as Musial is little known outside his hometown and certain retirement homes. (His SI cover this summer was, believe it or not, his first as a solo, although he was the magazine’s Sportsman of the Year for 1957.)

Now, if I can extricate myself from my own childhood we’ll go on.

Despite the nonsense of our Supreme Court (they also think tomatoes are vegetables) companies are not people. They can be immortal, renewing themselves with every generation, adapting constantly, changing with the times.

IBM has proven this. The Watsons are dead. Lou Gerstner is long gone. Elvis has left the building. Yet IBM goes on, its market cap still bigger than Google’s or Oracle’s. If open source needs a hero to step up, IBM is best positioned for the job.

One might even argue that IBM owes this to open source. Having benefited from open source for over a decade, unifying its product lines under Linux, sharing development costs with rivals, and making a ton of money, IBM really should give back.

This is something open source teaches all of us. You benefit more from open source when you give than when you just take. In fact the more you give the more you benefit.

I’m not asking IBM to do something against its interests here. Quite the contrary. It is very much in IBM’s own interest that it step up and lead the open source movement. That’s something IBM representatives have been telling their customers and business partners for some time, that you give in order to get.

OpenBravo recently released an upgraded open source ERP solution aimed at small and medium-sized businesses.

QuickStart 2 features more than 12 business process flows including Order to Cash, Procure to Pay, and Bank Statement to Bank Reconciliation.

The Barcelona-based developer also offers subscription-based pricing as well as cloud deployment for SMBs that want automation and optimization of accounting, sales order processing, inventory, and procurement processes without deploying on-premise.

Version 2 also provides Advanced Payable and Receivables anagement module and a revised payment system that makes data and cash flow mopre transparent to users.

The Great Google is wearing sackcloth and ashes this week, whipping up public resentment against legal rival Oracle by staying away from JavaOne, and quietly encouraging sales of James Gosling’s nifty anti-Oracle t-shirts. (Picture from Cafepress.)

But in publicly portraying itself as the Luke Skywalker of open source (and Larry Ellison as Darth Vader) Google is taking a risk. That’s right, someone might find out Oracle is its father. That would be a real disturbance in the force.

The problem, as Bruce Perens makes clear at his blog, is that this lawsuit isn’t really about open source. Google deliberately violated the patent freedom grant given by Sun, using a user interface toolkit not found in Java ME or Java SE.

Java on the web doesn’t seem to have the problems that Google built into Android, its users can stay within the patent grant without trouble.

Oops. Instead, Android implements the Dalvik Virtual Machine, recompiling  the Harmony class libraries on Apache’s version of Java SE. It then targets the new version at the same markets Oracle has identified.

Or, as Charles Nutter notes in his excellent summation of the issues, “Dalvik is not a JVM…it just plays one on TV.” Google made Java better, which is technically a good thing. But it did so in a legally questionable way.

One point even the fiercest open source advocates will insist on is that your rights to change code are not unlimited. They are defined by a license. If Google tweaked a proprietary version of Java it may lack the commercial rights to what it has done.

In other words, as painful as it may be admit this, Oracle may indeed have a case even Richard Stallman is bound to respect.

Google, who’s your daddy?

Dell is making nice-nice with open source as it seeks a way to compete with a headless HP.

It seems a wise choice. (How did Rabbi Moshe Feinstein get from Wikipedia to here? All will be explained.)

The media focus is currently on 3Par, for which HP has bid $30/share. Dell, which had an offer accepted at $27/share, says it is considering its next move. ($31, anyone, asks our Larry Dignan.)

Analysts say the bidding has reached Crazytown, that it’s now all about corporate ego. (Can a headless company have an ego? Apparently so.) My advice would be to let HP overpay. There is more than one way to skin a server farm.

The issue with 3Par is that both Dell and HP long ago hit upon similar strategies, high-end hardware tied to services. They have been on a collision course ever since Dell overpaid for Perot Systems to match HP’s EDS buy.

But Dell has a second strategy, maybe a better one. Dell is chasing HP out the back end of the “s” curve, looking to offer bargain prices with narrow margins, which pricing theory says is the way to go in a mature market.

Thus Dell is looking for the lowest-cost manufacturing environment it can find, whether in western China or even in India. The idea is if it’s about raw cost Dell is determined to win. (Cheap money is another element in the strategy.) It’s a long way from its old strategy of build-to-order, but it’s a different world.

The Dell Streak fits well into this world. It’s an Android tablet, run under the GPL, which apparently Dell has run afoul of. Rather than argue the point, Dell promises to comply with the license.

Critics are dumping on the Android strategy, but a better play might be to double-down.

Small and medium sized businesses would love to save with open source, but many remain suspicious about support. What they need is not a big bill, but an arm around the shoulder, what we New Yorkers call a rabbi.

A rabbi in this case doesn’t have to be a Jewish teacher. He doesn’t have to be Jewish. He doesn’t even have to be a he. A rabbi in this case means a friend, a trusted adviser, someone who will guide you and sponsor you.

That’s what a lot of medium-sized businesses need if they are to make a true commitment to open source, a rabbi, a friend, an adviser. Someone who knows and will tell them the truth.

By expanding its commitment to open source communities and software, by becoming knowledgeable and offering that knowledge, by sponsoring its customers to the open source world, answering questions, Dell could win a lot of customers at very low cost.

Rabbi Michael?

The MPEG Licensing Association will no longer charge royalties for use of its H.264 codec, when it’s put online for free.

Since the group already had a moratorium on such fees until 2015 the practical impact of this is minimal.

But the business impact could be large, if it decreases interest in WebM.

H.264 is now free as in beer, as opposed to WebM’s open format aimed at HTML5. (Free Beer from Denmark has a royalty-free recipe, and was enjoyed at the 2008 FSCONS launch party. Image from Digital-Rights.net.)

It’s not free for everyone. Those who charge for their video, whether it’s a service like Hulu, a Blu-Ray disc company, or Apple’s iTunes, which wants to charge 99 cents to see shows from free TV, will still pay. (If you’re charging for free beer it’s no longer free to you, is the idea.)

Feh, replied Mozilla, and Google too said feh. Free as in beer is not the issue anyway. Free as in freedom is the issue.

And there, MPEG LA is still spreading the FUD, claiming members of its association hold patents that would cover Google’s VP8, the heart of WebM, and any other video codec programmers might seek to create.

The issue of free as in freedom, in other words, remains, subject to litigation. Whether MPEG LA holds patents on some specific mousetrap designs, or the whole idea of catching a mouse, has yet to be determined.

It is within this cone of uncertainty that videographers now walk. The H.264 codec has been around for a long time, and while WebM offers freedom, that’s merely a declaration that has yet to be tested on a courtroom battlefield.

There is one other issue that bears notice here, namely HTML5. MPEG LA’s royalty scheme has long made it inappropriate for use as a Web standard, which by its nature wants to be royalty free. But now H.264 is royalty free, and defended by a moat of lawyers.

The current HTML5 standards document includes support code for H.264, MPEG 4 and Theora. It does not specify a format, although the group wants to specify one. WebM was created as a project that could be specified, being complete and free as in freedom.

Will free as in beer trump it?

Red Hat announced a strategy for its cloud stack, now called Cloud Foundations Edition One.

It’s about portability and interoperability. In other words it’s about standards. In line with that, Red Hat has submitted its cloud platform as a potential standard for interoperability.

At the heart of the cloud movement was always this idea that you would abstract the complexity of operating systems through virtualization, thus it wouldn’t matter on what specific piece of hardware your data and programs actually lived.

Of course that’s not how computer rivalries work. There are multiple hypervisors, multiple routes to virtualization, multiple ways to manage clouds, and multiple cloud stacks.

When seen in comparison to the ideal of a fully interoperable environment open source has a distinct advantage. When you can see the code, you can link to it more easily than if you can’t. (Try it at home. Wire up your computer with your eyes open, then do it with your eyes shut.)

The cloud strategy puts Red Hat on a collision course with Microsoft, whose Azure cloud says you should trust its portability, and trust its interoperability. Just to turn things up another notch, Red Hat said it would support its business software a full 10 years, as opposed to Microsoft’s five.

Logically Red Hat’s cloud strategy should work. Red Hat is seeking to be the center of the cloud world, while larger vendors swirl around it, and when all the rushing around is done the center is where you want to be.

But the real world is not the ideal plane. Red Hat marketing is indeed Switzerland, if you want to compare the Swiss army to that of, say, Russia. Yes it’s neutral, but if it comes to a fight I’m betting on the bear. Can Red Hat succeed without being, say, bought by IBM?

That’s the risk. It will take more than winning the Dreamworks account to assure a happy ending.

When I wrote disparagingly of Illumos yesterday, I got some well-deserved pushback.

So it may be useful to discuss what Illumos is, and is not.

Illumos is a fork of Open Solaris, but it’s a fork of a special type.

A normal fork takes today’s code and goes off in a new direction with it. Gradually the forked program and the root diverge.

Illumos isn’t going to be like that. It will expand on and support what Oracle offers under the CDDL as “OpenSolaris,” but if Oracle did a complete rewrite over the next few years, rendering what Illumos does in the meantime irrelevant, then Illumos will adapt to the new version.

It’s more of a redneck fork, a fork whose family tree doesn’t branch.

Nexenta is doing all this to push NexentaStor, which depends on some OpenSolaris capabilities. It’s working with two other OpenSolaris distros, BeleniX and Schilli. It’s a semi-independent operation, a software archipelago.

Illumos is sort of a cut-out for disgruntled Solaris customers who don’t want to do business directly with Oracle but still depend on capabilities of the old OpenSolaris for their business models. As Nextenta does.

Not that there’s anything wrong with that. But it doesn’t make OpenSolaris truly open. Contributions to Illumos are subject to being cut off at the knees by the next Oracle release. Illumos will play no part in the Oracle development roadmap.

If you like that and need that, Illumos offers a welcome home for you. But if you like and need true open source software, to which you can contribute and have an equal relationship with, look elsewhere.